feat: add podcast-transcript-fetcher skill - #48
Merged
Conversation
Contributor
Greptile SummaryThis PR adds a new
Confidence Score: 5/5The changes are additive and focused on a new skill entry, documentation, and a standalone CLI. No blocking correctness or security issues were identified in the reviewed changes, and the modified registry and skill files appear consistent with the described behavior.
 T-Rex LogsWhat T-Rex did
Reviews (7): Last reviewed commit: "fix(try_tier2): save fallback audio to p..." | Re-trigger Greptile |
Contributor
|
T-Rex pricing update — T-Rex was free through June 2026. Effective July 1, 2026, T-Rex adds 2 credits on top of the standard 1-credit review (3 total). T-Rex settings |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
There was a problem hiding this comment.
Pull request overview
Adds a new OpenDirectory skill, podcast-transcript-fetcher, including a Python CLI plus supporting registry/docs, and registers it in the repo-wide skill registry + root README skill list.
Changes:
- Introduces
podcast-transcript-fetcherskill documentation and metadata (SKILL.md, README.md, package.json, .env.example). - Adds a transcript-fetching CLI (
scripts/get_transcript.py) and a podcast registry (scripts/podcasts.json). - Updates repo registries (
packages/cli/registry.json, rootREADME.md) to include the new skill and increment skill count.
Reviewed changes
Copilot reviewed 9 out of 9 changed files in this pull request and generated 11 comments.
Show a summary per file
| File | Description |
|---|---|
| skills/podcast-transcript-fetcher/SKILL.md | Skill instructions and usage docs for the new transcript fetcher. |
| skills/podcast-transcript-fetcher/scripts/podcasts.json | Podcast registry (RSS feeds + transcript source metadata). |
| skills/podcast-transcript-fetcher/scripts/get_transcript.py | Core CLI implementing RSS search/batch + Groq/local transcription + Taddy fallback. |
| skills/podcast-transcript-fetcher/references/podcasts.md | Reference documentation for source availability and access methods. |
| skills/podcast-transcript-fetcher/README.md | Skill README with install/usage/how-it-works overview. |
| skills/podcast-transcript-fetcher/package.json | Skill metadata (name/version/category/description). |
| skills/podcast-transcript-fetcher/.env.example | Example environment variables for API keys. |
| README.md | Updates global skill count and adds the new skill to the All Skills table. |
| packages/cli/registry.json | Registers the new skill in the CLI registry. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
… and search pipeline fixes
- Add ThreadPoolExecutor parallel transcription via --parallel N flag
- Add _transcribe_job() helper for independent per-episode processing
- Add _get_api_keys() to scan GROQ_API_KEY, GROQ_API_KEY_2, ... from env
- Add _pick_key() round-robin key assignment across workers
- Refactor _transcribe_groq() to accept optional api_key parameter
- Rewrite batch_transcribe() and _run_search_pipeline() with concurrent.futures
- Fix search pipeline deduplication: ep.get('url') -> ep.get('link')
- Fix search pipeline filename index collision: idx + 1 instead of len(picked)
- Fix podcast_filter resolution in search pipeline via find_podcast()
- Fix try_tier1 github_archive fallthrough (no longer early-returns None)
- Fix try_tier2 temp dir cleanup with try/finally block
- Fix find_episode() returning episodes[0] on miss -> returns None
- Update .env.example with multi-key docs (GROQ_API_KEY_2, ...)
- Add parallel transcription section to SKILL.md, README.md, podcasts.md
…sources, Tier 2 recommended - Remove broken free sources: youtubetranscript.com, Spoken.md API, YouTube transcript handler - Tier 2 (RSS+Groq Whisper) is now the recommended approach - fast, free, and reliable - Tier 1 sources are best-effort (limited to GitHub archive for Lenny's) - Tier 3 Taddy API remains as commercial/premium option - Update README, SKILL.md, and registry to reflect new tier structure
…h, batch last-N ordering, search dedup - find_podcast: use key=lambda for sort to avoid TypeError when scores are tied (Python 3 can't compare dicts as tiebreaker) - batch_transcribe: sort episodes by pub_date descending before selecting the most recent N, so order doesn't vary with RSS feed ordering - search pipeline dedup: track both GUID and link independently with separate seen sets to catch duplicates sharing the same URL
Contributor
Author
|
Addressed all 3 Greptile issues:
All verified: syntax clean, LSP clean, search/alias/lookup working correctly. |
…dir leak When download_audio() failed, the function returned None before entering the try/finally block, leaving tmp_dir on disk forever. Moved the download call inside try so finally always cleans up.
…doc references batch_transcribe: - Replace confusing indices/total-i math with enumerate(reversed(selected), start=1) per Greptile suggestion - no more off-by-one risk - Add ISO 8601 date parsing fallback in _parse_pub_date so sort works with non-RFC-2822 date formats (T-Rex tests use ISO dates) Cleanup stale references: - podcasts.json: fix description from 'podcast-transcript' to 'podcast-transcript-fetcher' - SKILL.md: remove youtubetranscript.com and Spoken.md references (sources removed) - references/podcasts.md: remove dead Spoken.md API section from Cheeky Pint
… cleanup
When no transcription backend is available (no GROQ_API_KEY and no
faster-whisper), the code printed 'Audio saved at: {tmp_path}' but
the finally block immediately deleted tmp_dir. Users could never
actually use the file. Now copies audio to output/{podcast}/audio/
before cleanup so the recovery message points to a real file.
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Summary
Adds the podcast-transcript-fetcher skill to the OpenDirectory registry with Tier 2 (RSS+Groq Whisper) as the recommended approach.
What's included
Cleanup changes
Fixes included
Testing
Category
Research - helps find and retrieve podcast transcript content for analysis.