SK-2824 release/26.5.1 by skyflow-bharti · Pull Request #305 · skyflowapi/skyflow-java

skyflow-bharti · 2026-05-14T06:46:31Z

No description provided.

SK-2824 Request headers support

@deprecated

…ard compat deprecations, and docs (#308) * docs: add Java SDK nomenclature cleanup design spec Captures the design for public interface renames per the server-side SDK nomenclature changes spec: credential field fallbacks (clientId/keyId/tokenUri), skyflow_id→skyflowId in Get/Query responses, and QueryResponse errors/tokenizedData field additions. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: clarify tokenizedData reasoning in nomenclature cleanup spec Adds explanation for why tokenizedData change is valid despite the Query API currently not returning tokens — based on V1FieldRecords schema support and cross-SDK consistency requirement. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: expand reasoning in Java nomenclature cleanup spec Adds detailed rationale to each design section: why the naming convention matters, why the fallback strategy was chosen over a hard cut, why skyflow_id normalization is inconsistent today, the tokenizedData API schema vs docs discrepancy, and why getErrors() is missing only from QueryResponse. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: correct tokenizedData implementation rationale in spec Clarifies that ignoring record.getTokens() in getFormattedQueryRecord is intentional (Query API cannot return tokens), and that the fix is to promote the toString() hack into a real always-empty field rather than reading from the API response. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: remove tokenizedData from scope in nomenclature cleanup spec Query API cannot return tokens; the toString() inconsistency is not worth fixing since callers have no reason to access tokenizedData programmatically on query results. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add implementation plan for Java SDK nomenclature cleanup 5-task TDD plan covering credential field renames with fallback, skyflow_id normalisation in Get/Query responses, and QueryResponse getErrors() accessor. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId/tokenUri in BearerToken with fallback to old form Add fallback lookup logic so getBearerTokenFromCredentials tries new camelCase keys (clientId, keyId, tokenUri) first and falls back to the legacy all-caps forms (clientID, keyID, tokenURI) for backward compatibility during migration. Add testBearerTokenWithNewFormCredentialKeys to verify the new key form is recognized end-to-end. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId in SignedDataTokens with fallback to old form Add fallback logic to GenerateSignedTokensFromCredentials so both new-form keys (clientId/keyId) and legacy all-caps keys (clientID/keyID) are accepted during migration. Mirrors the pattern already applied to BearerToken.java. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: normalise skyflow_id to skyflowId in Get and Query response maps Insert and Update responses already used camelCase skyflowId; Get and Query were passing through the raw wire-format snake_case key. Add the rename in getFormattedGetRecord and getFormattedQueryRecord, and add reflection-based unit tests to cover both formatters. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: add getErrors() accessor to QueryResponse Adds a private final errors field (always null) and its public accessor to QueryResponse, matching the pattern in GetResponse and InsertResponse. Removes the hardcoded responseObject.add("errors", null) from toString() since serializeNulls on the declared field handles it automatically. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: audit confirms no setFooID/getFooID violations in public API * chore: add Claude Code setup (CLAUDE.md + .claude/) Adapted from skyflow-node PR #305. Includes: - CLAUDE.md with project overview, structure, naming conventions, build commands - .claude/settings.json with PostToolUse compile+checkstyle hooks, PreToolUse generated-code guard, Stop notification; paths are relative (no hardcoded user dirs) - .claude/commands/: code-review, code-security, sdk-sample, test slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: fix gaps and inaccuracies in Claude setup files - CLAUDE.md: add vault/bin/ package, all 5 controllers, pre-existing test failure baseline - settings.json: fix checkstyle hook to print violations (was silently swallowing output with capture_output=True) - sdk-sample.md: fix InsertOptions (doesn't exist), correct sample package structure, correct credential type per feature - code-review.md: fix validation location (controller not build()), fix HashMap rule (SDK pattern is raw HashMaps) - test.md: document pre-existing failures, note checkstyle failsOnError Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 backward compat + deprecation warnings implementation plan 6-task TDD plan: restore skyflow_id key alongside skyflowId in Get/Query responses, add WARN deprecation logs for old credential fields, Javadoc on affected response methods. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation messages to say 'upcoming release' not 'v3' Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add [DEPRECATED] prefix to deprecation log messages per industry standard Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add PM-facing document for v2 public interface changes and deprecation Non-technical overview of credential field renames and skyflow_id response key deprecation — covers customer impact, deprecation warnings, migration guide, and what is NOT changing. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add downloadURL→downloadUrl deprecation to plan and PM doc - Deprecation plan: add Task 6 for GetRequest + DetokenizeRequest with @deprecated annotation approach (compile-time signal vs runtime log) - PM doc: add section 3 for downloadURL rename with migration example Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: remove SDK-level field value null/empty validation from Insert and Update The Skyflow API accepts additionalProperties of Any type including null and empty strings. SDK should not add validation on top of BE — pass through and let BE decide. Removed: - value == null/isEmpty check in validateInsertRequest - value == null/isEmpty check in validateUpdateRequest - value == null/isEmpty check in validateTokensMapWithTokenStrict - values.isEmpty() check in validateInsertRequest (no minItems in API spec) Kept: - values == null check (NPE guard — cannot iterate null array) - key == null/isEmpty check (null keys cannot be JSON-serialized) Deleted 6 tests that asserted on the removed behaviour. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add IDE autocomplete behavior for deprecation signals - PM doc: new section explaining how @deprecated(forRemoval) shows in IntelliJ/VS Code autocomplete (strikethrough, orange underline, tooltip with clickable link to new method) vs runtime WARN log for map keys - Deprecation plan: update downloadURL tasks to use @deprecated(since="2.1", forRemoval=true) + {link} for stronger IDE signal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: segregate code smells into dedicated section in code-review command Splits old Section 6 into: - Section 6: Code quality (actionable correctness checks) - Section 7: Code smells (structural signals, flagged at Smell severity) Code smell catalogue covers: long methods/classes, business logic in data classes, toString() with logic, deep nesting, magic numbers, raw HashMap chains, dead code, stale comments, temporary fields. Severity table clarified: Critical/Bug/Edge Case/Quality = fix before merge; Smell = flag and track. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: add DEPRECATED_SKYFLOW_ID_KEY log entry to InfoLogs * fix: restore skyflow_id key in Get/Query responses for v2 backward compat Both skyflow_id (deprecated) and skyflowId (new form) are now present in response maps simultaneously. WARN log emitted per record. * docs: add deprecation Javadoc for skyflow_id key in GetResponse and QueryResponse * feat: deprecate downloadURL in favour of downloadUrl in GetRequest and DetokenizeRequest Old downloadURL() methods kept as @deprecated(forRemoval=true) delegates. Runtime WARN log emitted on old form usage. 100% test coverage: new form, deprecated form, default value for both classes. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update CLAUDE.md — add code-smell command, update slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation plan and PM doc - credentials permanently supported Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * test: add new-form downloadUrl tests alongside deprecated downloadURL tests Both old (downloadURL) and new (downloadUrl) builder methods tested: - GetTests: 2 new tests for downloadUrl() with cross-assertion on deprecated getDownloadURL() returning same value - DetokenizeTests: 1 new test same pattern - VaultClientTests: 1 new integration test for DetokenizeRequest.downloadUrl() Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: changes to claude * docs: migrate V1-to-V2 guide from README to docs/, update CHANGELOG - Extract 260-line migration section from README.md to docs/migrate_to_v2.md following the pattern established in skyflow-node PR #258 - README now links to docs/migrate_to_v2.md instead of inline content - docs/migrate_to_v2.md adds v2.1+ sections for credential field renames and skyflow_id deprecation (new content) - CHANGELOG.md: add v2.0.4 release notes covering nomenclature changes, backward compat deprecations, and validation removal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG — remove v1 entries, keep only v2.0.4 Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG to point to GitHub and Maven releases Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 banner to README with migration link and EOL notice Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: use release notes link instead of CHANGELOG in banner Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update README banner to v2.1.x announcement Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * revert: remove .claude/ and CLAUDE.md — will be raised as separate PR from main Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove superpowers planning docs from repo Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: ignore docs/superpowers/ — keep planning docs local only Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update cspell config — British English words, Maven flags, ignore paths Added words: serialise/d/s, normalise/d/s/Normalises, behaviour/s/Behaviour, sanitisation, recognised, unrecognised, prioritised Added regex: /-D[A-Za-z][A-Za-z0-9.]*/g to ignore Maven -D flags Added ignorePaths: RUNNING_SAMPLES.md, docs/superpowers/** Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove v2-public-interface-changes.md Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: replace real RSA key with fake key in BearerTokenTests Real 2048-bit RSA key replaced with fake base64 value. Assertion updated from InvalidTokenUri to InvalidKeySpec — still proves all credential fields were resolved (failure is at RSA parsing, not field lookup). Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: guard against null in DetokenizeRequest.downloadUrl(null) Calling .downloadUrl(null) previously stored null in the field, creating an NPE risk for callers who read getDownloadUrl() back without a null check. Now null -> false (matching the default), consistent with the continueOnError(null) guard in the same builder. Added test: testDetokenizeRequestDownloadUrlNullTreatedAsFalse Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove dead error constants after validation removal EmptyValues, EmptyValueInValues, EmptyValueInTokens (ErrorMessage) and EMPTY_VALUES, EMPTY_OR_NULL_VALUE_IN_VALUES, EMPTY_OR_NULL_VALUE_IN_TOKENS (ErrorLogs) are unreachable since the SDK-level null/empty field validation was removed. Deleted to prevent accidental re-wiring. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * Revert "fix: guard against null in DetokenizeRequest.downloadUrl(null)" This reverts commit a802668. * test: add positive tests for permissive Insert validation behaviour Three new tests assert that previously-blocked inputs now pass SDK validation (SDK defers to BE per API spec additionalProperties: Any type): - Empty values array [] passes - Null field value passes - Empty string field value passes Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix banner * feat: port PR #273 changes — raw body support, URL encoding, null-safe request ID - Add EMPTY_STRING, QUOTE, HTTPS_PROTOCOL, CURLY_PLACEHOLDER and HttpUtilityExtra (RAW_BODY_KEY, SDK_GENERATED_PREFIX) to Constants - HttpUtility: conditional content-type header, __raw_body__ passthrough, UUID fallback when server omits x-request-id, URL-encoded form params - Utils: URL-encode path and query params with graceful fallback - Validations: accept String request bodies for non-JSON content types - ConnectionController: wrap String bodies in __raw_body__ for non-JSON content types; fall back to raw string when response is not JSON - InfoLogs: "Bearer token is expired" → "Bearer token is invalid or expired" - HttpUtilityTests: add raw body, no content-type, null request ID and special-character form-encoding tests Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard unsafe Optional.get() calls and remove dead statement in DetectController - Remove dead statement `response.getEntities().get(0).getFile()` that discarded its result and threw IndexOutOfBoundsException/NPE when entities was null or empty (C1); the guarded block below already handles entity processing correctly - Replace unguarded `response.getOutput().get()` in getFirstOutput() with `.orElse(null)` so an absent output Optional returns null instead of throwing NoSuchElementException (C2) - Replace unguarded `firstOutput.getProcessedFileExtension().get().toString()` with `.map(Object::toString).orElse(UNKNOWN)`, reusing the already-computed Optional and matching the safe pattern used for processedFileType one line above (C3) - Add 4 unit tests covering all three fixes via reflection Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard Optional.get() in delete, close FileReader, guard empty form-encode - VaultController.delete: replace unguarded getRecordIdResponse().get() with .orElse(Collections.emptyList()) so an absent field in the API response returns an empty list instead of throwing NoSuchElementException (M2) - BearerToken, SignedDataTokens: wrap FileReader in try/finally to guarantee close() is called even when JsonParser throws JsonSyntaxException; close IOException is intentionally swallowed since the parse already completed (M3) - HttpUtility.formatJsonToFormEncodedString: guard against empty entry set so substring(0, -1) is not called on an empty StringBuilder (M4) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: update README for v2.1 public interface changes - Replace deprecated credential key names in examples: clientID→clientId, keyID→keyId, TokenURI→tokenUri - Update Insert, Get, and Query response examples to use skyflowId (the SDK has always returned skyflowId for Insert; Get/Query now return skyflowId as the primary key) - Add deprecation note after each affected response block: skyflow_id is deprecated and will be removed in an upcoming release Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: do not generate SDK-side requestId when server omits x-request-id header When the response carries no x-request-id header, set requestID to null instead of fabricating a UUID with an SDK-Generated- prefix. A null value is the honest signal to callers that no server request ID is available. Also removes the now-unused UUID import. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: add printWarningLogOnce to suppress per-record deprecation spam Replace per-record printWarningLog calls for DEPRECATED_SKYFLOW_ID_KEY with printWarningLogOnce so the deprecation notice fires at most once per JVM session regardless of result set size. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: fix testSendRequestWithNullRequestId to assert null requestId The SDK no longer generates a fallback UUID when the server omits x-request-id; getRequestID() now returns null. Update the assertion to match the new behaviour. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: fix request_index to requestIndex in README response examples The SDK puts camelCase requestIndex in response maps; the README examples incorrectly showed snake_case request_index. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: readme --------- Co-authored-by: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

