feat!: Add AI online evaluations (Judge, Evaluator, IRunner) #301
+1,945
−40
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -2,6 +2,7 @@ | |
| using System.Collections.Generic; | ||
| using System.Collections.Immutable; | ||
| using System.Linq; | ||
| using LaunchDarkly.Sdk.Server.Ai.Evals; | ||
| using LaunchDarkly.Sdk.Server.Ai.Interfaces; | ||
| using Mustache; | ||
|
|
||
|
|
@@ -21,11 +22,14 @@ internal sealed class ConfigFactory | |
|
|
||
| private readonly ILaunchDarklyClient _client; | ||
| private readonly ILogger _logger; | ||
| private readonly Func<LdAiJudgeConfig, IRunner> _runnerFactory; | ||
|
|
||
| public ConfigFactory(ILaunchDarklyClient client, ILogger logger, | ||
| Func<LdAiJudgeConfig, IRunner> runnerFactory = null) | ||
| { | ||
| _client = client; | ||
| _logger = logger; | ||
| _runnerFactory = runnerFactory; | ||
| } | ||
|
|
||
| public LdAiCompletionConfig BuildCompletionConfig( | ||
|
|
@@ -44,7 +48,7 @@ public LdAiCompletionConfig BuildCompletionConfig( | |
| _logger.Error( | ||
| "AI Config '{0}': variation result is not an object (got {1}); using caller's default.", | ||
| key, ldValue.Type); | ||
| return BuildCompletionFromDefault(key, defaultValue, mergedVars, trackerFactory, context, variables, interpolate); | ||
| } | ||
|
|
||
| var (enabled, variationKey, version, mode) = ParseMeta(ldValue); | ||
|
|
@@ -54,7 +58,7 @@ public LdAiCompletionConfig BuildCompletionConfig( | |
| _logger.Warn( | ||
| "AI Config mode mismatch for {0}: expected {1}, got {2}. Returning caller's default.", | ||
| key, LdAiCompletionConfig.Mode, mode); | ||
| return BuildCompletionFromDefault(key, defaultValue, mergedVars, trackerFactory, context, variables, interpolate); | ||
| } | ||
|
|
||
| var model = ParseModel(ldValue.Get("model")); | ||
|
|
@@ -64,6 +68,7 @@ public LdAiCompletionConfig BuildCompletionConfig( | |
| : ParseMessages(ldValue.Get("messages")); | ||
| var tools = ParseTools(ldValue.Get("tools")); | ||
| var judgeConfiguration = ParseJudgeConfiguration(ldValue.Get("judgeConfiguration")); | ||
| var evaluator = BuildEvaluator(judgeConfiguration, context, variables); | ||
|
|
||
| return new LdAiCompletionConfig( | ||
| key, | ||
|
|
@@ -75,21 +80,25 @@ public LdAiCompletionConfig BuildCompletionConfig( | |
| judgeConfiguration, | ||
| model, | ||
| provider, | ||
| trackerFactory, | ||
| evaluator); | ||
| } | ||
|
|
||
| private LdAiCompletionConfig BuildCompletionFromDefault( | ||
| string key, | ||
| LdAiCompletionConfigDefault defaultValue, | ||
| IReadOnlyDictionary<string, object> mergedVars, | ||
| Func<LdAiConfig, ILdAiConfigTracker> trackerFactory, | ||
| Context context, | ||
| IReadOnlyDictionary<string, object> variables = null, | ||
| bool interpolate = true) | ||
| { | ||
| // Caller-supplied default messages can contain Mustache templates too; interpolate | ||
| // with the same per-message fallback as server-returned configs. | ||
| var messages = interpolate | ||
| ? InterpolateMessages(defaultValue.Messages, mergedVars, key) | ||
| : (defaultValue.Messages ?? new List<LdAiConfigTypes.Message>()); | ||
| var evaluator = BuildEvaluator(defaultValue.JudgeConfiguration, context, variables); | ||
| return new LdAiCompletionConfig( | ||
| key, | ||
| defaultValue.Enabled ?? true, | ||
|
|
@@ -100,7 +109,8 @@ private LdAiCompletionConfig BuildCompletionFromDefault( | |
| defaultValue.JudgeConfiguration, | ||
| defaultValue.Model, | ||
| defaultValue.Provider, | ||
| trackerFactory, | ||
| evaluator); | ||
| } | ||
|
|
||
| public LdAiAgentConfig BuildAgentConfig( | ||
|
|
@@ -120,7 +130,7 @@ public LdAiAgentConfig BuildAgentConfig( | |
| _logger.Error( | ||
| "AI Config '{0}': variation result is not an object (got {1}); using caller's default.", | ||
| key, ldValue.Type); | ||
