How ravennova1/channelscraper works
ChannelScraper turns a YouTube channel, playlist, or video into timestamped transcripts and a standalone offline search page. The main workflow can start in the browser dashboard or run as a pipeline: discover videos, fetch captions, align words and timestamps, compile transcript files, then build the search page. Separate capabilities update existing workspaces, resume interrupted processing, and detect duplicate or overlapping videos. The README documents the pipeline stages; the sampled source confirms key orchestration and compilation links, but does not expose every stage implementation or dashboard route.
Main entry points: pipeline.py Pipeline Orchestrator · build_search.py Search Compiler · runs Transcript Files · continue_run.py Run Recovery
Components 17
Dashboard
- Dashboard UI
core/ui/index.html - Dashboard Server
core/app.py
Transcription Pipeline
- Pipeline Orchestrator
core/pipeline.py - Video Discovery
core/pipeline.py - Caption Fetching
core/pipeline.py - Word Timing
core/pipeline.py - Transcript Compilation
core/pipeline.py - Transcript Files
runs
Offline Search
- Search Compiler
core/build_search.py - Offline Search Page
core/build_search.py
Workspace Operations
- Incremental Updater
core/update_channel.py - Run Recovery
core/continue_run.py - Duplicate Detection
core/find_duplicates.py - Duplicate Report
core/find_duplicates.py
Other components
- User
- YouTube
- Browser
Connections 24
- User → Dashboard UI: uses
README.md - Dashboard UI → Dashboard Server: communicates with
- Dashboard Server → Pipeline Orchestrator: starts pipeline
core/app.py - Pipeline Orchestrator → Video Discovery: runs stage
README.md - Video Discovery → YouTube: lists videos
README.md - Pipeline Orchestrator → Caption Fetching: runs stage
README.md - Caption Fetching → YouTube: fetches captions
README.md - Pipeline Orchestrator → Word Timing: runs stage
README.md - Pipeline Orchestrator → Transcript Compilation: runs stage
README.md - Pipeline Orchestrator → Transcript Files: writes
README.md - Pipeline Orchestrator → Search Compiler: calls
core/pipeline.py - Transcript Compilation → Transcript Files: produces
README.md - Search Compiler → Transcript Files: reads
README.md - Search Compiler → Offline Search Page: generates
core/build_search.py - User → Incremental Updater: requests update
README.md - Incremental Updater → Pipeline Orchestrator: uses workspace engine
core/update_channel.py - User → Run Recovery: resumes run
README.md - Run Recovery → Pipeline Orchestrator: uses workspace engine
core/continue_run.py - Run Recovery → Search Compiler: builds search page
core/continue_run.py - User → Duplicate Detection: runs analysis
README.md - Duplicate Detection → Transcript Files: reads transcript data
core/find_duplicates.py - Duplicate Detection → Duplicate Report: exports report
core/find_duplicates.py - User → Browser: opens page
README.md - Browser → Offline Search Page: searches locally
README.md