@deprecated

* docs: add Java SDK nomenclature cleanup design spec Captures the design for public interface renames per the server-side SDK nomenclature changes spec: credential field fallbacks (clientId/keyId/tokenUri), skyflow_id→skyflowId in Get/Query responses, and QueryResponse errors/tokenizedData field additions. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: clarify tokenizedData reasoning in nomenclature cleanup spec Adds explanation for why tokenizedData change is valid despite the Query API currently not returning tokens — based on V1FieldRecords schema support and cross-SDK consistency requirement. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: expand reasoning in Java nomenclature cleanup spec Adds detailed rationale to each design section: why the naming convention matters, why the fallback strategy was chosen over a hard cut, why skyflow_id normalization is inconsistent today, the tokenizedData API schema vs docs discrepancy, and why getErrors() is missing only from QueryResponse. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: correct tokenizedData implementation rationale in spec Clarifies that ignoring record.getTokens() in getFormattedQueryRecord is intentional (Query API cannot return tokens), and that the fix is to promote the toString() hack into a real always-empty field rather than reading from the API response. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: remove tokenizedData from scope in nomenclature cleanup spec Query API cannot return tokens; the toString() inconsistency is not worth fixing since callers have no reason to access tokenizedData programmatically on query results. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add implementation plan for Java SDK nomenclature cleanup 5-task TDD plan covering credential field renames with fallback, skyflow_id normalisation in Get/Query responses, and QueryResponse getErrors() accessor. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId/tokenUri in BearerToken with fallback to old form Add fallback lookup logic so getBearerTokenFromCredentials tries new camelCase keys (clientId, keyId, tokenUri) first and falls back to the legacy all-caps forms (clientID, keyID, tokenURI) for backward compatibility during migration. Add testBearerTokenWithNewFormCredentialKeys to verify the new key form is recognized end-to-end. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId in SignedDataTokens with fallback to old form Add fallback logic to GenerateSignedTokensFromCredentials so both new-form keys (clientId/keyId) and legacy all-caps keys (clientID/keyID) are accepted during migration. Mirrors the pattern already applied to BearerToken.java. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: normalise skyflow_id to skyflowId in Get and Query response maps Insert and Update responses already used camelCase skyflowId; Get and Query were passing through the raw wire-format snake_case key. Add the rename in getFormattedGetRecord and getFormattedQueryRecord, and add reflection-based unit tests to cover both formatters. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: add getErrors() accessor to QueryResponse Adds a private final errors field (always null) and its public accessor to QueryResponse, matching the pattern in GetResponse and InsertResponse. Removes the hardcoded responseObject.add("errors", null) from toString() since serializeNulls on the declared field handles it automatically. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: audit confirms no setFooID/getFooID violations in public API * chore: add Claude Code setup (CLAUDE.md + .claude/) Adapted from skyflow-node PR #305. Includes: - CLAUDE.md with project overview, structure, naming conventions, build commands - .claude/settings.json with PostToolUse compile+checkstyle hooks, PreToolUse generated-code guard, Stop notification; paths are relative (no hardcoded user dirs) - .claude/commands/: code-review, code-security, sdk-sample, test slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: fix gaps and inaccuracies in Claude setup files - CLAUDE.md: add vault/bin/ package, all 5 controllers, pre-existing test failure baseline - settings.json: fix checkstyle hook to print violations (was silently swallowing output with capture_output=True) - sdk-sample.md: fix InsertOptions (doesn't exist), correct sample package structure, correct credential type per feature - code-review.md: fix validation location (controller not build()), fix HashMap rule (SDK pattern is raw HashMaps) - test.md: document pre-existing failures, note checkstyle failsOnError Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 backward compat + deprecation warnings implementation plan 6-task TDD plan: restore skyflow_id key alongside skyflowId in Get/Query responses, add WARN deprecation logs for old credential fields, Javadoc on affected response methods. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation messages to say 'upcoming release' not 'v3' Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add [DEPRECATED] prefix to deprecation log messages per industry standard Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add PM-facing document for v2 public interface changes and deprecation Non-technical overview of credential field renames and skyflow_id response key deprecation — covers customer impact, deprecation warnings, migration guide, and what is NOT changing. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add downloadURL→downloadUrl deprecation to plan and PM doc - Deprecation plan: add Task 6 for GetRequest + DetokenizeRequest with @deprecated annotation approach (compile-time signal vs runtime log) - PM doc: add section 3 for downloadURL rename with migration example Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: remove SDK-level field value null/empty validation from Insert and Update The Skyflow API accepts additionalProperties of Any type including null and empty strings. SDK should not add validation on top of BE — pass through and let BE decide. Removed: - value == null/isEmpty check in validateInsertRequest - value == null/isEmpty check in validateUpdateRequest - value == null/isEmpty check in validateTokensMapWithTokenStrict - values.isEmpty() check in validateInsertRequest (no minItems in API spec) Kept: - values == null check (NPE guard — cannot iterate null array) - key == null/isEmpty check (null keys cannot be JSON-serialized) Deleted 6 tests that asserted on the removed behaviour. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add IDE autocomplete behavior for deprecation signals - PM doc: new section explaining how @deprecated(forRemoval) shows in IntelliJ/VS Code autocomplete (strikethrough, orange underline, tooltip with clickable link to new method) vs runtime WARN log for map keys - Deprecation plan: update downloadURL tasks to use @deprecated(since="2.1", forRemoval=true) + {link} for stronger IDE signal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: segregate code smells into dedicated section in code-review command Splits old Section 6 into: - Section 6: Code quality (actionable correctness checks) - Section 7: Code smells (structural signals, flagged at Smell severity) Code smell catalogue covers: long methods/classes, business logic in data classes, toString() with logic, deep nesting, magic numbers, raw HashMap chains, dead code, stale comments, temporary fields. Severity table clarified: Critical/Bug/Edge Case/Quality = fix before merge; Smell = flag and track. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: add DEPRECATED_SKYFLOW_ID_KEY log entry to InfoLogs * fix: restore skyflow_id key in Get/Query responses for v2 backward compat Both skyflow_id (deprecated) and skyflowId (new form) are now present in response maps simultaneously. WARN log emitted per record. * docs: add deprecation Javadoc for skyflow_id key in GetResponse and QueryResponse * feat: deprecate downloadURL in favour of downloadUrl in GetRequest and DetokenizeRequest Old downloadURL() methods kept as @deprecated(forRemoval=true) delegates. Runtime WARN log emitted on old form usage. 100% test coverage: new form, deprecated form, default value for both classes. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update CLAUDE.md — add code-smell command, update slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation plan and PM doc - credentials permanently supported Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * test: add new-form downloadUrl tests alongside deprecated downloadURL tests Both old (downloadURL) and new (downloadUrl) builder methods tested: - GetTests: 2 new tests for downloadUrl() with cross-assertion on deprecated getDownloadURL() returning same value - DetokenizeTests: 1 new test same pattern - VaultClientTests: 1 new integration test for DetokenizeRequest.downloadUrl() Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: changes to claude * docs: migrate V1-to-V2 guide from README to docs/, update CHANGELOG - Extract 260-line migration section from README.md to docs/migrate_to_v2.md following the pattern established in skyflow-node PR #258 - README now links to docs/migrate_to_v2.md instead of inline content - docs/migrate_to_v2.md adds v2.1+ sections for credential field renames and skyflow_id deprecation (new content) - CHANGELOG.md: add v2.0.4 release notes covering nomenclature changes, backward compat deprecations, and validation removal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG — remove v1 entries, keep only v2.0.4 Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG to point to GitHub and Maven releases Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 banner to README with migration link and EOL notice Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: use release notes link instead of CHANGELOG in banner Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update README banner to v2.1.x announcement Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * revert: remove .claude/ and CLAUDE.md — will be raised as separate PR from main Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove superpowers planning docs from repo Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: ignore docs/superpowers/ — keep planning docs local only Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update cspell config — British English words, Maven flags, ignore paths Added words: serialise/d/s, normalise/d/s/Normalises, behaviour/s/Behaviour, sanitisation, recognised, unrecognised, prioritised Added regex: /-D[A-Za-z][A-Za-z0-9.]*/g to ignore Maven -D flags Added ignorePaths: RUNNING_SAMPLES.md, docs/superpowers/** Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove v2-public-interface-changes.md Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: replace real RSA key with fake key in BearerTokenTests Real 2048-bit RSA key replaced with fake base64 value. Assertion updated from InvalidTokenUri to InvalidKeySpec — still proves all credential fields were resolved (failure is at RSA parsing, not field lookup). Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: guard against null in DetokenizeRequest.downloadUrl(null) Calling .downloadUrl(null) previously stored null in the field, creating an NPE risk for callers who read getDownloadUrl() back without a null check. Now null -> false (matching the default), consistent with the continueOnError(null) guard in the same builder. Added test: testDetokenizeRequestDownloadUrlNullTreatedAsFalse Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove dead error constants after validation removal EmptyValues, EmptyValueInValues, EmptyValueInTokens (ErrorMessage) and EMPTY_VALUES, EMPTY_OR_NULL_VALUE_IN_VALUES, EMPTY_OR_NULL_VALUE_IN_TOKENS (ErrorLogs) are unreachable since the SDK-level null/empty field validation was removed. Deleted to prevent accidental re-wiring. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * Revert "fix: guard against null in DetokenizeRequest.downloadUrl(null)" This reverts commit a802668. * test: add positive tests for permissive Insert validation behaviour Three new tests assert that previously-blocked inputs now pass SDK validation (SDK defers to BE per API spec additionalProperties: Any type): - Empty values array [] passes - Null field value passes - Empty string field value passes Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix banner * feat: port PR #273 changes — raw body support, URL encoding, null-safe request ID - Add EMPTY_STRING, QUOTE, HTTPS_PROTOCOL, CURLY_PLACEHOLDER and HttpUtilityExtra (RAW_BODY_KEY, SDK_GENERATED_PREFIX) to Constants - HttpUtility: conditional content-type header, __raw_body__ passthrough, UUID fallback when server omits x-request-id, URL-encoded form params - Utils: URL-encode path and query params with graceful fallback - Validations: accept String request bodies for non-JSON content types - ConnectionController: wrap String bodies in __raw_body__ for non-JSON content types; fall back to raw string when response is not JSON - InfoLogs: "Bearer token is expired" → "Bearer token is invalid or expired" - HttpUtilityTests: add raw body, no content-type, null request ID and special-character form-encoding tests Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard unsafe Optional.get() calls and remove dead statement in DetectController - Remove dead statement `response.getEntities().get(0).getFile()` that discarded its result and threw IndexOutOfBoundsException/NPE when entities was null or empty (C1); the guarded block below already handles entity processing correctly - Replace unguarded `response.getOutput().get()` in getFirstOutput() with `.orElse(null)` so an absent output Optional returns null instead of throwing NoSuchElementException (C2) - Replace unguarded `firstOutput.getProcessedFileExtension().get().toString()` with `.map(Object::toString).orElse(UNKNOWN)`, reusing the already-computed Optional and matching the safe pattern used for processedFileType one line above (C3) - Add 4 unit tests covering all three fixes via reflection Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard Optional.get() in delete, close FileReader, guard empty form-encode - VaultController.delete: replace unguarded getRecordIdResponse().get() with .orElse(Collections.emptyList()) so an absent field in the API response returns an empty list instead of throwing NoSuchElementException (M2) - BearerToken, SignedDataTokens: wrap FileReader in try/finally to guarantee close() is called even when JsonParser throws JsonSyntaxException; close IOException is intentionally swallowed since the parse already completed (M3) - HttpUtility.formatJsonToFormEncodedString: guard against empty entry set so substring(0, -1) is not called on an empty StringBuilder (M4) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: update README for v2.1 public interface changes - Replace deprecated credential key names in examples: clientID→clientId, keyID→keyId, TokenURI→tokenUri - Update Insert, Get, and Query response examples to use skyflowId (the SDK has always returned skyflowId for Insert; Get/Query now return skyflowId as the primary key) - Add deprecation note after each affected response block: skyflow_id is deprecated and will be removed in an upcoming release Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: do not generate SDK-side requestId when server omits x-request-id header When the response carries no x-request-id header, set requestID to null instead of fabricating a UUID with an SDK-Generated- prefix. A null value is the honest signal to callers that no server request ID is available. Also removes the now-unused UUID import. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: add printWarningLogOnce to suppress per-record deprecation spam Replace per-record printWarningLog calls for DEPRECATED_SKYFLOW_ID_KEY with printWarningLogOnce so the deprecation notice fires at most once per JVM session regardless of result set size. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: fix testSendRequestWithNullRequestId to assert null requestId The SDK no longer generates a fallback UUID when the server omits x-request-id; getRequestID() now returns null. Update the assertion to match the new behaviour. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: fix request_index to requestIndex in README response examples The SDK puts camelCase requestIndex in response maps; the README examples incorrectly showed snake_case request_index. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: readme * docs: add v2.1.0 upgrade banner and migration guide to v1 README (#310) Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: accept camelCase skyflowId in UpdateRequest and fire deprecation warnings per-request - UpdateRequest.data now accepts 'skyflowId' (preferred) alongside deprecated 'skyflow_id' - extractUpdateSkyflowId() prefers camelCase key, emits WARN-level deprecation on snake_case fallback - Remove printWarningLogOnce — all deprecation warnings now fire on every request via printWarningLog - Add DEPRECATED_SKYFLOW_ID_REQUEST_KEY log entry for request-side key deprecation Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * Revert "docs: add v2.1.0 upgrade banner and migration guide to v1 README (#310)" (#311) This reverts commit 6727625. * feat: prefer skyflowId over skyflow_id in UpdateRequest with deprecation warning When both keys are present, skyflowId is used and a warning is emitted. Adds unit tests for all four cases: camelCase only, snake_case only, both keys (preference), and both keys (map cleanup). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: add getByot() as canonical form; keep getBYOT() as deprecated delegate - TokenMode.getByot() is the new preferred accessor; getBYOT() retained as @deprecated(since="2.1",forRemoval=true) delegate for back-compat - VaultClient updated to call getByot() at all three insert/update sites - UpdateTests: 5 new tests covering camelCase skyflowId key in UpdateRequest - LogUtilLevelTests: 5 new tests verifying WARN log fires at DEBUG/INFO/WARN levels and is suppressed at ERROR (matches Skyflow default) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: add TokenModeTest covering getByot(), deprecated getBYOT() delegate, and toString() Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: emit runtime deprecation warning in getBYOT(); test that it fires - InfoLogs: add DEPRECATED_GET_BYOT constant - TokenMode.getBYOT(): call LogUtil.printWarningLog() so callers see a runtime warning (consistent with downloadURL() and other deprecated methods) - TokenModeTest: replace @SuppressWarnings-only tests with assertions that the warning fires at INFO level and is suppressed at ERROR level Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * refactor: consolidate both-keys deprecation warning into DEPRECATED_SKYFLOW_ID_REQUEST_KEY Remove DEPRECATED_SKYFLOW_ID_BOTH_KEYS; both the snake_case-only and both-keys-present paths in extractUpdateSkyflowId now emit the same DEPRECATED_SKYFLOW_ID_REQUEST_KEY message. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

