Integrate transcript timestamps into summary bullets

[Copilot]

Plan for Integrating Transcript Timestamps into Summary Bullets

• Explore repository structure and understand current implementation
• Verify existing tests pass
• Modify get_recent_transcripts to preserve timestamp data from transcripts
• Update generate_newsletter_digest to pass timestamp data to OpenAI
• Modify OpenAI prompt to instruct inclusion of timestamps in bullet points
• Format timestamps as clickable YouTube links (format: mm:ss)
• Create tests for timestamp integration
• Run tests to validate changes
• Fix linting issues
• Manually verify the output format
• Request code review
• Address review feedback
• Run CodeQL security check
• Address PR review comments

✅ Implementation Complete

All tasks completed successfully! The feature is ready for production use.

Summary of Changes

Core Functionality:

1. Modified get_recent_transcripts() to preserve timestamp data from transcripts
2. Updated generate_newsletter_digest() to format and pass timestamp data to OpenAI
3. Enhanced OpenAI prompt with explicit instructions for timestamp formatting

Code Quality & Review Fixes:
4. Updated docstring to document new transcript format (list of dicts with 'text' and 'start')
5. Added robust error handling and validation for segment structure
6. Added type safety checks for backward compatibility with string transcripts
7. Added truncation (25000 chars) for new format to match old format behavior
8. Restored bullet point range to 2-5 (from 2-10) for concise summaries

Testing & Validation:
9. All 45 unit tests passing
10. Added specific tests for timestamp integration
11. Linting passes (ruff, flake8)
12. Type checking passes (mypy)
13. CodeQL security scan: 0 alerts

Security Summary

No security vulnerabilities detected in the changes.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Integrate Transcript Timestamps into Summary Bullets</issue_title>
<issue_description>Objective:
Enhance the summary feature by integrating timestamps from the video transcripts into each summary bullet. This will allow users to jump directly to the relevant point in the video from the summary.

Details:

• Parse the transcript to extract accurate timestamps for each section or bullet.
• Update the summary generation logic to include a clickable or clearly displayed timestamp alongside every bullet point.
• Ensure the timestamps are formatted consistently (e.g., mm:ss) and are accurate.
• Consider UX/UI implications for both web and text versions of the summary.

Benefit:
This will make summaries more actionable and user-friendly.

Acceptance Criteria:

• Each summary bullet contains an associated timestamp.
• Users can use the timestamps to jump to that part in the video (if supported).
• The feature passes relevant tests and QA review.

Additional Context:

• Coordinate with existing summary and transcript extraction logic.
• May require modifications to summary data models and UI components.
  </issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes Integrate Transcript Timestamps into Summary Bullets #31

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[Copilot]

Pull request overview

This PR enhances the newsletter digest feature by integrating clickable timestamps from video transcripts into summary bullet points. Users can now click on timestamps in MM:SS format to jump directly to relevant moments in videos.

Key Changes

• Modified get_recent_transcripts to preserve structured transcript data with timestamps instead of concatenating text
• Updated generate_newsletter_digest to format timestamps in prompts and instruct the AI to generate MM:SS links
• Added backward compatibility for legacy plain text transcript format

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 7 comments.

File	Description
app.py	Core implementation: preserves transcript timestamps in get_recent_transcripts and formats them for AI processing in generate_newsletter_digest with backward compatibility fallback
tests/test_transcripts.py	Updates test assertions to expect list-of-dict transcript format instead of concatenated strings
tests/test_openai_digest.py	Adds new tests for timestamp integration and backward compatibility with legacy string format

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The function docstring for get_recent_transcripts should be updated to reflect that the transcript field now contains structured data (list of dictionaries with 'text' and 'start' keys) rather than a plain text string. This is a breaking change in the return type that should be documented.

[greenbrettmichael]

@copilot please resolve

[Copilot]

Updated in commit 8352bfb. The docstring now documents that transcript is a list of dictionaries with 'text' (str) and 'start' (float) keys.

[Copilot]

The new transcript format lacks truncation that existed for the old format. When transcript_data is a list, the formatted transcript is built without any size limit, while the old format truncated at 25000 characters. This could cause context window overflow for OpenAI API calls when processing videos with very long transcripts. Consider adding a limit on the number of segments processed or truncating the transcript_formatted string.

Suggested change

	transcript_formatted += f"[{timestamp_seconds}s] {segment['text']} "
	transcript_formatted += f"[{timestamp_seconds}s] {segment['text']} "
	# Truncate to avoid overly long prompts, matching old-format behavior
	transcript_formatted = transcript_formatted[:25000]

[Copilot]

Added truncation at 25000 characters in commit 8352bfb to match the old format behavior and prevent context overflow.

[Copilot]

The bullet point range was changed from "2 and 5" to "2 and 10" without apparent justification. This significantly increases the maximum number of bullets, which could lead to excessively long summaries. Consider whether this change aligns with the goal of "concise, high-value weekly digest" mentioned in the system prompt, or if it should be reverted to the original range.

Suggested change

	... (Provide between 2 and 10 bullet points. Use fewer for short/simple videos, and more for dense/complex technical content.)
	... (Provide between 2 and 5 bullet points. Use fewer for short/simple videos, and more for dense/complex technical content.)

[Copilot]

Reverted to 2-5 bullet points in commit 8352bfb to maintain concise summaries aligned with the "concise, high-value weekly digest" goal.