Render Read/Write of any Markdown file as Markdown, not Pygments

[cboos]

Closes #232.

Problem

A follow-up to the auto-memory work (#192). We already render auto-memory file bodies as rendered Markdown instead of Pygments-highlighted source — but that specialization was keyed on the memory path, not on the file being Markdown. Any other .md file (a README, a doc, a plan) read or written during a session still rendered as highlighted source.

Change

Generalize the body rendering to every Markdown file:

• New is_markdown_path() (.md/.markdown, case- and separator-insensitive) in html/utils.py. Memory paths are a subset: is_memory_path ⊂ is_markdown_path.
• Write of any .md → always rendered as Markdown. A Write carries the file's full content, so there's no partial-slice concern.
• Read of a .md → rendered as Markdown only for a full read (start_line == 1 and not truncated, via _is_full_read). A partial slice (offset/limit) can begin or end mid-code-fence and would render garbled, so partial reads keep Pygments — where a line-numbered source view is more useful for a slice anyway.
• Reuses the HTML-escaping markdown helper (render_user_markdown_collapsible): file content is untrusted regardless of being a memory file, so raw <script>/HTML renders as text, not live DOM.

Deliberately unchanged

• Memory keeps its extra specialization: the 🧠 title, relative-link resolution, and always-Markdown body (even for partial reads). General .md files get neither the 🧠 title nor link rewriting.
• The Markdown output renderer is untouched (still fenced code blocks) — embedding raw Markdown-in-Markdown risks heading bleed and fence collisions, and the issue is framed around HTML pygmentizing.
• Edit stays as-is (diff view + partial snippet).

Testing

New test/test_markdown_file_rendering.py (15 tests): is_markdown_path (.md/.markdown, case, Windows, memory-subset, negatives incl. .mdx); Read full→Markdown / partial→Pygments / truncated→Pygments / non-md→Pygments / escaping / no-link-resolution; Write md→Markdown / non-md→Pygments / escaping / no-link-resolution. The full/partial/truncated split is pinned behaviorally through the public format_read_output (no private-symbol import). test_memory_rendering.py regressions all pass — memory behavior unchanged. No snapshot churn (no existing fixture does a non-memory .md Read/Write).

just ci green (unit + TUI + browser + snapshots, ruff, pyright, ty).

🤖 Generated with Claude Code

Summary by CodeRabbit

• New Features

  • Improved Markdown rendering: complete .md/.markdown reads and writes now display as formatted Markdown with safer HTML escaping.
  • Enhanced memory file handling, including Markdown link resolution.
  • Partial or truncated Markdown reads continue using syntax-highlighted rendering to avoid corrupted Markdown structure.

• Tests

  • Added coverage for Markdown path detection (including Windows paths and case-insensitivity), read/write rendering behavior, and link rewriting rules.

[coderabbitai]

📝 Walkthrough

Walkthrough

Adds Markdown path detection and updates read/write tool formatting so .md and .markdown files render as Markdown HTML, with partial reads still using syntax-highlighted output. Adds tests for detection and rendering branches.

Changes

Markdown File Rendering

Layer / File(s)	Summary
Markdown path detection utility claude_code_log/html/utils.py, claude_code_log/html/tool_formatters.py	Defines _MARKDOWN_EXTS = (".md", ".markdown") and is_markdown_path(file_path) in utils.py; imports is_markdown_path into tool_formatters.py.
Read formatter: full-read guard and Markdown rendering claude_code_log/html/tool_formatters.py	Introduces _is_full_read(ReadOutput) -> bool (checks start_line == 1 and not is_truncated), then reworks format_read_output to render memory paths and full .md reads as Markdown HTML, falling back to render_file_content_collapsible for partial/truncated reads.
Write formatter: Markdown rendering claude_code_log/html/tool_formatters.py	Updates format_write_input so non-memory .md/.markdown write targets call render_user_markdown_collapsible and return early; non-Markdown writes continue to render_file_content_collapsible.
Tests for Markdown detection and rendering test/test_markdown_file_rendering.py	Adds tests for extension detection, case-insensitivity, Windows paths, full/partial/truncated read branches, HTML escaping, and no file:// rewriting for non-memory links.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Possibly related PRs

• daaain/claude-code-log#172: Changes how ReadOutput.start_line and is_truncated are computed from parsed tool results, which affects the new full-read rendering branch.
• daaain/claude-code-log#204: Introduced the memory-path Markdown rendering and link-resolution behavior that this PR extends to non-memory .md/.markdown files.

Poem

🐇 I hop through .md with careful cheer,
Full reads bloom Markdown, crisp and clear.
Partial slices stay safe and sound,
While writes in rabbit-mode are Markdown-bound.
Tiny links remain just where they be —
Hooray for docs that render merrily! 🌿

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly summarizes the main change: Markdown reads and writes now render as Markdown instead of Pygments.
Linked Issues check	✅ Passed	The changes match #232 by rendering Markdown writes as Markdown and full Markdown reads safely, while keeping partial reads on Pygments.
Out of Scope Changes check	✅ Passed	The added path detection, formatter updates, and tests all support the Markdown rendering scope described in the linked issue.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch dev/md-file-rendering

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[cboos]

@coderabbitai full review

[coderabbitai]

✅ Action performed

Full review finished.

[coderabbitai]

🧹 Nitpick comments (1)

test/test_markdown_file_rendering.py (1)

91-94: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

Decouple partial-read coverage from truncation in this test.

Line 93 currently makes the case both partial and truncated, so this can pass even if the start_line branch regresses. Make it non-truncated to pin the partial-read condition specifically.

Suggested diff

-        html = format_read_output(_read(MD, MD_BODY, start_line=10, total=999))
+        html = format_read_output(_read(MD, MD_BODY, start_line=2, total=3))

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@test/test_markdown_file_rendering.py` around lines 91 - 94, The partial
markdown read test currently mixes the start_line partial-read path with
truncation via total=999, so it can still pass even if the partial-read behavior
regresses. Update test_partial_md_read_keeps_pygments to use a non-truncated
read while keeping start_line > 1, and keep the assertion on
format_read_output(_read(...)) focused on the partial-read branch in MD/MD_BODY
rather than truncation.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@test/test_markdown_file_rendering.py`:
- Around line 91-94: The partial markdown read test currently mixes the
start_line partial-read path with truncation via total=999, so it can still pass
even if the partial-read behavior regresses. Update
test_partial_md_read_keeps_pygments to use a non-truncated read while keeping
start_line > 1, and keep the assertion on format_read_output(_read(...)) focused
on the partial-read branch in MD/MD_BODY rather than truncation.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: c1be2361-a271-405b-8561-c191b3bf91aa

📥 Commits

Reviewing files that changed from the base of the PR and between 8013b61 and 9b0a164.

📒 Files selected for processing (3)

• claude_code_log/html/tool_formatters.py
• claude_code_log/html/utils.py
• test/test_markdown_file_rendering.py