@deprecated

* docs: add Java SDK nomenclature cleanup design spec Captures the design for public interface renames per the server-side SDK nomenclature changes spec: credential field fallbacks (clientId/keyId/tokenUri), skyflow_id→skyflowId in Get/Query responses, and QueryResponse errors/tokenizedData field additions. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: clarify tokenizedData reasoning in nomenclature cleanup spec Adds explanation for why tokenizedData change is valid despite the Query API currently not returning tokens — based on V1FieldRecords schema support and cross-SDK consistency requirement. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: expand reasoning in Java nomenclature cleanup spec Adds detailed rationale to each design section: why the naming convention matters, why the fallback strategy was chosen over a hard cut, why skyflow_id normalization is inconsistent today, the tokenizedData API schema vs docs discrepancy, and why getErrors() is missing only from QueryResponse. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: correct tokenizedData implementation rationale in spec Clarifies that ignoring record.getTokens() in getFormattedQueryRecord is intentional (Query API cannot return tokens), and that the fix is to promote the toString() hack into a real always-empty field rather than reading from the API response. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: remove tokenizedData from scope in nomenclature cleanup spec Query API cannot return tokens; the toString() inconsistency is not worth fixing since callers have no reason to access tokenizedData programmatically on query results. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add implementation plan for Java SDK nomenclature cleanup 5-task TDD plan covering credential field renames with fallback, skyflow_id normalisation in Get/Query responses, and QueryResponse getErrors() accessor. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId/tokenUri in BearerToken with fallback to old form Add fallback lookup logic so getBearerTokenFromCredentials tries new camelCase keys (clientId, keyId, tokenUri) first and falls back to the legacy all-caps forms (clientID, keyID, tokenURI) for backward compatibility during migration. Add testBearerTokenWithNewFormCredentialKeys to verify the new key form is recognized end-to-end. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId in SignedDataTokens with fallback to old form Add fallback logic to GenerateSignedTokensFromCredentials so both new-form keys (clientId/keyId) and legacy all-caps keys (clientID/keyID) are accepted during migration. Mirrors the pattern already applied to BearerToken.java. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: normalise skyflow_id to skyflowId in Get and Query response maps Insert and Update responses already used camelCase skyflowId; Get and Query were passing through the raw wire-format snake_case key. Add the rename in getFormattedGetRecord and getFormattedQueryRecord, and add reflection-based unit tests to cover both formatters. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: add getErrors() accessor to QueryResponse Adds a private final errors field (always null) and its public accessor to QueryResponse, matching the pattern in GetResponse and InsertResponse. Removes the hardcoded responseObject.add("errors", null) from toString() since serializeNulls on the declared field handles it automatically. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: audit confirms no setFooID/getFooID violations in public API * chore: add Claude Code setup (CLAUDE.md + .claude/) Adapted from skyflow-node PR #305. Includes: - CLAUDE.md with project overview, structure, naming conventions, build commands - .claude/settings.json with PostToolUse compile+checkstyle hooks, PreToolUse generated-code guard, Stop notification; paths are relative (no hardcoded user dirs) - .claude/commands/: code-review, code-security, sdk-sample, test slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: fix gaps and inaccuracies in Claude setup files - CLAUDE.md: add vault/bin/ package, all 5 controllers, pre-existing test failure baseline - settings.json: fix checkstyle hook to print violations (was silently swallowing output with capture_output=True) - sdk-sample.md: fix InsertOptions (doesn't exist), correct sample package structure, correct credential type per feature - code-review.md: fix validation location (controller not build()), fix HashMap rule (SDK pattern is raw HashMaps) - test.md: document pre-existing failures, note checkstyle failsOnError Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 backward compat + deprecation warnings implementation plan 6-task TDD plan: restore skyflow_id key alongside skyflowId in Get/Query responses, add WARN deprecation logs for old credential fields, Javadoc on affected response methods. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation messages to say 'upcoming release' not 'v3' Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add [DEPRECATED] prefix to deprecation log messages per industry standard Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add PM-facing document for v2 public interface changes and deprecation Non-technical overview of credential field renames and skyflow_id response key deprecation — covers customer impact, deprecation warnings, migration guide, and what is NOT changing. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add downloadURL→downloadUrl deprecation to plan and PM doc - Deprecation plan: add Task 6 for GetRequest + DetokenizeRequest with @deprecated annotation approach (compile-time signal vs runtime log) - PM doc: add section 3 for downloadURL rename with migration example Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: remove SDK-level field value null/empty validation from Insert and Update The Skyflow API accepts additionalProperties of Any type including null and empty strings. SDK should not add validation on top of BE — pass through and let BE decide. Removed: - value == null/isEmpty check in validateInsertRequest - value == null/isEmpty check in validateUpdateRequest - value == null/isEmpty check in validateTokensMapWithTokenStrict - values.isEmpty() check in validateInsertRequest (no minItems in API spec) Kept: - values == null check (NPE guard — cannot iterate null array) - key == null/isEmpty check (null keys cannot be JSON-serialized) Deleted 6 tests that asserted on the removed behaviour. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add IDE autocomplete behavior for deprecation signals - PM doc: new section explaining how @deprecated(forRemoval) shows in IntelliJ/VS Code autocomplete (strikethrough, orange underline, tooltip with clickable link to new method) vs runtime WARN log for map keys - Deprecation plan: update downloadURL tasks to use @deprecated(since="2.1", forRemoval=true) + {link} for stronger IDE signal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: segregate code smells into dedicated section in code-review command Splits old Section 6 into: - Section 6: Code quality (actionable correctness checks) - Section 7: Code smells (structural signals, flagged at Smell severity) Code smell catalogue covers: long methods/classes, business logic in data classes, toString() with logic, deep nesting, magic numbers, raw HashMap chains, dead code, stale comments, temporary fields. Severity table clarified: Critical/Bug/Edge Case/Quality = fix before merge; Smell = flag and track. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: add DEPRECATED_SKYFLOW_ID_KEY log entry to InfoLogs * fix: restore skyflow_id key in Get/Query responses for v2 backward compat Both skyflow_id (deprecated) and skyflowId (new form) are now present in response maps simultaneously. WARN log emitted per record. * docs: add deprecation Javadoc for skyflow_id key in GetResponse and QueryResponse * feat: deprecate downloadURL in favour of downloadUrl in GetRequest and DetokenizeRequest Old downloadURL() methods kept as @deprecated(forRemoval=true) delegates. Runtime WARN log emitted on old form usage. 100% test coverage: new form, deprecated form, default value for both classes. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update CLAUDE.md — add code-smell command, update slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation plan and PM doc - credentials permanently supported Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * test: add new-form downloadUrl tests alongside deprecated downloadURL tests Both old (downloadURL) and new (downloadUrl) builder methods tested: - GetTests: 2 new tests for downloadUrl() with cross-assertion on deprecated getDownloadURL() returning same value - DetokenizeTests: 1 new test same pattern - VaultClientTests: 1 new integration test for DetokenizeRequest.downloadUrl() Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: changes to claude * docs: migrate V1-to-V2 guide from README to docs/, update CHANGELOG - Extract 260-line migration section from README.md to docs/migrate_to_v2.md following the pattern established in skyflow-node PR #258 - README now links to docs/migrate_to_v2.md instead of inline content - docs/migrate_to_v2.md adds v2.1+ sections for credential field renames and skyflow_id deprecation (new content) - CHANGELOG.md: add v2.0.4 release notes covering nomenclature changes, backward compat deprecations, and validation removal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG — remove v1 entries, keep only v2.0.4 Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG to point to GitHub and Maven releases Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 banner to README with migration link and EOL notice Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: use release notes link instead of CHANGELOG in banner Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update README banner to v2.1.x announcement Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * revert: remove .claude/ and CLAUDE.md — will be raised as separate PR from main Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove superpowers planning docs from repo Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: ignore docs/superpowers/ — keep planning docs local only Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update cspell config — British English words, Maven flags, ignore paths Added words: serialise/d/s, normalise/d/s/Normalises, behaviour/s/Behaviour, sanitisation, recognised, unrecognised, prioritised Added regex: /-D[A-Za-z][A-Za-z0-9.]*/g to ignore Maven -D flags Added ignorePaths: RUNNING_SAMPLES.md, docs/superpowers/** Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove v2-public-interface-changes.md Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: replace real RSA key with fake key in BearerTokenTests Real 2048-bit RSA key replaced with fake base64 value. Assertion updated from InvalidTokenUri to InvalidKeySpec — still proves all credential fields were resolved (failure is at RSA parsing, not field lookup). Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: guard against null in DetokenizeRequest.downloadUrl(null) Calling .downloadUrl(null) previously stored null in the field, creating an NPE risk for callers who read getDownloadUrl() back without a null check. Now null -> false (matching the default), consistent with the continueOnError(null) guard in the same builder. Added test: testDetokenizeRequestDownloadUrlNullTreatedAsFalse Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove dead error constants after validation removal EmptyValues, EmptyValueInValues, EmptyValueInTokens (ErrorMessage) and EMPTY_VALUES, EMPTY_OR_NULL_VALUE_IN_VALUES, EMPTY_OR_NULL_VALUE_IN_TOKENS (ErrorLogs) are unreachable since the SDK-level null/empty field validation was removed. Deleted to prevent accidental re-wiring. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * Revert "fix: guard against null in DetokenizeRequest.downloadUrl(null)" This reverts commit a802668. * test: add positive tests for permissive Insert validation behaviour Three new tests assert that previously-blocked inputs now pass SDK validation (SDK defers to BE per API spec additionalProperties: Any type): - Empty values array [] passes - Null field value passes - Empty string field value passes Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix banner * feat: port PR #273 changes — raw body support, URL encoding, null-safe request ID - Add EMPTY_STRING, QUOTE, HTTPS_PROTOCOL, CURLY_PLACEHOLDER and HttpUtilityExtra (RAW_BODY_KEY, SDK_GENERATED_PREFIX) to Constants - HttpUtility: conditional content-type header, __raw_body__ passthrough, UUID fallback when server omits x-request-id, URL-encoded form params - Utils: URL-encode path and query params with graceful fallback - Validations: accept String request bodies for non-JSON content types - ConnectionController: wrap String bodies in __raw_body__ for non-JSON content types; fall back to raw string when response is not JSON - InfoLogs: "Bearer token is expired" → "Bearer token is invalid or expired" - HttpUtilityTests: add raw body, no content-type, null request ID and special-character form-encoding tests Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard unsafe Optional.get() calls and remove dead statement in DetectController - Remove dead statement `response.getEntities().get(0).getFile()` that discarded its result and threw IndexOutOfBoundsException/NPE when entities was null or empty (C1); the guarded block below already handles entity processing correctly - Replace unguarded `response.getOutput().get()` in getFirstOutput() with `.orElse(null)` so an absent output Optional returns null instead of throwing NoSuchElementException (C2) - Replace unguarded `firstOutput.getProcessedFileExtension().get().toString()` with `.map(Object::toString).orElse(UNKNOWN)`, reusing the already-computed Optional and matching the safe pattern used for processedFileType one line above (C3) - Add 4 unit tests covering all three fixes via reflection Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard Optional.get() in delete, close FileReader, guard empty form-encode - VaultController.delete: replace unguarded getRecordIdResponse().get() with .orElse(Collections.emptyList()) so an absent field in the API response returns an empty list instead of throwing NoSuchElementException (M2) - BearerToken, SignedDataTokens: wrap FileReader in try/finally to guarantee close() is called even when JsonParser throws JsonSyntaxException; close IOException is intentionally swallowed since the parse already completed (M3) - HttpUtility.formatJsonToFormEncodedString: guard against empty entry set so substring(0, -1) is not called on an empty StringBuilder (M4) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: update README for v2.1 public interface changes - Replace deprecated credential key names in examples: clientID→clientId, keyID→keyId, TokenURI→tokenUri - Update Insert, Get, and Query response examples to use skyflowId (the SDK has always returned skyflowId for Insert; Get/Query now return skyflowId as the primary key) - Add deprecation note after each affected response block: skyflow_id is deprecated and will be removed in an upcoming release Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: do not generate SDK-side requestId when server omits x-request-id header When the response carries no x-request-id header, set requestID to null instead of fabricating a UUID with an SDK-Generated- prefix. A null value is the honest signal to callers that no server request ID is available. Also removes the now-unused UUID import. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: add printWarningLogOnce to suppress per-record deprecation spam Replace per-record printWarningLog calls for DEPRECATED_SKYFLOW_ID_KEY with printWarningLogOnce so the deprecation notice fires at most once per JVM session regardless of result set size. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: fix testSendRequestWithNullRequestId to assert null requestId The SDK no longer generates a fallback UUID when the server omits x-request-id; getRequestID() now returns null. Update the assertion to match the new behaviour. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: fix request_index to requestIndex in README response examples The SDK puts camelCase requestIndex in response maps; the README examples incorrectly showed snake_case request_index. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: readme * docs: add v2.1.0 upgrade banner and migration guide to v1 README (#310) Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: accept camelCase skyflowId in UpdateRequest and fire deprecation warnings per-request - UpdateRequest.data now accepts 'skyflowId' (preferred) alongside deprecated 'skyflow_id' - extractUpdateSkyflowId() prefers camelCase key, emits WARN-level deprecation on snake_case fallback - Remove printWarningLogOnce — all deprecation warnings now fire on every request via printWarningLog - Add DEPRECATED_SKYFLOW_ID_REQUEST_KEY log entry for request-side key deprecation Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * Revert "docs: add v2.1.0 upgrade banner and migration guide to v1 README (#310)" (#311) This reverts commit 6727625. * feat: prefer skyflowId over skyflow_id in UpdateRequest with deprecation warning When both keys are present, skyflowId is used and a warning is emitted. Adds unit tests for all four cases: camelCase only, snake_case only, both keys (preference), and both keys (map cleanup). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: add getByot() as canonical form; keep getBYOT() as deprecated delegate - TokenMode.getByot() is the new preferred accessor; getBYOT() retained as @deprecated(since="2.1",forRemoval=true) delegate for back-compat - VaultClient updated to call getByot() at all three insert/update sites - UpdateTests: 5 new tests covering camelCase skyflowId key in UpdateRequest - LogUtilLevelTests: 5 new tests verifying WARN log fires at DEBUG/INFO/WARN levels and is suppressed at ERROR (matches Skyflow default) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: add TokenModeTest covering getByot(), deprecated getBYOT() delegate, and toString() Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: emit runtime deprecation warning in getBYOT(); test that it fires - InfoLogs: add DEPRECATED_GET_BYOT constant - TokenMode.getBYOT(): call LogUtil.printWarningLog() so callers see a runtime warning (consistent with downloadURL() and other deprecated methods) - TokenModeTest: replace @SuppressWarnings-only tests with assertions that the warning fires at INFO level and is suppressed at ERROR level Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * refactor: consolidate both-keys deprecation warning into DEPRECATED_SKYFLOW_ID_REQUEST_KEY Remove DEPRECATED_SKYFLOW_ID_BOTH_KEYS; both the snake_case-only and both-keys-present paths in extractUpdateSkyflowId now emit the same DEPRECATED_SKYFLOW_ID_REQUEST_KEY message. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: deprecate updateLogLevel(); add setLogLevel() as canonical replacement - Skyflow.setLogLevel() is the new preferred method on the live client - updateLogLevel() retained as @deprecated(since="2.1",forRemoval=true) delegate; emits runtime warning via DEPRECATED_UPDATE_LOG_LEVEL log - InfoLogs: add DEPRECATED_UPDATE_LOG_LEVEL constant - SkyflowException: add JavaDoc documenting validation vs API error paths, null behaviour of requestId/grpcCode, and typical catch pattern - SkyflowTests: add setLogLevel test, deprecation-warning-fires test, and warning-suppressed-at-ERROR test for updateLogLevel() Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * samples: migrate UpdateExample to skyflowId; preserve deprecated skyflow_id form - UpdateExample.java: replace skyflow_id with skyflowId in both data maps - deprecated/UpdateExample.java: retain original skyflow_id pattern with @deprecated annotation and pointer to the current example Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: update migration guide with v2.1 changes Add sections for UpdateRequest skyflowId preference, updateLogLevel and getBYOT deprecations. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: achieve 100% instruction/branch coverage on all public interfaces Closes all reachable coverage gaps across 16 test files. Adds tests for detect response types (ReidentifyTextResponse, DeidentifyTextResponse, EntityInfo, TextIndex, TokenFormat), enum values (DetectOutputTranscriptions, DetectEntities, MaskingMethod, TokenType, DeidentifyFileStatus), InvokeConnectionResponse.getErrors(), DeidentifyFileRequest/Response constructors and builder null-handling, SkyflowException missing JSON branches, InsertRequest.continueOnError(false), and VaultConfig/ ConnectionConfig null-credentials fallback. Two residual JaCoCo false negatives remain: Optional.filter lambda in DetokenizeRecordResponse (unreachable null check) and dead ternary branches in Skyflow$SkyflowClientBuilder whose guards are always non-null after validation. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: emit DEPRECATED_SKYFLOW_ID_KEY warning in getFormattedBatchInsertRecord getFormattedGetRecord and getFormattedQueryRecord both emit the deprecation warning when the API response contains the snake_case skyflow_id key. getFormattedBatchInsertRecord was inconsistent — it silently mapped skyflow_id to skyflowId with no warning. Also future-proofs the batch insert path to handle camelCase skyflowId if the API wire format ever migrates. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: add Error Handling section to README Documents SkyflowException — its two categories (validation vs API errors), all six properties (httpCode, message, httpStatus, grpcCode, requestId, details), the recommended try/catch pattern, and a table distinguishing what is null/empty for validation errors versus API errors. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: remove redundant null-for-validation-errors notes from property table The subsection below the table already covers null/empty behaviour for validation errors; repeating it inline on every row was noise. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