| return BuildAgentFromDefault(key, defaultValue, mergedVars, trackerFactory, context, variables, interpolate, graphKey); | ||
| } | ||
|
|
||
| var (enabled, variationKey, version, mode) = ParseMeta(ldValue); | ||
|
|
@@ -130,7 +140,7 @@ public LdAiAgentConfig BuildAgentConfig( | |
| _logger.Warn( | ||
| "AI Config mode mismatch for {0}: expected {1}, got {2}. Returning caller's default.", | ||
| key, LdAiAgentConfig.Mode, mode); | ||
| return BuildAgentFromDefault(key, defaultValue, mergedVars, trackerFactory, context, variables, interpolate, graphKey); | ||
| } | ||
|
|
||
| var model = ParseModel(ldValue.Get("model")); | ||
|
|
@@ -140,6 +150,7 @@ public LdAiAgentConfig BuildAgentConfig( | |
| ? InterpolateInstructions(ParseInstructions(ldValue.Get("instructions")), mergedVars, key) | ||
| : ParseInstructions(ldValue.Get("instructions")); | ||
| var judgeConfiguration = ParseJudgeConfiguration(ldValue.Get("judgeConfiguration")); | ||
| var evaluator = BuildEvaluator(judgeConfiguration, context, variables, graphKey); | ||
|
|
||
| return new LdAiAgentConfig( | ||
| key, | ||
|
|
@@ -151,19 +162,24 @@ public LdAiAgentConfig BuildAgentConfig( | |
| model, | ||
| provider, | ||
| judgeConfiguration, | ||
| trackerFactory, | ||
| evaluator); | ||
| } | ||
|
|
||
| internal LdAiAgentConfig BuildAgentFromDefault( | ||
| string key, | ||
| LdAiAgentConfigDefault defaultValue, | ||
| IReadOnlyDictionary<string, object> mergedVars, | ||
| Func<LdAiConfig, ILdAiConfigTracker> trackerFactory, | ||
| Context context, | ||
| IReadOnlyDictionary<string, object> variables = null, | ||
| bool interpolate = true, | ||
| string graphKey = null) | ||
| { | ||
| var instructions = interpolate | ||
| ? InterpolateInstructions(defaultValue.Instructions, mergedVars, key) | ||
| : defaultValue.Instructions; | ||
| var evaluator = BuildEvaluator(defaultValue.JudgeConfiguration, context, variables, graphKey); | ||
| return new LdAiAgentConfig( | ||
| key, | ||
| defaultValue.Enabled ?? true, | ||
|
|
@@ -174,7 +190,8 @@ internal LdAiAgentConfig BuildAgentFromDefault( | |
| defaultValue.Model, | ||
| defaultValue.Provider, | ||
| defaultValue.JudgeConfiguration, | ||
| trackerFactory, | ||
| evaluator); | ||
| } | ||
|
|
||
| public LdAiJudgeConfig BuildJudgeConfig( | ||
|
|
@@ -183,10 +200,11 @@ public LdAiJudgeConfig BuildJudgeConfig( | |
| Context context, | ||
| LdAiJudgeConfigDefault defaultValue, | ||
| IReadOnlyDictionary<string, object> variables, | ||
| bool interpolate = true, | ||
| string graphKey = null) | ||
| { | ||
| var mergedVars = interpolate ? MergeVariables(variables, context) : null; | ||
| var trackerFactory = TrackerFactoryFor(context, graphKey); | ||
|
|
||
| if (ldValue.Type != LdValueType.Object) | ||
| { | ||
|
|
@@ -247,6 +265,49 @@ private LdAiJudgeConfig BuildJudgeFromDefault( | |
| trackerFactory); | ||
| } | ||
|
|
||
| private Evaluator BuildEvaluator(LdAiConfigTypes.JudgeConfiguration judgeConfiguration, Context context, | ||
| IReadOnlyDictionary<string, object> variables = null, string graphKey = null) | ||
| { | ||
| if (_runnerFactory == null || judgeConfiguration == null || judgeConfiguration.Judges.Count == 0) | ||
| { | ||
| return Evaluator.Noop(); | ||
| } | ||
|
|
||
| var judges = new Dictionary<string, Judge>(); | ||
| foreach (var judgeEntry in judgeConfiguration.Judges) | ||
| { | ||
| try | ||
| { | ||
| var defaultValue = LdAiJudgeConfigDefault.Disabled; | ||
| var ldValue = _client.JsonVariation(judgeEntry.Key, context, defaultValue.ToLdValue()); | ||
| var judgeConfig = BuildJudgeConfig(judgeEntry.Key, ldValue, context, defaultValue, variables, graphKey: graphKey); | ||
|
There was a problem hiding this comment. Judge sampling ignores interpolate flagMedium Severity
Additional Locations (2)  Reviewed by Cursor Bugbot for commit 6f7d88d. Configure here.
