Spaces:
Runtime error
Runtime error
Commit History
-
Changes
1. Created utils.py Extracted the core fraud detection logic from the original app.py . Purpose: Serve as a shared utility module. Contents: Mock Data Generators ( generate_mock_application ) Fraud Tools ( get_application_summary , explain_fraud_score , etc.) SYSTEM_PROMPT Removed: Gradio UI code, create_agent (Agent Setup), and query functions. 2. Renamed app_with_confluence.py to app.py This is now the main application file. Refactoring: Updated imports to pull tools and prompts from utils.py instead of the old app.py . Command: mv app_with_confluence.py app.py 3. Deleted Original app.py The original app.py was replaced by the renamed file. Verification Confirmed file existence: app.py (Main FastAPI app) utils.py (Shared tools) Validated imports in app.py point to utils.py . Attempted to run app.py (failed due to missing fastapi in the specific shell environment, but confirmed file structure is correct). Next Steps Run pip install -r requirements.txt to ensure all dependen…6b3286a -
Update app_with_confluence.py
Verified signaturebcc0880 -
Delete confluence_integration_example_2.py
Verified signaturebe4df70 -
Delete confluence_integration_example.py
Verified signature74401e7 -
Delete vendor
Verified signatured9d2a08 -
Update README.md
Verified signature3ee010f -
Update requirements.txt
Verified signature2b210b5 -
Migrating Gradio app to Docker-based HF space
Verified signature1a6cb34 -
Use API asset URL (.assets[0].url) instead of browser_download_url - this requires authentication properly
Verified signature5ce5a2b -
Remove confluence-ingestor library
Verified signature5e55e89 -
The RUN stage with --mount=type=secret successfully authenticates (HTTP 200, token length 40 chars, detects releases array with length 2), extracts the tag, and prints "✅ Release found, continuing...",
Verified signaturea264751 -
Update Dockerfile
Verified signature56723fc -
Update Dockerfile
Verified signature78b8ab6 -
Quick test version (add debug)
Verified signaturee954f40 -
Update Dockerfile
Verified signature3ecd75e -
Update Dockerfile
Verified signature8c9db16 -
Update Dockerfile
Verified signature42fe909 -
Update Dockerfile
Verified signature8389a98 -
Update Dockerfile
Verified signature1c35116 -
Update README.md
Verified signature74f3f71 -
Update Dockerfile
Verified signaturefa06c07 -
Update Dockerfile
Verified signature0890467 -
Update Dockerfile
Verified signature99cd0d7 -
GITHUB_TOKEN as a build argument
Verified signature77d0ad5 -
Update Dockerfile
Verified signature34d36dd -
Remove private lib
Verified signature3846331 -
Update Dockerfile
Verified signature13b0a31 -
Update Dockerfile
Verified signaturef7ced4c -
Add a Dockerfile to repo
Verified signature175b1ad -
Migrate to Docker based space instead of Gradio
Verified signature7fb9f78 -
Create .space/config.yaml
Verified signature13958c3 -
Try fixing import
Verified signature68b63fe -
Obtain package from GitHub confluence-ingestor repo
Verified signature878f9ca
-
Fixed THREE critical issues:
Issue 1: Content Truncation ✅ FIXED - Problem: Only 200 chars preview returned - Fix: Modified confluence_loader to return full content Issue 2: Wrong Pages Loaded ✅ FIXED - Problem: Tool loaded random 3 pages, might miss the target page - Fix: Added page_title parameter for specific page filtering Issue 3: HTML Processor Bug ✅ FIXED (This was the root cause - Problem: Buggy _extract_structured_text method only extracted 77 chars out of 49,319 - Fix: Rewrote to use recursive processing instead of flawed soup.descendants iteration - Result: Now extracts 13,360 chars (173x improvement)a0f7fc3 -
Root Cause: Severe Content Truncation
The confluence_loader tool was only returning: - ❌ First 5 pages (out of N loaded) - ❌ 200 characters per page - ❌ Called it a preview This caused the LLM to hallucinate because it didn't have the actual information to extract.bf60203 -
Update System Prompt to improve LLM behavior
7562969 -
Tweaked code comment
5ff387f -
Add additional libraries
6411d0c -
Remove phase phrasing
ae06c7d -
No PyPi install for confluence-ingestor lib yet
0697233 -
Only one requirements file
f095b7d -
Now the page titles will be displayed regardless of whether the space is newly ingested or already cached.
f3d79f4 -
Add mock Confluence pages for the 'cquisition-and-Credit-Risk' space, designed specifically to demonstrate the Fraud Model Explainability Assistant with Confluence integration.
ecfbd81
-
Try a different approach for build wheel
39a338c -
Update to confluence_ingestor-0.3.0-py3-none-any.whl version
729efb4 -
Add a pip build wheel
ad817ec -
Production features:
- ✅ Comprehensive structured logging - ✅ Performance metrics tracking (searches, cache hits, query times, errors) - ✅ Scheduled daily re-ingestion at 2 AM - ✅ Error monitoring and recovery - ✅ Audit trail for regulatory compliance77b035a -
The fraud assistant is ready to run with Confluence integration
87c817a -
Confluence integration would transform the fraud assistant from a demo tool into a production-ready compliance and investigation platform.
bae9ca8
-
add to launch ssr_mode=False
Verified signature77bdb14