@deprecated

* docs: add Java SDK nomenclature cleanup design spec Captures the design for public interface renames per the server-side SDK nomenclature changes spec: credential field fallbacks (clientId/keyId/tokenUri), skyflow_id→skyflowId in Get/Query responses, and QueryResponse errors/tokenizedData field additions. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: clarify tokenizedData reasoning in nomenclature cleanup spec Adds explanation for why tokenizedData change is valid despite the Query API currently not returning tokens — based on V1FieldRecords schema support and cross-SDK consistency requirement. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: expand reasoning in Java nomenclature cleanup spec Adds detailed rationale to each design section: why the naming convention matters, why the fallback strategy was chosen over a hard cut, why skyflow_id normalization is inconsistent today, the tokenizedData API schema vs docs discrepancy, and why getErrors() is missing only from QueryResponse. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: correct tokenizedData implementation rationale in spec Clarifies that ignoring record.getTokens() in getFormattedQueryRecord is intentional (Query API cannot return tokens), and that the fix is to promote the toString() hack into a real always-empty field rather than reading from the API response. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: remove tokenizedData from scope in nomenclature cleanup spec Query API cannot return tokens; the toString() inconsistency is not worth fixing since callers have no reason to access tokenizedData programmatically on query results. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add implementation plan for Java SDK nomenclature cleanup 5-task TDD plan covering credential field renames with fallback, skyflow_id normalisation in Get/Query responses, and QueryResponse getErrors() accessor. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId/tokenUri in BearerToken with fallback to old form Add fallback lookup logic so getBearerTokenFromCredentials tries new camelCase keys (clientId, keyId, tokenUri) first and falls back to the legacy all-caps forms (clientID, keyID, tokenURI) for backward compatibility during migration. Add testBearerTokenWithNewFormCredentialKeys to verify the new key form is recognized end-to-end. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId in SignedDataTokens with fallback to old form Add fallback logic to GenerateSignedTokensFromCredentials so both new-form keys (clientId/keyId) and legacy all-caps keys (clientID/keyID) are accepted during migration. Mirrors the pattern already applied to BearerToken.java. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: normalise skyflow_id to skyflowId in Get and Query response maps Insert and Update responses already used camelCase skyflowId; Get and Query were passing through the raw wire-format snake_case key. Add the rename in getFormattedGetRecord and getFormattedQueryRecord, and add reflection-based unit tests to cover both formatters. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: add getErrors() accessor to QueryResponse Adds a private final errors field (always null) and its public accessor to QueryResponse, matching the pattern in GetResponse and InsertResponse. Removes the hardcoded responseObject.add("errors", null) from toString() since serializeNulls on the declared field handles it automatically. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: audit confirms no setFooID/getFooID violations in public API * chore: add Claude Code setup (CLAUDE.md + .claude/) Adapted from skyflow-node PR #305. Includes: - CLAUDE.md with project overview, structure, naming conventions, build commands - .claude/settings.json with PostToolUse compile+checkstyle hooks, PreToolUse generated-code guard, Stop notification; paths are relative (no hardcoded user dirs) - .claude/commands/: code-review, code-security, sdk-sample, test slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: fix gaps and inaccuracies in Claude setup files - CLAUDE.md: add vault/bin/ package, all 5 controllers, pre-existing test failure baseline - settings.json: fix checkstyle hook to print violations (was silently swallowing output with capture_output=True) - sdk-sample.md: fix InsertOptions (doesn't exist), correct sample package structure, correct credential type per feature - code-review.md: fix validation location (controller not build()), fix HashMap rule (SDK pattern is raw HashMaps) - test.md: document pre-existing failures, note checkstyle failsOnError Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 backward compat + deprecation warnings implementation plan 6-task TDD plan: restore skyflow_id key alongside skyflowId in Get/Query responses, add WARN deprecation logs for old credential fields, Javadoc on affected response methods. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation messages to say 'upcoming release' not 'v3' Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add [DEPRECATED] prefix to deprecation log messages per industry standard Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add PM-facing document for v2 public interface changes and deprecation Non-technical overview of credential field renames and skyflow_id response key deprecation — covers customer impact, deprecation warnings, migration guide, and what is NOT changing. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add downloadURL→downloadUrl deprecation to plan and PM doc - Deprecation plan: add Task 6 for GetRequest + DetokenizeRequest with @deprecated annotation approach (compile-time signal vs runtime log) - PM doc: add section 3 for downloadURL rename with migration example Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: remove SDK-level field value null/empty validation from Insert and Update The Skyflow API accepts additionalProperties of Any type including null and empty strings. SDK should not add validation on top of BE — pass through and let BE decide. Removed: - value == null/isEmpty check in validateInsertRequest - value == null/isEmpty check in validateUpdateRequest - value == null/isEmpty check in validateTokensMapWithTokenStrict - values.isEmpty() check in validateInsertRequest (no minItems in API spec) Kept: - values == null check (NPE guard — cannot iterate null array) - key == null/isEmpty check (null keys cannot be JSON-serialized) Deleted 6 tests that asserted on the removed behaviour. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add IDE autocomplete behavior for deprecation signals - PM doc: new section explaining how @deprecated(forRemoval) shows in IntelliJ/VS Code autocomplete (strikethrough, orange underline, tooltip with clickable link to new method) vs runtime WARN log for map keys - Deprecation plan: update downloadURL tasks to use @deprecated(since="2.1", forRemoval=true) + {link} for stronger IDE signal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: segregate code smells into dedicated section in code-review command Splits old Section 6 into: - Section 6: Code quality (actionable correctness checks) - Section 7: Code smells (structural signals, flagged at Smell severity) Code smell catalogue covers: long methods/classes, business logic in data classes, toString() with logic, deep nesting, magic numbers, raw HashMap chains, dead code, stale comments, temporary fields. Severity table clarified: Critical/Bug/Edge Case/Quality = fix before merge; Smell = flag and track. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: add DEPRECATED_SKYFLOW_ID_KEY log entry to InfoLogs * fix: restore skyflow_id key in Get/Query responses for v2 backward compat Both skyflow_id (deprecated) and skyflowId (new form) are now present in response maps simultaneously. WARN log emitted per record. * docs: add deprecation Javadoc for skyflow_id key in GetResponse and QueryResponse * feat: deprecate downloadURL in favour of downloadUrl in GetRequest and DetokenizeRequest Old downloadURL() methods kept as @deprecated(forRemoval=true) delegates. Runtime WARN log emitted on old form usage. 100% test coverage: new form, deprecated form, default value for both classes. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update CLAUDE.md — add code-smell command, update slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation plan and PM doc - credentials permanently supported Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * test: add new-form downloadUrl tests alongside deprecated downloadURL tests Both old (downloadURL) and new (downloadUrl) builder methods tested: - GetTests: 2 new tests for downloadUrl() with cross-assertion on deprecated getDownloadURL() returning same value - DetokenizeTests: 1 new test same pattern - VaultClientTests: 1 new integration test for DetokenizeRequest.downloadUrl() Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: changes to claude * docs: migrate V1-to-V2 guide from README to docs/, update CHANGELOG - Extract 260-line migration section from README.md to docs/migrate_to_v2.md following the pattern established in skyflow-node PR #258 - README now links to docs/migrate_to_v2.md instead of inline content - docs/migrate_to_v2.md adds v2.1+ sections for credential field renames and skyflow_id deprecation (new content) - CHANGELOG.md: add v2.0.4 release notes covering nomenclature changes, backward compat deprecations, and validation removal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG — remove v1 entries, keep only v2.0.4 Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG to point to GitHub and Maven releases Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 banner to README with migration link and EOL notice Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: use release notes link instead of CHANGELOG in banner Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update README banner to v2.1.x announcement Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * revert: remove .claude/ and CLAUDE.md — will be raised as separate PR from main Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove superpowers planning docs from repo Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: ignore docs/superpowers/ — keep planning docs local only Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update cspell config — British English words, Maven flags, ignore paths Added words: serialise/d/s, normalise/d/s/Normalises, behaviour/s/Behaviour, sanitisation, recognised, unrecognised, prioritised Added regex: /-D[A-Za-z][A-Za-z0-9.]*/g to ignore Maven -D flags Added ignorePaths: RUNNING_SAMPLES.md, docs/superpowers/** Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove v2-public-interface-changes.md Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: replace real RSA key with fake key in BearerTokenTests Real 2048-bit RSA key replaced with fake base64 value. Assertion updated from InvalidTokenUri to InvalidKeySpec — still proves all credential fields were resolved (failure is at RSA parsing, not field lookup). Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: guard against null in DetokenizeRequest.downloadUrl(null) Calling .downloadUrl(null) previously stored null in the field, creating an NPE risk for callers who read getDownloadUrl() back without a null check. Now null -> false (matching the default), consistent with the continueOnError(null) guard in the same builder. Added test: testDetokenizeRequestDownloadUrlNullTreatedAsFalse Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove dead error constants after validation removal EmptyValues, EmptyValueInValues, EmptyValueInTokens (ErrorMessage) and EMPTY_VALUES, EMPTY_OR_NULL_VALUE_IN_VALUES, EMPTY_OR_NULL_VALUE_IN_TOKENS (ErrorLogs) are unreachable since the SDK-level null/empty field validation was removed. Deleted to prevent accidental re-wiring. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * Revert "fix: guard against null in DetokenizeRequest.downloadUrl(null)" This reverts commit a802668. * test: add positive tests for permissive Insert validation behaviour Three new tests assert that previously-blocked inputs now pass SDK validation (SDK defers to BE per API spec additionalProperties: Any type): - Empty values array [] passes - Null field value passes - Empty string field value passes Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix banner * feat: port PR #273 changes — raw body support, URL encoding, null-safe request ID - Add EMPTY_STRING, QUOTE, HTTPS_PROTOCOL, CURLY_PLACEHOLDER and HttpUtilityExtra (RAW_BODY_KEY, SDK_GENERATED_PREFIX) to Constants - HttpUtility: conditional content-type header, __raw_body__ passthrough, UUID fallback when server omits x-request-id, URL-encoded form params - Utils: URL-encode path and query params with graceful fallback - Validations: accept String request bodies for non-JSON content types - ConnectionController: wrap String bodies in __raw_body__ for non-JSON content types; fall back to raw string when response is not JSON - InfoLogs: "Bearer token is expired" → "Bearer token is invalid or expired" - HttpUtilityTests: add raw body, no content-type, null request ID and special-character form-encoding tests Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard unsafe Optional.get() calls and remove dead statement in DetectController - Remove dead statement `response.getEntities().get(0).getFile()` that discarded its result and threw IndexOutOfBoundsException/NPE when entities was null or empty (C1); the guarded block below already handles entity processing correctly - Replace unguarded `response.getOutput().get()` in getFirstOutput() with `.orElse(null)` so an absent output Optional returns null instead of throwing NoSuchElementException (C2) - Replace unguarded `firstOutput.getProcessedFileExtension().get().toString()` with `.map(Object::toString).orElse(UNKNOWN)`, reusing the already-computed Optional and matching the safe pattern used for processedFileType one line above (C3) - Add 4 unit tests covering all three fixes via reflection Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard Optional.get() in delete, close FileReader, guard empty form-encode - VaultController.delete: replace unguarded getRecordIdResponse().get() with .orElse(Collections.emptyList()) so an absent field in the API response returns an empty list instead of throwing NoSuchElementException (M2) - BearerToken, SignedDataTokens: wrap FileReader in try/finally to guarantee close() is called even when JsonParser throws JsonSyntaxException; close IOException is intentionally swallowed since the parse already completed (M3) - HttpUtility.formatJsonToFormEncodedString: guard against empty entry set so substring(0, -1) is not called on an empty StringBuilder (M4) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: update README for v2.1 public interface changes - Replace deprecated credential key names in examples: clientID→clientId, keyID→keyId, TokenURI→tokenUri - Update Insert, Get, and Query response examples to use skyflowId (the SDK has always returned skyflowId for Insert; Get/Query now return skyflowId as the primary key) - Add deprecation note after each affected response block: skyflow_id is deprecated and will be removed in an upcoming release Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: do not generate SDK-side requestId when server omits x-request-id header When the response carries no x-request-id header, set requestID to null instead of fabricating a UUID with an SDK-Generated- prefix. A null value is the honest signal to callers that no server request ID is available. Also removes the now-unused UUID import. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: add printWarningLogOnce to suppress per-record deprecation spam Replace per-record printWarningLog calls for DEPRECATED_SKYFLOW_ID_KEY with printWarningLogOnce so the deprecation notice fires at most once per JVM session regardless of result set size. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: fix testSendRequestWithNullRequestId to assert null requestId The SDK no longer generates a fallback UUID when the server omits x-request-id; getRequestID() now returns null. Update the assertion to match the new behaviour. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: fix request_index to requestIndex in README response examples The SDK puts camelCase requestIndex in response maps; the README examples incorrectly showed snake_case request_index. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: readme * docs: add v2.1.0 upgrade banner and migration guide to v1 README (#310) Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: accept camelCase skyflowId in UpdateRequest and fire deprecation warnings per-request - UpdateRequest.data now accepts 'skyflowId' (preferred) alongside deprecated 'skyflow_id' - extractUpdateSkyflowId() prefers camelCase key, emits WARN-level deprecation on snake_case fallback - Remove printWarningLogOnce — all deprecation warnings now fire on every request via printWarningLog - Add DEPRECATED_SKYFLOW_ID_REQUEST_KEY log entry for request-side key deprecation Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * Revert "docs: add v2.1.0 upgrade banner and migration guide to v1 README (#310)" (#311) This reverts commit 6727625. * feat: prefer skyflowId over skyflow_id in UpdateRequest with deprecation warning When both keys are present, skyflowId is used and a warning is emitted. Adds unit tests for all four cases: camelCase only, snake_case only, both keys (preference), and both keys (map cleanup). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: add getByot() as canonical form; keep getBYOT() as deprecated delegate - TokenMode.getByot() is the new preferred accessor; getBYOT() retained as @deprecated(since="2.1",forRemoval=true) delegate for back-compat - VaultClient updated to call getByot() at all three insert/update sites - UpdateTests: 5 new tests covering camelCase skyflowId key in UpdateRequest - LogUtilLevelTests: 5 new tests verifying WARN log fires at DEBUG/INFO/WARN levels and is suppressed at ERROR (matches Skyflow default) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: add TokenModeTest covering getByot(), deprecated getBYOT() delegate, and toString() Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: emit runtime deprecation warning in getBYOT(); test that it fires - InfoLogs: add DEPRECATED_GET_BYOT constant - TokenMode.getBYOT(): call LogUtil.printWarningLog() so callers see a runtime warning (consistent with downloadURL() and other deprecated methods) - TokenModeTest: replace @SuppressWarnings-only tests with assertions that the warning fires at INFO level and is suppressed at ERROR level Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * refactor: consolidate both-keys deprecation warning into DEPRECATED_SKYFLOW_ID_REQUEST_KEY Remove DEPRECATED_SKYFLOW_ID_BOTH_KEYS; both the snake_case-only and both-keys-present paths in extractUpdateSkyflowId now emit the same DEPRECATED_SKYFLOW_ID_REQUEST_KEY message. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: deprecate updateLogLevel(); add setLogLevel() as canonical replacement - Skyflow.setLogLevel() is the new preferred method on the live client - updateLogLevel() retained as @deprecated(since="2.1",forRemoval=true) delegate; emits runtime warning via DEPRECATED_UPDATE_LOG_LEVEL log - InfoLogs: add DEPRECATED_UPDATE_LOG_LEVEL constant - SkyflowException: add JavaDoc documenting validation vs API error paths, null behaviour of requestId/grpcCode, and typical catch pattern - SkyflowTests: add setLogLevel test, deprecation-warning-fires test, and warning-suppressed-at-ERROR test for updateLogLevel() Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * samples: migrate UpdateExample to skyflowId; preserve deprecated skyflow_id form - UpdateExample.java: replace skyflow_id with skyflowId in both data maps - deprecated/UpdateExample.java: retain original skyflow_id pattern with @deprecated annotation and pointer to the current example Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: update migration guide with v2.1 changes Add sections for UpdateRequest skyflowId preference, updateLogLevel and getBYOT deprecations. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: achieve 100% instruction/branch coverage on all public interfaces Closes all reachable coverage gaps across 16 test files. Adds tests for detect response types (ReidentifyTextResponse, DeidentifyTextResponse, EntityInfo, TextIndex, TokenFormat), enum values (DetectOutputTranscriptions, DetectEntities, MaskingMethod, TokenType, DeidentifyFileStatus), InvokeConnectionResponse.getErrors(), DeidentifyFileRequest/Response constructors and builder null-handling, SkyflowException missing JSON branches, InsertRequest.continueOnError(false), and VaultConfig/ ConnectionConfig null-credentials fallback. Two residual JaCoCo false negatives remain: Optional.filter lambda in DetokenizeRecordResponse (unreachable null check) and dead ternary branches in Skyflow$SkyflowClientBuilder whose guards are always non-null after validation. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: emit DEPRECATED_SKYFLOW_ID_KEY warning in getFormattedBatchInsertRecord getFormattedGetRecord and getFormattedQueryRecord both emit the deprecation warning when the API response contains the snake_case skyflow_id key. getFormattedBatchInsertRecord was inconsistent — it silently mapped skyflow_id to skyflowId with no warning. Also future-proofs the batch insert path to handle camelCase skyflowId if the API wire format ever migrates. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: add Error Handling section to README Documents SkyflowException — its two categories (validation vs API errors), all six properties (httpCode, message, httpStatus, grpcCode, requestId, details), the recommended try/catch pattern, and a table distinguishing what is null/empty for validation errors versus API errors. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: remove redundant null-for-validation-errors notes from property table The subsection below the table already covers null/empty behaviour for validation errors; repeating it inline on every row was noise. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: add downloadURL→downloadUrl rename to migration guide method renames table Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * samples: add deprecated DetokenizeExample; show updateLogLevel in deprecated UpdateExample DetokenizeExample demonstrates the deprecated downloadURL() builder method. UpdateExample now also demonstrates the deprecated updateLogLevel() client method alongside the existing skyflow_id key deprecation. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * samples: add deprecated GetExample showing legacy skyflow_id response key Demonstrates reading the deprecated "skyflow_id" key alongside the preferred "skyflowId" key in the Get response map, retained for reference during the deprecation window. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: add unit tests for VaultController, DetectController, ConnectionController Adds comprehensive unit tests across the three controller classes to improve JaCoCo instruction coverage from ~25% to 87%/43%/87%: - VaultController: full happy-path + API-error tests for insert (bulk and batch), detokenize, get, update, delete, query, tokenize; also unit tests for extractUpdateSkyflowId, getFormattedGetRecord, getFormattedQueryRecord, downloadUrl accessors, and skyflowId normalisation. All private helpers exercised via reflection. - ConnectionController: happy-path tests for all HTTP methods (GET, POST, PUT, DELETE) plus non-JSON response wrapping, requestId in metadata, and all validation-failure paths (empty headers/params/ body, null header value, IOException, SkyflowException propagation). - DetectController: happy-path + API-error tests for deidentifyText, reidentifyText, and getDetectRun; drives extractBodyAsString to 100%. - pom.xml: prefix surefire argLine with ${argLine} so JaCoCo's agent is properly injected into the test JVM and coverage data is accurate. Total test count: 462 → 514. Pre-existing 5 failures unchanged. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: boost instruction coverage for key public-interface methods VaultController: - getFormattedBatchInsertRecord: 77% → 100% (deprecated skyflow_id key path, error-body path, mixed success+error batch result) - detokenize: 95% → 100% (mixed success+error response, empty records list) - get: 97% → 100% (added redactionType to cover non-null ternary branch) - insert: 85% → 99% (batch all-error path, deprecated-key path, mixed result; remaining 2 instr are unreachable dead-code ternary branch) DetectController: - parseDeidentifyFileResponse: 66% → 95% (wordCharacterCount branch L272-273, processedFile decode+write branch L283-291; IOException catch at L290-291 requires mocking NIO and is left uncovered) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: remove unreachable ternary in insert return path After the `if (insertedFields.isEmpty()) { return; }` guard, insertedFields is guaranteed non-empty, making `insertedFields.isEmpty() ? null : insertedFields` dead code. Simplify to just `insertedFields`, bringing insert to 100% coverage. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: add DetectController deidentifyFile/processFileByType coverage tests Adds 19 new test methods covering deidentifyFile validation, happy paths, IN_PROGRESS timeout, all processFileByType extension branches, and error paths. Also fixes null-dereference NPE in Validations.validateDeidentifyFileRequest when waitTime is null. Brings DetectController instruction coverage from 49% to 92%. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: add pollForResults retry and null-outputDir coverage for DetectController Adds three more tests: - processedFile present with no outputDirectory (covers line 146/168 branches) - IN_PROGRESS → SUCCESS with waitTime=2 (covers lines 218-229 if-branch retry) - IN_PROGRESS → SUCCESS with waitTime=3 (covers lines 225-226 else-branch retry) Brings DetectController to 95% instruction / 96% line / 86% branch coverage. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: add uploadFile coverage for all three FileInput paths in VaultController Tests cover getFileForFileUpload via filePath, base64, and fileObject inputs, plus an API error path. Brings VaultClient instruction coverage from 89% to 91%. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: cover entityUniqueCounter branches in VaultClient request builders Adds 6 tests (txt/mp3/pdf/jpg/csv/dcm extensions with tokenFormat.entityUniqueCounter) to cover the entityUniqueCounter mapping branches across all deidentify request-builder methods. Brings VaultClient instruction coverage to 96%. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: add JWT decoded() coverage to TokenTests Adds two tests with hand-crafted fake JWTs (no env var dependency): - 3-part token with expired payload (exp=1) → covers isExpired=true path - 3-part token with far-future payload (exp=9999999999) → covers isExpired=false path Brings Token instruction coverage from 58% to 95%. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: add appendRequestId and nested JSON body coverage for HttpUtility Tests appendRequestId with non-null, null, and empty requestId; also covers convertJsonToMap recursive path with nested JSON objects via sendRequest. Brings HttpUtility from 87% to 91% instruction coverage. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: emit deprecation WARN when legacy clientID/keyID/tokenURI credential fields are used Adds WARN logs in BearerToken and SignedDataTokens when the old all-caps field names (clientID, keyID, tokenURI) trigger the fallback path, so callers know to migrate to the camelCase forms (clientId, keyId, tokenUri). Also adds the three corresponding InfoLogs entries. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: achieve 100% instruction coverage on all public interface packages Adds tests across ConnectionClient, VaultClient, and Skyflow builder to cover all previously missed code paths and reach 100% JaCoCo instruction coverage on every public-facing package (com.skyflow, config, enums, errors, vault.*). New test files: - ConnectionClientDotenvTests: PowerMock tests for dotenv credential path - VaultClientDotenvTests: actual .env file approach (Java 21 compatible) - BinAuditTests: GetBin/ListEvent constructor coverage Extended tests: - ConnectionClientTests: apiKey reuse, bearer token reuse, credential change reset, and no-credentials (DotenvException) paths - SkyflowTests: findAndUpdateVaultConfig/findAndUpdateConnectionConfig null-fallback branches via anonymous subclasses and reflection - VaultClientTests: file upload null return, masking method, entity scores lambda, credential change reset, HTTP interceptor, and getDeidentifyGenericFileRequest branch coverage pom.xml: switch surefire argLine to @{argLine} (late-binding) so JaCoCo agent arg is injected correctly; add --add-opens java.base/java.io for PowerMock on Java 21. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * chore: add test abbreviations and sample IDs to cspell allowlist Fixes cspell failures in CI caused by abbreviations used in test string literals (nocreds, nodir, detok) and a hardcoded cluster ID in tracked sample files (qhdmceurtnlz). Also adds ngrok and obac which appear in sample config files. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: propagate SkyflowException from prioritiseCredentials without wrapping When Dotenv loads successfully but SKYFLOW_CREDENTIALS is absent, SkyflowException thrown inside the try block was being caught by the generic catch(Exception e) handler and wrapped in RuntimeException. Add an explicit catch(SkyflowException e) before the generic handler so it propagates directly, matching the declared throws signature. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: properly save/restore .env content in dotenv tests and remove PowerMock VaultClientDotenvTests and ConnectionClientDotenvTests previously only tracked whether .env existed, never restoring its original content. In CI the workflow pre-creates .env with SKYFLOW_CREDENTIALS and TEST_REUSABLE_TOKEN secrets; overwriting it without restore corrupted subsequent tests (ConnectionClientTests.testSetBearerToken et al). Fix: save the full byte content of .env in @before and restore it in @after, so CI's .env is intact after each test. Also removed PowerMock from ConnectionClientDotenvTests — the static mock of Dotenv.class leaked class-loader state into the subsequent ConnectionClientTests class. Replaced with the same real .env file approach used by VaultClientDotenvTests. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: relax noCredentials test assertion to be environment-agnostic In CI the .env file contains SKYFLOW_CREDENTIALS so prioritiseCredentials succeeds, and the subsequent generateBearerToken fails with a private-key error instead of EmptyCredentials. The test was correctly catching SkyflowException but failing on the specific message comparison. Remove the assertEquals on the message — the important invariant is that a SkyflowException is thrown when no explicit credentials are configured, regardless of which specific credential error the environment produces. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

