Tests cover:
- GET /health - health check
- GET /companies - list companies
- GET /companies/{id}/stats - company statistics
- POST /reports/generate - report generation
- Helper functions: get_kg_client, get_llm_provider
21 tests total, all passing.
https://claude.ai/code/session_01PQ6QMnmh1DbR6PH9DaiPpS
Add FastAPI server for KG-based report generation
Claude
Unverified signature
0f06ab1
- Add api/server.py with endpoints:
- GET /companies - list all companies
- GET /companies/{id}/stats - get doc/chunk counts
- POST /reports/generate - generate report sections from KG
- Add fastapi and uvicorn dependencies
- Add `make api` and `make api-prod` Makefile targets
https://claude.ai/code/session_01PQ6QMnmh1DbR6PH9DaiPpS
add key debates generation to gradio kg based tab; fix failing unit tests
Implements knowledge graph-based report generation that allows generating
reports directly from documents stored in Supabase, without requiring file uploads.
Key changes:
- Reorganize KG services into services/kg/ module (client.py, retrieval.py)
- Add KGRetrievalService with 3 retrieval modes:
* chunks: Vector search for relevant chunks
* full_documents: Use complete parsed_text from documents
* hybrid: Find top docs via chunks, then use full text (default)
- Add api/kg_routes.py with run_kg_generation() and helper functions
- Add "KG Reports" tab to Gradio UI (app.py) with:
* Company dropdown selector
* Data availability stats display
* Retrieval mode selector
* Section generation buttons (TAM, Product, Signals, Team)
- Add get_all_companies() method to KGClient
- Support embedding queries and vector search via OpenAIEmbedding
- Integrate with existing LLM pipeline and tracing infrastructure
The new flow allows users to select a company from the knowledge graph,
choos…
Plan: Add multi-company document linking via mentioned_company_ids
Add company extraction strategy to multi-company linking plan
Claude
Unverified signature
e8eea15
Adds section 6 detailing how ingestion flows will detect primary vs
mentioned companies:
- Primary: folder structure, metadata, or LLM fallback
- Mentioned: fast string matching + LLM extraction during claim extraction
- Integration point with existing claim extraction to get mentions "for free"
https://claude.ai/code/session_01UikxPeNTorkkCPaA5mhJAY
Design document for extending the KG schema to support documents that
mention multiple companies (e.g., expert interview transcripts discussing
competitors). Approach: keep existing company_id as primary, add new
mentioned_company_ids JSONB array column.
https://claude.ai/code/session_01UikxPeNTorkkCPaA5mhJAY
Merge pull request #3 from altisvc/claude/organize-repo-structure-Fcq80
docs: Add comprehensive repository reorganization plan
Update plan: multi-app selector for single root app.py
Claude
Unverified signature
576271d
- Use APP env var to select which Gradio app to launch
- Supports report_gen, kg_dashboard, and future apps from one entry point
- Works with HF Spaces by setting APP in Space settings
- Each app exports a `demo` object for consistent interface
https://claude.ai/code/session_01E5LyrMhGsH5DeHbLZWHPP8
Merge pull request #4 from altisvc/claude/plan-scraping-ingestion-fdXP9
Plan: Add public scraping ingestion routes for news, YouTube, and podcasts
simplify plan: one class per platform, backend via env
Claude
Unverified signature
deee3f2
Simplified architecture:
- Each scraper is self-contained (CrunchbaseScraper, YouTubeScraper, etc.)
- Scraping backend (ScrapingDog/ScraperAPI) configured via SCRAPING_BACKEND env var
- No --provider/--backend CLI flags, just --source
- CLI: `scrape --source crunchbase --company "Anthropic"`
https://claude.ai/code/session_01RPiNB6tC5gyNrUDNz4F2in
Add HuggingFace Spaces compatibility to reorganization plan
Claude
Unverified signature
6793687
- Keep thin app.py wrapper at root for HF Spaces auto-deployment
- Actual app logic lives in apps/report_gen/app.py
- Document options for future Gradio apps (separate wrappers, separate repos, or combined tabs)
https://claude.ai/code/session_01E5LyrMhGsH5DeHbLZWHPP8
Add repository reorganization plan for multi-project structure
Claude
Unverified signature
cc298ba
Documents the plan to split the repo into:
- core/ - shared library (services, models, tracing, utils)
- apps/ - Gradio applications (report_gen, kg_dashboard, future apps)
- cli/ - CLI tools (ingest pipeline for GCP container)
Includes step-by-step phases, import cheatsheet, and verification checklist.
https://claude.ai/code/session_01E5LyrMhGsH5DeHbLZWHPP8
refactor plan: separate classes per API/platform
Claude
Unverified signature
8b147a8
Architecture now has clear separation:
- ScrapingBackend ABC: ScrapingDogBackend, ScraperAPIBackend, DirectFetchBackend
- NewsSearchProvider ABC: CrunchbaseProvider, NewsAPIProvider, GoogleNewsProvider
- YouTube: YouTubeSearchClient, YouTubeTranscriptClient
- Podcast: ListenNotesClient, SpotifyPodcastClient, WhisperTranscriber
Sources compose these clients via dependency injection for better
testability and maintainability.
https://claude.ai/code/session_01RPiNB6tC5gyNrUDNz4F2in
Simplified approach:
- No RSS/pub-sub, just periodic CLI-triggered batch runs
- Use ScrapingDog/ScraperAPI for web content
- Company-centric: scrape for companies already in KG
- Same pipeline as GDrive: Source → GCS → Supabase
https://claude.ai/code/session_01RPiNB6tC5gyNrUDNz4F2in
add plan for public scraping ingestion routes
Claude
Unverified signature
46e2946
Comprehensive plan for adding WebScraperSource, YouTubeSource, and
PodcastSource to ingest news coverage, founder interviews, and podcast
transcripts into the knowledge graph.
https://claude.ai/code/session_01RPiNB6tC5gyNrUDNz4F2in
v0 claim extraction and various code health improvements
Implements V1 ingestion system that monitors a Google Drive folder,
stores raw files to GCS, and populates Supabase KG with documents,
chunks, and embeddings.
Key components:
- Abstract ContentSource/StorageBackend interfaces for extensibility
- GoogleDriveSource: folder listing, file download, Google Docs export
- GCSStorageBackend: raw file storage with content hashing
- IngestionOrchestrator: coordinates full sync pipeline
- CLI for manual sync operations (python -m cli.ingest sync --full)
Schema changes:
- Add ingestion_source, external_id, external_path columns to documents
- Add sync_cursors table for tracking sync state
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
add an initial knowledge graph interface that's connected to supabase
Update README with HuggingFace Space tracking info, update models
Update README: GitHub as source of truth, clarify HF Space deployment, update models
Claude
Unverified signature
04799d1
- Add Deployment section explaining GitHub repo is source of truth for future migrations
- Document workflow: pull from main GitHub repo, push to HF Space
- Remove HuggingFace Space metadata header
- Update model list to match current implementation (gemini-2.5-flash-lite, gemini-3-pro-preview, gemini-flash-latest, gpt-4o, gpt-5.2-chat-latest, gpt-5.2-pro)
- Simplify API Keys section by removing HF-specific setup instructions
port hf space demo version to git for future source of truth