Contributor
Author
There was a problem hiding this comment. Not a bug. Attached judges are operational -- they execute against a model provider via IRunner.RunAsync and require interpolated prompts to function. A judge receiving raw {{variable}} tokens would produce nonsense. When fetching a judge directly with
Contributor
Author
There was a problem hiding this comment. Messages when fetching toxicity judge as template: [
{
"content": "You are a business information accuracy and safety expert. Evaluate....",
"role": "system"
},
{
"content": "MESSAGE HISTORY:\n{{message_history}}\n",
"role": "assistant"
},
{
"content": "RESPONSE TO EVALUATE:\n{{response_to_evaluate}}\n",
"role": "user"
}
] |
||
|
|
||
| if (!judgeConfig.Enabled) | ||
| { | ||
| _logger?.Debug("Judge '{0}' is disabled; skipping.", judgeEntry.Key); | ||
| continue; | ||
|
cursor[bot] marked this conversation as resolved.
|
||
| } | ||
|
|
||
| var runner = _runnerFactory(judgeConfig); | ||
| if (runner == null) | ||
| { | ||
| _logger?.Warn("Runner factory returned null for judge '{0}'; skipping.", judgeEntry.Key); | ||
| continue; | ||
| } | ||
|
|
||
| judges[judgeEntry.Key] = new Judge(judgeConfig, runner, _logger); | ||
| } | ||
| catch (Exception ex) | ||
| { | ||
| _logger?.Warn("Failed to initialize judge '{0}': {1}", judgeEntry.Key, ex.Message); | ||
| } | ||
| } | ||
|
|
||
| var filteredConfig = new LdAiConfigTypes.JudgeConfiguration( | ||
| judgeConfiguration.Judges.Where(j => judges.ContainsKey(j.Key)).ToList()); | ||
| return new Evaluator(judges, filteredConfig, _logger); | ||
| } | ||
|
|
||
| private string InterpolateInstructions( | ||
| string instructions, | ||
| IReadOnlyDictionary<string, object> mergedVars, | ||
|
|
@@ -401,9 +462,13 @@ internal static LdAiConfigTypes.JudgeConfiguration ParseJudgeConfiguration(LdVal | |
| { | ||
| var j = judgesArray.Get(i); | ||
| if (j.Type != LdValueType.Object) continue; | ||
| var samplingRateValue = j.Get("samplingRate"); | ||
| double? samplingRate = samplingRateValue.Type == LdValueType.Number | ||
| ? samplingRateValue.AsDouble | ||
| : (double?)null; | ||
| entries.Add(new LdAiConfigTypes.JudgeConfiguration.Judge( | ||
| j.Get("key").AsString ?? "", | ||
| samplingRate)); | ||
| } | ||
| return new LdAiConfigTypes.JudgeConfiguration(entries); | ||
| } | ||
|
|
||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Template fetch builds live evaluators
Medium Severity
CompletionConfigTemplateandAgentConfigTemplatecall the factory withinterpolate: false, butBuildCompletionConfigandBuildAgentConfigalways invokeBuildEvaluatorregardless of that flag. With a non-nullrunnerFactory, template retrieval still evaluates judge flags, interpolates judge prompts, and attaches a runnableEvaluatorwhile completion/agent content stays un-interpolated.Additional Locations (2)
pkgs/sdk/server-ai/src/Config/ConfigFactory.cs#L152-L153pkgs/sdk/server-ai/src/Config/ConfigFactory.cs#L100-L101Reviewed by Cursor Bugbot for commit 636858c. Configure here.