* docs: add Java SDK nomenclature cleanup design spec Captures the design for public interface renames per the server-side SDK nomenclature changes spec: credential field fallbacks (clientId/keyId/tokenUri), skyflow_id→skyflowId in Get/Query responses, and QueryResponse errors/tokenizedData field additions. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: clarify tokenizedData reasoning in nomenclature cleanup spec Adds explanation for why tokenizedData change is valid despite the Query API currently not returning tokens — based on V1FieldRecords schema support and cross-SDK consistency requirement. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: expand reasoning in Java nomenclature cleanup spec Adds detailed rationale to each design section: why the naming convention matters, why the fallback strategy was chosen over a hard cut, why skyflow_id normalization is inconsistent today, the tokenizedData API schema vs docs discrepancy, and why getErrors() is missing only from QueryResponse. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: correct tokenizedData implementation rationale in spec Clarifies that ignoring record.getTokens() in getFormattedQueryRecord is intentional (Query API cannot return tokens), and that the fix is to promote the toString() hack into a real always-empty field rather than reading from the API response. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: remove tokenizedData from scope in nomenclature cleanup spec Query API cannot return tokens; the toString() inconsistency is not worth fixing since callers have no reason to access tokenizedData programmatically on query results. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add implementation plan for Java SDK nomenclature cleanup 5-task TDD plan covering credential field renames with fallback, skyflow_id normalisation in Get/Query responses, and QueryResponse getErrors() accessor. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId/tokenUri in BearerToken with fallback to old form Add fallback lookup logic so getBearerTokenFromCredentials tries new camelCase keys (clientId, keyId, tokenUri) first and falls back to the legacy all-caps forms (clientID, keyID, tokenURI) for backward compatibility during migration. Add testBearerTokenWithNewFormCredentialKeys to verify the new key form is recognized end-to-end. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId in SignedDataTokens with fallback to old form Add fallback logic to GenerateSignedTokensFromCredentials so both new-form keys (clientId/keyId) and legacy all-caps keys (clientID/keyID) are accepted during migration. Mirrors the pattern already applied to BearerToken.java. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: normalise skyflow_id to skyflowId in Get and Query response maps Insert and Update responses already used camelCase skyflowId; Get and Query were passing through the raw wire-format snake_case key. Add the rename in getFormattedGetRecord and getFormattedQueryRecord, and add reflection-based unit tests to cover both formatters. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: add getErrors() accessor to QueryResponse Adds a private final errors field (always null) and its public accessor to QueryResponse, matching the pattern in GetResponse and InsertResponse. Removes the hardcoded responseObject.add("errors", null) from toString() since serializeNulls on the declared field handles it automatically. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: audit confirms no setFooID/getFooID violations in public API * chore: add Claude Code setup (CLAUDE.md + .claude/) Adapted from skyflow-node PR #305. Includes: - CLAUDE.md with project overview, structure, naming conventions, build commands - .claude/settings.json with PostToolUse compile+checkstyle hooks, PreToolUse generated-code guard, Stop notification; paths are relative (no hardcoded user dirs) - .claude/commands/: code-review, code-security, sdk-sample, test slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: fix gaps and inaccuracies in Claude setup files - CLAUDE.md: add vault/bin/ package, all 5 controllers, pre-existing test failure baseline - settings.json: fix checkstyle hook to print violations (was silently swallowing output with capture_output=True) - sdk-sample.md: fix InsertOptions (doesn't exist), correct sample package structure, correct credential type per feature - code-review.md: fix validation location (controller not build()), fix HashMap rule (SDK pattern is raw HashMaps) - test.md: document pre-existing failures, note checkstyle failsOnError Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 backward compat + deprecation warnings implementation plan 6-task TDD plan: restore skyflow_id key alongside skyflowId in Get/Query responses, add WARN deprecation logs for old credential fields, Javadoc on affected response methods. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation messages to say 'upcoming release' not 'v3' Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add [DEPRECATED] prefix to deprecation log messages per industry standard Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add PM-facing document for v2 public interface changes and deprecation Non-technical overview of credential field renames and skyflow_id response key deprecation — covers customer impact, deprecation warnings, migration guide, and what is NOT changing. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add downloadURL→downloadUrl deprecation to plan and PM doc - Deprecation plan: add Task 6 for GetRequest + DetokenizeRequest with @Deprecated annotation approach (compile-time signal vs runtime log) - PM doc: add section 3 for downloadURL rename with migration example Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: remove SDK-level field value null/empty validation from Insert and Update The Skyflow API accepts additionalProperties of Any type including null and empty strings. SDK should not add validation on top of BE — pass through and let BE decide. Removed: - value == null/isEmpty check in validateInsertRequest - value == null/isEmpty check in validateUpdateRequest - value == null/isEmpty check in validateTokensMapWithTokenStrict - values.isEmpty() check in validateInsertRequest (no minItems in API spec) Kept: - values == null check (NPE guard — cannot iterate null array) - key == null/isEmpty check (null keys cannot be JSON-serialized) Deleted 6 tests that asserted on the removed behaviour. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add IDE autocomplete behavior for deprecation signals - PM doc: new section explaining how @Deprecated(forRemoval) shows in IntelliJ/VS Code autocomplete (strikethrough, orange underline, tooltip with clickable link to new method) vs runtime WARN log for map keys - Deprecation plan: update downloadURL tasks to use @Deprecated(since="2.1", forRemoval=true) + {link} for stronger IDE signal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: segregate code smells into dedicated section in code-review command Splits old Section 6 into: - Section 6: Code quality (actionable correctness checks) - Section 7: Code smells (structural signals, flagged at Smell severity) Code smell catalogue covers: long methods/classes, business logic in data classes, toString() with logic, deep nesting, magic numbers, raw HashMap chains, dead code, stale comments, temporary fields. Severity table clarified: Critical/Bug/Edge Case/Quality = fix before merge; Smell = flag and track. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: add DEPRECATED_SKYFLOW_ID_KEY log entry to InfoLogs * fix: restore skyflow_id key in Get/Query responses for v2 backward compat Both skyflow_id (deprecated) and skyflowId (new form) are now present in response maps simultaneously. WARN log emitted per record. * docs: add deprecation Javadoc for skyflow_id key in GetResponse and QueryResponse * feat: deprecate downloadURL in favour of downloadUrl in GetRequest and DetokenizeRequest Old downloadURL() methods kept as @Deprecated(forRemoval=true) delegates. Runtime WARN log emitted on old form usage. 100% test coverage: new form, deprecated form, default value for both classes. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update CLAUDE.md — add code-smell command, update slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation plan and PM doc - credentials permanently supported Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * test: add new-form downloadUrl tests alongside deprecated downloadURL tests Both old (downloadURL) and new (downloadUrl) builder methods tested: - GetTests: 2 new tests for downloadUrl() with cross-assertion on deprecated getDownloadURL() returning same value - DetokenizeTests: 1 new test same pattern - VaultClientTests: 1 new integration test for DetokenizeRequest.downloadUrl() Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: changes to claude * docs: migrate V1-to-V2 guide from README to docs/, update CHANGELOG - Extract 260-line migration section from README.md to docs/migrate_to_v2.md following the pattern established in skyflow-node PR #258 - README now links to docs/migrate_to_v2.md instead of inline content - docs/migrate_to_v2.md adds v2.1+ sections for credential field renames and skyflow_id deprecation (new content) - CHANGELOG.md: add v2.0.4 release notes covering nomenclature changes, backward compat deprecations, and validation removal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG — remove v1 entries, keep only v2.0.4 Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG to point to GitHub and Maven releases Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 banner to README with migration link and EOL notice Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: use release notes link instead of CHANGELOG in banner Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update README banner to v2.1.x announcement Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * revert: remove .claude/ and CLAUDE.md — will be raised as separate PR from main Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove superpowers planning docs from repo Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: ignore docs/superpowers/ — keep planning docs local only Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update cspell config — British English words, Maven flags, ignore paths Added words: serialise/d/s, normalise/d/s/Normalises, behaviour/s/Behaviour, sanitisation, recognised, unrecognised, prioritised Added regex: /-D[A-Za-z][A-Za-z0-9.]*/g to ignore Maven -D flags Added ignorePaths: RUNNING_SAMPLES.md, docs/superpowers/** Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove v2-public-interface-changes.md Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: replace real RSA key with fake key in BearerTokenTests Real 2048-bit RSA key replaced with fake base64 value. Assertion updated from InvalidTokenUri to InvalidKeySpec — still proves all credential fields were resolved (failure is at RSA parsing, not field lookup). Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: guard against null in DetokenizeRequest.downloadUrl(null) Calling .downloadUrl(null) previously stored null in the field, creating an NPE risk for callers who read getDownloadUrl() back without a null check. Now null -> false (matching the default), consistent with the continueOnError(null) guard in the same builder. Added test: testDetokenizeRequestDownloadUrlNullTreatedAsFalse Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove dead error constants after validation removal EmptyValues, EmptyValueInValues, EmptyValueInTokens (ErrorMessage) and EMPTY_VALUES, EMPTY_OR_NULL_VALUE_IN_VALUES, EMPTY_OR_NULL_VALUE_IN_TOKENS (ErrorLogs) are unreachable since the SDK-level null/empty field validation was removed. Deleted to prevent accidental re-wiring. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * Revert "fix: guard against null in DetokenizeRequest.downloadUrl(null)" This reverts commit a8026686fa071419e52fea438f97846845de9618. * test: add positive tests for permissive Insert validation behaviour Three new tests assert that previously-blocked inputs now pass SDK validation (SDK defers to BE per API spec additionalProperties: Any type): - Empty values array [] passes - Null field value passes - Empty string field value passes Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix banner * feat: port PR #273 changes — raw body support, URL encoding, null-safe request ID - Add EMPTY_STRING, QUOTE, HTTPS_PROTOCOL, CURLY_PLACEHOLDER and HttpUtilityExtra (RAW_BODY_KEY, SDK_GENERATED_PREFIX) to Constants - HttpUtility: conditional content-type header, __raw_body__ passthrough, UUID fallback when server omits x-request-id, URL-encoded form params - Utils: URL-encode path and query params with graceful fallback - Validations: accept String request bodies for non-JSON content types - ConnectionController: wrap String bodies in __raw_body__ for non-JSON content types; fall back to raw string when response is not JSON - InfoLogs: "Bearer token is expired" → "Bearer token is invalid or expired" - HttpUtilityTests: add raw body, no content-type, null request ID and special-character form-encoding tests Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard unsafe Optional.get() calls and remove dead statement in DetectController - Remove dead statement `response.getEntities().get(0).getFile()` that discarded its result and threw IndexOutOfBoundsException/NPE when entities was null or empty (C1); the guarded block below already handles entity processing correctly - Replace unguarded `response.getOutput().get()` in getFirstOutput() with `.orElse(null)` so an absent output Optional returns null instead of throwing NoSuchElementException (C2) - Replace unguarded `firstOutput.getProcessedFileExtension().get().toString()` with `.map(Object::toString).orElse(UNKNOWN)`, reusing the already-computed Optional and matching the safe pattern used for processedFileType one line above (C3) - Add 4 unit tests covering all three fixes via reflection Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard Optional.get() in delete, close FileReader, guard empty form-encode - VaultController.delete: replace unguarded getRecordIdResponse().get() with .orElse(Collections.emptyList()) so an absent field in the API response returns an empty list instead of throwing NoSuchElementException (M2) - BearerToken, SignedDataTokens: wrap FileReader in try/finally to guarantee close() is called even when JsonParser throws JsonSyntaxException; close IOException is intentionally swallowed since the parse already completed (M3) - HttpUtility.formatJsonToFormEncodedString: guard against empty entry set so substring(0, -1) is not called on an empty StringBuilder (M4) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: update README for v2.1 public interface changes - Replace deprecated credential key names in examples: clientID→clientId, keyID→keyId, TokenURI→tokenUri - Update Insert, Get, and Query response examples to use skyflowId (the SDK has always returned skyflowId for Insert; Get/Query now return skyflowId as the primary key) - Add deprecation note after each affected response block: skyflow_id is deprecated and will be removed in an upcoming release Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: do not generate SDK-side requestId when server omits x-request-id header When the response carries no x-request-id header, set requestID to null instead of fabricating a UUID with an SDK-Generated- prefix. A null value is the honest signal to callers that no server request ID is available. Also removes the now-unused UUID import. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: add printWarningLogOnce to suppress per-record deprecation spam Replace per-record printWarningLog calls for DEPRECATED_SKYFLOW_ID_KEY with printWarningLogOnce so the deprecation notice fires at most once per JVM session regardless of result set size. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: fix testSendRequestWithNullRequestId to assert null requestId The SDK no longer generates a fallback UUID when the server omits x-request-id; getRequestID() now returns null. Update the assertion to match the new behaviour. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: fix request_index to requestIndex in README response examples The SDK puts camelCase requestIndex in response maps; the README examples incorrectly showed snake_case request_index. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: readme * SK-2814: Java SDK v2.1 — Public interface nomenclature cleanup, backward compat deprecations, and docs (#308) * docs: add Java SDK nomenclature cleanup design spec Captures the design for public interface renames per the server-side SDK nomenclature changes spec: credential field fallbacks (clientId/keyId/tokenUri), skyflow_id→skyflowId in Get/Query responses, and QueryResponse errors/tokenizedData field additions. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: clarify tokenizedData reasoning in nomenclature cleanup spec Adds explanation for why tokenizedData change is valid despite the Query API currently not returning tokens — based on V1FieldRecords schema support and cross-SDK consistency requirement. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: expand reasoning in Java nomenclature cleanup spec Adds detailed rationale to each design section: why the naming convention matters, why the fallback strategy was chosen over a hard cut, why skyflow_id normalization is inconsistent today, the tokenizedData API schema vs docs discrepancy, and why getErrors() is missing only from QueryResponse. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: correct tokenizedData implementation rationale in spec Clarifies that ignoring record.getTokens() in getFormattedQueryRecord is intentional (Query API cannot return tokens), and that the fix is to promote the toString() hack into a real always-empty field rather than reading from the API response. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: remove tokenizedData from scope in nomenclature cleanup spec Query API cannot return tokens; the toString() inconsistency is not worth fixing since callers have no reason to access tokenizedData programmatically on query results. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add implementation plan for Java SDK nomenclature cleanup 5-task TDD plan covering credential field renames with fallback, skyflow_id normalisation in Get/Query responses, and QueryResponse getErrors() accessor. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId/tokenUri in BearerToken with fallback to old form Add fallback lookup logic so getBearerTokenFromCredentials tries new camelCase keys (clientId, keyId, tokenUri) first and falls back to the legacy all-caps forms (clientID, keyID, tokenURI) for backward compatibility during migration. Add testBearerTokenWithNewFormCredentialKeys to verify the new key form is recognized end-to-end. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId in SignedDataTokens with fallback to old form Add fallback logic to GenerateSignedTokensFromCredentials so both new-form keys (clientId/keyId) and legacy all-caps keys (clientID/keyID) are accepted during migration. Mirrors the pattern already applied to BearerToken.java. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: normalise skyflow_id to skyflowId in Get and Query response maps Insert and Update responses already used camelCase skyflowId; Get and Query were passing through the raw wire-format snake_case key. Add the rename in getFormattedGetRecord and getFormattedQueryRecord, and add reflection-based unit tests to cover both formatters. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: add getErrors() accessor to QueryResponse Adds a private final errors field (always null) and its public accessor to QueryResponse, matching the pattern in GetResponse and InsertResponse. Removes the hardcoded responseObject.add("errors", null) from toString() since serializeNulls on the declared field handles it automatically. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: audit confirms no setFooID/getFooID violations in public API * chore: add Claude Code setup (CLAUDE.md + .claude/) Adapted from skyflow-node PR #305. Includes: - CLAUDE.md with project overview, structure, naming conventions, build commands - .claude/settings.json with PostToolUse compile+checkstyle hooks, PreToolUse generated-code guard, Stop notification; paths are relative (no hardcoded user dirs) - .claude/commands/: code-review, code-security, sdk-sample, test slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: fix gaps and inaccuracies in Claude setup files - CLAUDE.md: add vault/bin/ package, all 5 controllers, pre-existing test failure baseline - settings.json: fix checkstyle hook to print violations (was silently swallowing output with capture_output=True) - sdk-sample.md: fix InsertOptions (doesn't exist), correct sample package structure, correct credential type per feature - code-review.md: fix validation location (controller not build()), fix HashMap rule (SDK pattern is raw HashMaps) - test.md: document pre-existing failures, note checkstyle failsOnError Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 backward compat + deprecation warnings implementation plan 6-task TDD plan: restore skyflow_id key alongside skyflowId in Get/Query responses, add WARN deprecation logs for old credential fields, Javadoc on affected response methods. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation messages to say 'upcoming release' not 'v3' Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add [DEPRECATED] prefix to deprecation log messages per industry standard Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add PM-facing document for v2 public interface changes and deprecation Non-technical overview of credential field renames and skyflow_id response key deprecation — covers customer impact, deprecation warnings, migration guide, and what is NOT changing. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add downloadURL→downloadUrl deprecation to plan and PM doc - Deprecation plan: add Task 6 for GetRequest + DetokenizeRequest with @Deprecated annotation approach (compile-time signal vs runtime log) - PM doc: add section 3 for downloadURL rename with migration example Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: remove SDK-level field value null/empty validation from Insert and Update The Skyflow API accepts additionalProperties of Any type including null and empty strings. SDK should not add validation on top of BE — pass through and let BE decide. Removed: - value == null/isEmpty check in validateInsertRequest - value == null/isEmpty check in validateUpdateRequest - value == null/isEmpty check in validateTokensMapWithTokenStrict - values.isEmpty() check in validateInsertRequest (no minItems in API spec) Kept: - values == null check (NPE guard — cannot iterate null array) - key == null/isEmpty check (null keys cannot be JSON-serialized) Deleted 6 tests that asserted on the removed behaviour. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add IDE autocomplete behavior for deprecation signals - PM doc: new section explaining how @Deprecated(forRemoval) shows in IntelliJ/VS Code autocomplete (strikethrough, orange underline, tooltip with clickable link to new method) vs runtime WARN log for map keys - Deprecation plan: update downloadURL tasks to use @Deprecated(since="2.1", forRemoval=true) + {link} for stronger IDE signal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: segregate code smells into dedicated section in code-review command Splits old Section 6 into: - Section 6: Code quality (actionable correctness checks) - Section 7: Code smells (structural signals, flagged at Smell severity) Code smell catalogue covers: long methods/classes, business logic in data classes, toString() with logic, deep nesting, magic numbers, raw HashMap chains, dead code, stale comments, temporary fields. Severity table clarified: Critical/Bug/Edge Case/Quality = fix before merge; Smell = flag and track. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: add DEPRECATED_SKYFLOW_ID_KEY log entry to InfoLogs * fix: restore skyflow_id key in Get/Query responses for v2 backward compat Both skyflow_id (deprecated) and skyflowId (new form) are now present in response maps simultaneously. WARN log emitted per record. * docs: add deprecation Javadoc for skyflow_id key in GetResponse and QueryResponse * feat: deprecate downloadURL in favour of downloadUrl in GetRequest and DetokenizeRequest Old downloadURL() methods kept as @Deprecated(forRemoval=true) delegates. Runtime WARN log emitted on old form usage. 100% test coverage: new form, deprecated form, default value for both classes. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update CLAUDE.md — add code-smell command, update slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation plan and PM doc - credentials permanently supported Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * test: add new-form downloadUrl tests alongside deprecated downloadURL tests Both old (downloadURL) and new (downloadUrl) builder methods tested: - GetTests: 2 new tests for downloadUrl() with cross-assertion on deprecated getDownloadURL() returning same value - DetokenizeTests: 1 new test same pattern - VaultClientTests: 1 new integration test for DetokenizeRequest.downloadUrl() Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: changes to claude * docs: migrate V1-to-V2 guide from README to docs/, update CHANGELOG - Extract 260-line migration section from README.md to docs/migrate_to_v2.md following the pattern established in skyflow-node PR #258 - README now links to docs/migrate_to_v2.md instead of inline content - docs/migrate_to_v2.md adds v2.1+ sections for credential field renames and skyflow_id deprecation (new content) - CHANGELOG.md: add v2.0.4 release notes covering nomenclature changes, backward compat deprecations, and validation removal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG — remove v1 entries, keep only v2.0.4 Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG to point to GitHub and Maven releases Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 banner to README with migration link and EOL notice Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: use release notes link instead of CHANGELOG in banner Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update README banner to v2.1.x announcement Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * revert: remove .claude/ and CLAUDE.md — will be raised as separate PR from main Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove superpowers planning docs from repo Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: ignore docs/superpowers/ — keep planning docs local only Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update cspell config — British English words, Maven flags, ignore paths Added words: serialise/d/s, normalise/d/s/Normalises, behaviour/s/Behaviour, sanitisation, recognised, unrecognised, prioritised Added regex: /-D[A-Za-z][A-Za-z0-9.]*/g to ignore Maven -D flags Added ignorePaths: RUNNING_SAMPLES.md, docs/superpowers/** Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove v2-public-interface-changes.md Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: replace real RSA key with fake key in BearerTokenTests Real 2048-bit RSA key replaced with fake base64 value. Assertion updated from InvalidTokenUri to InvalidKeySpec — still proves all credential fields were resolved (failure is at RSA parsing, not field lookup). Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: guard against null in DetokenizeRequest.downloadUrl(null) Calling .downloadUrl(null) previously stored null in the field, creating an NPE risk for callers who read getDownloadUrl() back without a null check. Now null -> false (matching the default), consistent with the continueOnError(null) guard in the same builder. Added test: testDetokenizeRequestDownloadUrlNullTreatedAsFalse Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove dead error constants after validation removal EmptyValues, EmptyValueInValues, EmptyValueInTokens (ErrorMessage) and EMPTY_VALUES, EMPTY_OR_NULL_VALUE_IN_VALUES, EMPTY_OR_NULL_VALUE_IN_TOKENS (ErrorLogs) are unreachable since the SDK-level null/empty field validation was removed. Deleted to prevent accidental re-wiring. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * Revert "fix: guard against null in DetokenizeRequest.downloadUrl(null)" This reverts commit a8026686fa071419e52fea438f97846845de9618. * test: add positive tests for permissive Insert validation behaviour Three new tests assert that previously-blocked inputs now pass SDK validation (SDK defers to BE per API spec additionalProperties: Any type): - Empty values array [] passes - Null field value passes - Empty string field value passes Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix banner * feat: port PR #273 changes — raw body support, URL encoding, null-safe request ID - Add EMPTY_STRING, QUOTE, HTTPS_PROTOCOL, CURLY_PLACEHOLDER and HttpUtilityExtra (RAW_BODY_KEY, SDK_GENERATED_PREFIX) to Constants - HttpUtility: conditional content-type header, __raw_body__ passthrough, UUID fallback when server omits x-request-id, URL-encoded form params - Utils: URL-encode path and query params with graceful fallback - Validations: accept String request bodies for non-JSON content types - ConnectionController: wrap String bodies in __raw_body__ for non-JSON content types; fall back to raw string when response is not JSON - InfoLogs: "Bearer token is expired" → "Bearer token is invalid or expired" - HttpUtilityTests: add raw body, no content-type, null request ID and special-character form-encoding tests Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard unsafe Optional.get() calls and remove dead statement in DetectController - Remove dead statement `response.getEntities().get(0).getFile()` that discarded its result and threw IndexOutOfBoundsException/NPE when entities was null or empty (C1); the guarded block below already handles entity processing correctly - Replace unguarded `response.getOutput().get()` in getFirstOutput() with `.orElse(null)` so an absent output Optional returns null instead of throwing NoSuchElementException (C2) - Replace unguarded `firstOutput.getProcessedFileExtension().get().toString()` with `.map(Object::toString).orElse(UNKNOWN)`, reusing the already-computed Optional and matching the safe pattern used for processedFileType one line above (C3) - Add 4 unit tests covering all three fixes via reflection Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard Optional.get() in delete, close FileReader, guard empty form-encode - VaultController.delete: replace unguarded getRecordIdResponse().get() with .orElse(Collections.emptyList()) so an absent field in the API response returns an empty list instead of throwing NoSuchElementException (M2) - BearerToken, SignedDataTokens: wrap FileReader in try/finally to guarantee close() is called even when JsonParser throws JsonSyntaxException; close IOException is intentionally swallowed since the parse already completed (M3) - HttpUtility.formatJsonToFormEncodedString: guard against empty entry set so substring(0, -1) is not called on an empty StringBuilder (M4) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: update README for v2.1 public interface changes - Replace deprecated credential key names in examples: clientID→clientId, keyID→keyId, TokenURI→tokenUri - Update Insert, Get, and Query response examples to use skyflowId (the SDK has always returned skyflowId for Insert; Get/Query now return skyflowId as the primary key) - Add deprecation note after each affected response block: skyflow_id is deprecated and will be removed in an upcoming release Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: do not generate SDK-side requestId when server omits x-request-id header When the response carries no x-request-id header, set requestID to null instead of fabricating a UUID with an SDK-Generated- prefix. A null value is the honest signal to callers that no server request ID is available. Also removes the now-unused UUID import. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: add printWarningLogOnce to suppress per-record deprecation spam Replace per-record printWarningLog calls for DEPRECATED_SKYFLOW_ID_KEY with printWarningLogOnce so the deprecation notice fires at most once per JVM session regardless of result set size. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: fix testSendRequestWithNullRequestId to assert null requestId The SDK no longer generates a fallback UUID when the server omits x-request-id; getRequestID() now returns null. Update the assertion to match the new behaviour. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: fix request_index to requestIndex in README response examples The SDK puts camelCase requestIndex in response maps; the README examples incorrectly showed snake_case request_index. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: readme --------- Co-authored-by: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * [AUTOMATED] Private Release 3.0.0-beta.11-dev-de8273b * feat: accept camelCase skyflowId in UpdateRequest and fire deprecation warnings per-request - UpdateRequest.data now accepts 'skyflowId' (preferred) alongside deprecated 'skyflow_id' - extractUpdateSkyflowId() prefers camelCase key, emits WARN-level deprecation on snake_case fallback - Remove printWarningLogOnce — all deprecation warnings now fire on every request via printWarningLog - Add DEPRECATED_SKYFLOW_ID_REQUEST_KEY log entry for request-side key deprecation Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: prefer skyflowId over skyflow_id in UpdateRequest with deprecation warning When both keys are present, skyflowId is used and a warning is emitted. Adds unit tests for all four cases: camelCase only, snake_case only, both keys (preference), and both keys (map cleanup). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: add getByot() as canonical form; keep getBYOT() as deprecated delegate - TokenMode.getByot() is the new preferred accessor; getBYOT() retained as @Deprecated(since="2.1",forRemoval=true) delegate for back-compat - VaultClient updated to call getByot() at all three insert/update sites - UpdateTests: 5 new tests covering camelCase skyflowId key in UpdateRequest - LogUtilLevelTests: 5 new tests verifying WARN log fires at DEBUG/INFO/WARN levels and is suppressed at ERROR (matches Skyflow default) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: add TokenModeTest covering getByot(), deprecated getBYOT() delegate, and toString() Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: emit runtime deprecation warning in getBYOT(); test that it fires - InfoLogs: add DEPRECATED_GET_BYOT constant - TokenMode.getBYOT(): call LogUtil.printWarningLog() so callers see a runtime warning (consistent with downloadURL() and other deprecated methods) - TokenModeTest: replace @SuppressWarnings-only tests with assertions that the warning fires at INFO level and is suppressed at ERROR level Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * refactor: consolidate both-keys deprecation warning into DEPRECATED_SKYFLOW_ID_REQUEST_KEY Remove DEPRECATED_SKYFLOW_ID_BOTH_KEYS; both the snake_case-only and both-keys-present paths in extractUpdateSkyflowId now emit the same DEPRECATED_SKYFLOW_ID_REQUEST_KEY message. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * SK-2814: java public interface cleanup (#314) * docs: add Java SDK nomenclature cleanup design spec Captures the design for public interface renames per the server-side SDK nomenclature changes spec: credential field fallbacks (clientId/keyId/tokenUri), skyflow_id→skyflowId in Get/Query responses, and QueryResponse errors/tokenizedData field additions. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: clarify tokenizedData reasoning in nomenclature cleanup spec Adds explanation for why tokenizedData change is valid despite the Query API currently not returning tokens — based on V1FieldRecords schema support and cross-SDK consistency requirement. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: expand reasoning in Java nomenclature cleanup spec Adds detailed rationale to each design section: why the naming convention matters, why the fallback strategy was chosen over a hard cut, why skyflow_id normalization is inconsistent today, the tokenizedData API schema vs docs discrepancy, and why getErrors() is missing only from QueryResponse. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: correct tokenizedData implementation rationale in spec Clarifies that ignoring record.getTokens() in getFormattedQueryRecord is intentional (Query API cannot return tokens), and that the fix is to promote the toString() hack into a real always-empty field rather than reading from the API response. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: remove tokenizedData from scope in nomenclature cleanup spec Query API cannot return tokens; the toString() inconsistency is not worth fixing since callers have no reason to access tokenizedData programmatically on query results. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add implementation plan for Java SDK nomenclature cleanup 5-task TDD plan covering credential field renames with fallback, skyflow_id normalisation in Get/Query responses, and QueryResponse getErrors() accessor. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId/tokenUri in BearerToken with fallback to old form Add fallback lookup logic so getBearerTokenFromCredentials tries new camelCase keys (clientId, keyId, tokenUri) first and falls back to the legacy all-caps forms (clientID, keyID, tokenURI) for backward compatibility during migration. Add testBearerTokenWithNewFormCredentialKeys to verify the new key form is recognized end-to-end. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId in SignedDataTokens with fallback to old form Add fallback logic to GenerateSignedTokensFromCredentials so both new-form keys (clientId/keyId) and legacy all-caps keys (clientID/keyID) are accepted during migration. Mirrors the pattern already applied to BearerToken.java. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: normalise skyflow_id to skyflowId in Get and Query response maps Insert and Update responses already used camelCase skyflowId; Get and Query were passing through the raw wire-format snake_case key. Add the rename in getFormattedGetRecord and getFormattedQueryRecord, and add reflection-based unit tests to cover both formatters. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: add getErrors() accessor to QueryResponse Adds a private final errors field (always null) and its public accessor to QueryResponse, matching the pattern in GetResponse and InsertResponse. Removes the hardcoded responseObject.add("errors", null) from toString() since serializeNulls on the declared field handles it automatically. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: audit confirms no setFooID/getFooID violations in public API * chore: add Claude Code setup (CLAUDE.md + .claude/) Adapted from skyflow-node PR #305. Includes: - CLAUDE.md with project overview, structure, naming conventions, build commands - .claude/settings.json with PostToolUse compile+checkstyle hooks, PreToolUse generated-code guard, Stop notification; paths are relative (no hardcoded user dirs) - .claude/commands/: code-review, code-security, sdk-sample, test slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: fix gaps and inaccuracies in Claude setup files - CLAUDE.md: add vault/bin/ package, all 5 controllers, pre-existing test failure baseline - settings.json: fix checkstyle hook to print violations (was silently swallowing output with capture_output=True) - sdk-sample.md: fix InsertOptions (doesn't exist), correct sample package structure, correct credential type per feature - code-review.md: fix validation location (controller not build()), fix HashMap rule (SDK pattern is raw HashMaps) - test.md: document pre-existing failures, note checkstyle failsOnError Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 backward compat + deprecation warnings implementation plan 6-task TDD plan: restore skyflow_id key alongside skyflowId in Get/Query responses, add WARN deprecation logs for old credential fields, Javadoc on affected response methods. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation messages to say 'upcoming release' not 'v3' Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add [DEPRECATED] prefix to deprecation log messages per industry standard Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add PM-facing document for v2 public interface changes and deprecation Non-technical overview of credential field renames and skyflow_id response key deprecation — covers customer impact, deprecation warnings, migration guide, and what is NOT changing. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add downloadURL→downloadUrl deprecation to plan and PM doc - Deprecation plan: add Task 6 for GetRequest + DetokenizeRequest with @Deprecated annotation approach (compile-time signal vs runtime log) - PM doc: add section 3 for downloadURL rename with migration example Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: remove SDK-level field value null/empty validation from Insert and Update The Skyflow API accepts additionalProperties of Any type including null and empty strings. SDK should not add validation on top of BE — pass through and let BE decide. Removed: - value == null/isEmpty check in validateInsertRequest - value == null/isEmpty check in validateUpdateRequest - value == null/isEmpty check in validateTokensMapWithTokenStrict - values.isEmpty() check in validateInsertRequest (no minItems in API spec) Kept: - values == null check (NPE guard — cannot iterate null array) - key == null/isEmpty check (null keys cannot be JSON-serialized) Deleted 6 tests that asserted on the removed behaviour. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add IDE autocomplete behavior for deprecation signals - PM doc: new section explaining how @Deprecated(forRemoval) shows in IntelliJ/VS Code autocomplete (strikethrough, orange underline, tooltip with clickable link to new method) vs runtime WARN log for map keys - Deprecation plan: update downloadURL tasks to use @Deprecated(since="2.1", forRemoval=true) + {link} for stronger IDE signal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: segregate code smells into dedicated section in code-review command Splits old Section 6 into: - Section 6: Code quality (actionable correctness checks) - Section 7: Code smells (structural signals, flagged at Smell severity) Code smell catalogue covers: long methods/classes, business logic in data classes, toString() with logic, deep nesting, magic numbers, raw HashMap chains, dead code, stale comments, temporary fields. Severity table clarified: Critical/Bug/Edge Case/Quality = fix before merge; Smell = flag and track. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: add DEPRECATED_SKYFLOW_ID_KEY log entry to InfoLogs * fix: restore skyflow_id key in Get/Query responses for v2 backward compat Both skyflow_id (deprecated) and skyflowId (new form) are now present in response maps simultaneously. WARN log emitted per record. * docs: add deprecation Javadoc for skyflow_id key in GetResponse and QueryResponse * feat: deprecate downloadURL in favour of downloadUrl in GetRequest and DetokenizeRequest Old downloadURL() methods kept as @Deprecated(forRemoval=true) delegates. Runtime WARN log emitted on old form usage. 100% test coverage: new form, deprecated form, default value for both classes. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update CLAUDE.md — add code-smell command, update slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update deprecation plan and PM doc - credentials permanently supported Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * test: add new-form downloadUrl tests alongside deprecated downloadURL tests Both old (downloadURL) and new (downloadUrl) builder methods tested: - GetTests: 2 new tests for downloadUrl() with cross-assertion on deprecated getDownloadURL() returning same value - DetokenizeTests: 1 new test same pattern - VaultClientTests: 1 new integration test for DetokenizeRequest.downloadUrl() Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: changes to claude * docs: migrate V1-to-V2 guide from README to docs/, update CHANGELOG - Extract 260-line migration section from README.md to docs/migrate_to_v2.md following the pattern established in skyflow-node PR #258 - README now links to docs/migrate_to_v2.md instead of inline content - docs/migrate_to_v2.md adds v2.1+ sections for credential field renames and skyflow_id deprecation (new content) - CHANGELOG.md: add v2.0.4 release notes covering nomenclature changes, backward compat deprecations, and validation removal Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG — remove v1 entries, keep only v2.0.4 Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: simplify CHANGELOG to point to GitHub and Maven releases Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add v2 banner to README with migration link and EOL notice Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: use release notes link instead of CHANGELOG in banner Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: update README banner to v2.1.x announcement Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * revert: remove .claude/ and CLAUDE.md — will be raised as separate PR from main Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove superpowers planning docs from repo Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: ignore docs/superpowers/ — keep planning docs local only Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: update cspell config — British English words, Maven flags, ignore paths Added words: serialise/d/s, normalise/d/s/Normalises, behaviour/s/Behaviour, sanitisation, recognised, unrecognised, prioritised Added regex: /-D[A-Za-z][A-Za-z0-9.]*/g to ignore Maven -D flags Added ignorePaths: RUNNING_SAMPLES.md, docs/superpowers/** Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove v2-public-interface-changes.md Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: replace real RSA key with fake key in BearerTokenTests Real 2048-bit RSA key replaced with fake base64 value. Assertion updated from InvalidTokenUri to InvalidKeySpec — still proves all credential fields were resolved (failure is at RSA parsing, not field lookup). Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix: guard against null in DetokenizeRequest.downloadUrl(null) Calling .downloadUrl(null) previously stored null in the field, creating an NPE risk for callers who read getDownloadUrl() back without a null check. Now null -> false (matching the default), consistent with the continueOnError(null) guard in the same builder. Added test: testDetokenizeRequestDownloadUrlNullTreatedAsFalse Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove dead error constants after validation removal EmptyValues, EmptyValueInValues, EmptyValueInTokens (ErrorMessage) and EMPTY_VALUES, EMPTY_OR_NULL_VALUE_IN_VALUES, EMPTY_OR_NULL_VALUE_IN_TOKENS (ErrorLogs) are unreachable since the SDK-level null/empty field validation was removed. Deleted to prevent accidental re-wiring. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * Revert "fix: guard against null in DetokenizeRequest.downloadUrl(null)" This reverts commit a8026686fa071419e52fea438f97846845de9618. * test: add positive tests for permissive Insert validation behaviour Three new tests assert that previously-blocked inputs now pass SDK validation (SDK defers to BE per API spec additionalProperties: Any type): - Empty values array [] passes - Null field value passes - Empty string field value passes Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix banner * feat: port PR #273 changes — raw body support, URL encoding, null-safe request ID - Add EMPTY_STRING, QUOTE, HTTPS_PROTOCOL, CURLY_PLACEHOLDER and HttpUtilityExtra (RAW_BODY_KEY, SDK_GENERATED_PREFIX) to Constants - HttpUtility: conditional content-type header, __raw_body__ passthrough, UUID fallback when server omits x-request-id, URL-encoded form params - Utils: URL-encode path and query params with graceful fallback - Validations: accept String request bodies for non-JSON content types - ConnectionController: wrap String bodies in __raw_body__ for non-JSON content types; fall back to raw string when response is not JSON - InfoLogs: "Bearer token is expired" → "Bearer token is invalid or expired" - HttpUtilityTests: add raw body, no content-type, null request ID and special-character form-encoding tests Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard unsafe Optional.get() calls and remove dead statement in DetectController - Remove dead statement `response.getEntities().get(0).getFile()` that discarded its result and threw IndexOutOfBoundsException/NPE when entities was null or empty (C1); the guarded block below already handles entity processing correctly - Replace unguarded `response.getOutput().get()` in getFirstOutput() with `.orElse(null)` so an absent output Optional returns null instead of throwing NoSuchElementException (C2) - Replace unguarded `firstOutput.getProcessedFileExtension().get().toString()` with `.map(Object::toString).orElse(UNKNOWN)`, reusing the already-computed Optional and matching the safe pattern used for processedFileType one line above (C3) - Add 4 unit tests covering all three fixes via reflection Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: guard Optional.get() in delete, close FileReader, guard empty form-encode - VaultController.delete: replace unguarded getRecordIdResponse().get() with .orElse(Collections.emptyList()) so an absent field in the API response returns an empty list instead of throwing NoSuchElementException (M2) - BearerToken, SignedDataTokens: wrap FileReader in try/finally to guarantee close() is called even when JsonParser throws JsonSyntaxException; close IOException is intentionally swallowed since the parse already completed (M3) - HttpUtility.formatJsonToFormEncodedString: guard against empty entry set so substring(0, -1) is not called on an empty StringBuilder (M4) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: update README for v2.1 public interface changes - Replace deprecated credential key names in examples: clientID→clientId, keyID→keyId, TokenURI→tokenUri - Update Insert, Get, and Query response examples to use skyflowId (the SDK has always returned skyflowId for Insert; Get/Query now return skyflowId as the primary key) - Add deprecation note after each affected response block: skyflow_id is deprecated and will be removed in an upcoming release Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: do not generate SDK-side requestId when server omits x-request-id header When the response carries no x-request-id header, set requestID to null instead of fabricating a UUID with an SDK-Generated- prefix. A null value is the honest signal to callers that no server request ID is available. Also removes the now-unused UUID import. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: add printWarningLogOnce to suppress per-record deprecation spam Replace per-record printWarningLog calls for DEPRECATED_SKYFLOW_ID_KEY with printWarningLogOnce so the deprecation notice fires at most once per JVM session regardless of result set size. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: fix testSendRequestWithNullRequestId to assert null requestId The SDK no longer generates a fallback UUID when the server omits x-request-id; getRequestID() now returns null. Update the assertion to match the new behaviour. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: fix request_index to requestIndex in README response examples The SDK puts camelCase requestIndex in response maps; the README examples incorrectly showed snake_case request_index. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: readme * docs: add v2.1.0 upgrade banner and migration guide to v1 README (#310) Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: accept camelCase skyflowId in UpdateRequest and fire deprecation warnings per-request - UpdateRequest.data now accepts 'skyflowId' (preferred) alongside deprecated 'skyflow_id' - extractUpdateSkyflowId() prefers camelCase key, emits WARN-level deprecation on snake_case fallback - Remove printWarningLogOnce — all deprecation warnings now fire on every request via printWarningLog - Add DEPRECATED_SKYFLOW_ID_REQUEST_KEY log entry for request-side key deprecation Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * Revert "docs: add v2.1.0 upgrade banner and migration guide to v1 README (#310)" (#311) This reverts commit 6727625e9bcecc3bcbb326920475239b2af33c3e. * feat: prefer skyflowId over skyflow_id in UpdateRequest with deprecation warning When both keys are present, skyflowId is used and a warning is emitted. Adds unit tests for all four cases: camelCase only, snake_case only, both keys (preference), and both keys (map cleanup). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: add getByot() as canonical form; keep getBYOT() as deprecated delegate - TokenMode.getByot() is the new preferred accessor; getBYOT() retained as @Deprecated(since="2.1",forRemoval=true) delegate for back-compat - VaultClient updated to call getByot() at all three insert/update sites - UpdateTests: 5 new tests covering camelCase skyflowId key in UpdateRequest - LogUtilLevelTests: 5 new tests verifying WARN log fires at DEBUG/INFO/WARN levels and is suppressed at ERROR (matches Skyflow default) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: add TokenModeTest covering getByot(), deprecated getBYOT() delegate, and toString() Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: emit runtime deprecation warning in getBYOT(); test that it fires - InfoLogs: add DEPRECATED_GET_BYOT constant - TokenMode.getBYOT(): call LogUtil.printWarningLog() so callers see a runtime warning (consistent with downloadURL() and other deprecated methods) - TokenModeTest: replace @SuppressWarnings-only tests with assertions that the warning fires at INFO level and is suppressed at ERROR level Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * refactor: consolidate both-keys deprecation warning into DEPRECATED_SKYFLOW_ID_REQUEST_KEY Remove DEPRECATED_SKYFLOW_ID_BOTH_KEYS; both the snake_case-only and both-keys-present paths in extractUpdateSkyflowId now emit the same DEPRECATED_SKYFLOW_ID_REQUEST_KEY message. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * [AUTOMATED] Private Release 3.0.0-beta.11-dev-e7b5167 * feat: deprecate updateLogLevel(); add setLogLevel() as canonical replacement - Skyflow.setLogLevel() is the new preferred method on the live client - updateLogLevel() retained as @Deprecated(since="2.1",forRemoval=true) delegate; emits runtime warning via DEPRECATED_UPDATE_LOG_LEVEL log - InfoLogs: add DEPRECATED_UPDATE_LOG_LEVEL constant - SkyflowException: add JavaDoc documenting validation vs API error paths, null behaviour of requestId/grpcCode, and typical catch pattern - SkyflowTests: add setLogLevel test, deprecation-warning-fires test, and warning-suppressed-at-ERROR test for updateLogLevel() Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * samples: migrate UpdateExample to skyflowId; preserve deprecated skyflow_id form - UpdateExample.java: replace skyflow_id with skyflowId in both data maps - deprecated/UpdateExample.java: retain original skyflow_id pattern with @Deprecated annotation and pointer to the current example Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: update migration guide with v2.1 changes Add sections for UpdateRequest skyflowId preference, updateLogLevel and getBYOT deprecations. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: achieve 100% instruction/branch coverage on all public interfaces Closes all reachable coverage gaps across 16 test files. Adds tests for detect response types (ReidentifyTextResponse, DeidentifyTextResponse, EntityInfo, TextIndex, TokenFormat), enum values (DetectOutputTranscriptions, DetectEntities, MaskingMethod, TokenType, DeidentifyFileStatus), InvokeConnectionResponse.getErrors(), DeidentifyFileRequest/Response constructors and builder null-handling, SkyflowException missing JSON branches, InsertRequest.continueOnError(false), and VaultConfig/ ConnectionConfig null-credentials fallback. Two residual JaCoCo false negatives remain: Optional.filter lambda in DetokenizeRecordResponse (unreachable null check) and dead ternary branches in Skyflow$SkyflowClientBuilder whose guards are always non-null after validation. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: emit DEPRECATED_SKYFLOW_ID_KEY warning in getFormattedBatchInsertRecord getFormattedGetRecord and getFormattedQueryRecord both emit the deprecation warning when the API response contains the snake_case skyflow_id key. getFormattedBatchInsertRecord was inconsistent — it silently mapped skyflow_id to skyflowId with no warning. Also future-proofs the batch insert path to handle camelCase skyflowId if the API wire format ever migrates. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: add Error Handling section to README Documents SkyflowException — its two categories (validation vs API errors), all six properties (httpCode, message, httpStatus, grpcCode, requestId, details), the recommended try/catch pattern, and a table distinguishing what is null/empty for validation errors versus API errors. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs: remove redundant null-for-validation-errors notes from property table The subsection below the table already covers null/empty behaviour for validation errors; repeating it inline on every row was noise. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * Devesh/sk 2814 java public interface cleanup (#315) * docs: add Java SDK nomenclature cleanup design spec Captures the design for public interface renames per the server-side SDK nomenclature changes spec: credential field fallbacks (clientId/keyId/tokenUri), skyflow_id→skyflowId in Get/Query responses, and QueryResponse errors/tokenizedData field additions. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: clarify tokenizedData reasoning in nomenclature cleanup spec Adds explanation for why tokenizedData change is valid despite the Query API currently not returning tokens — based on V1FieldRecords schema support and cross-SDK consistency requirement. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: expand reasoning in Java nomenclature cleanup spec Adds detailed rationale to each design section: why the naming convention matters, why the fallback strategy was chosen over a hard cut, why skyflow_id normalization is inconsistent today, the tokenizedData API schema vs docs discrepancy, and why getErrors() is missing only from QueryResponse. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: correct tokenizedData implementation rationale in spec Clarifies that ignoring record.getTokens() in getFormattedQueryRecord is intentional (Query API cannot return tokens), and that the fix is to promote the toString() hack into a real always-empty field rather than reading from the API response. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: remove tokenizedData from scope in nomenclature cleanup spec Query API cannot return tokens; the toString() inconsistency is not worth fixing since callers have no reason to access tokenizedData programmatically on query results. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add implementation plan for Java SDK nomenclature cleanup 5-task TDD plan covering credential field renames with fallback, skyflow_id normalisation in Get/Query responses, and QueryResponse getErrors() accessor. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * feat: accept clientId/keyId/tokenUri in BearerToken with fallback to old form Add fallback lookup logic so getBearerTokenFromCredentials tries new camelCase keys (clientId, keyId, tokenUri) first and falls back to the legacy all-caps forms (clientID, keyID, tokenURI) for backward compatibility during migration. Add testBearerTokenWithNewFormCredentialKeys to verify t…

Adapted from skyflow-node PR #305. Includes: - CLAUDE.md with project overview, structure, naming conventions, build commands - .claude/settings.json with PostToolUse compile+checkstyle hooks, PreToolUse generated-code guard, Stop notification; paths are relative (no hardcoded user dirs) - .claude/commands/: code-review, code-security, sdk-sample, test slash commands Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

skyflow-bharti and others added 8 commits May 12, 2026 13:56

SK-2824 add support for request id

0e42dcc

SK-2824 run v3 test on pr

3cb2d80

SK-2824 run v3 test on pr

a810343

SK-2824 request headers support

04c7085

SK-2824 add tests for request headers support

7bb2e00

SK-2824 removed unused logs

0d35743

Merge pull request #304 from skyflowapi/SK-2824-x-request-id-support

a98436e

SK-2824 Request headers support

[AUTOMATED] Private Release 2.0.4-dev-a98436e

57f6b4d

skyflow-bharti changed the title ~~V3 release/26.5.1~~ SK-2824 release/26.5.1 May 14, 2026

skyflow-bharti closed this May 14, 2026

skyflow-bharti reopened this May 14, 2026

Devesh-Skyflow approved these changes May 14, 2026

View reviewed changes

saileshwar-skyflow approved these changes May 14, 2026

View reviewed changes

skyflow-bharti merged commit 7fb2310 into beta-release/26.5.1 May 14, 2026
6 of 8 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

SK-2824 release/26.5.1#305

SK-2824 release/26.5.1#305
skyflow-bharti merged 8 commits into
beta-release/26.5.1from
v3-release/26.5.1

skyflow-bharti commented May 14, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

skyflow-bharti commented May 14, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants