PDF Stability & Infrastructure Fix
What’s New?
This patch brings a critical infrastructure fix to our PDF flashcard generation engine. We’ve completely refactored how the system handles uploaded documents to ensure high-reliability processing for large academic files.
đźš‘ Robust PDF Processing Engine
We’ve moved beyond the unstable “single-read” stream model to a high-memory-stability byte-array pipeline:
- Stream Exhaustion Fix: Resolved a bug where larger PDFs (5MB+) would fail to produce study traces because the document stream was prematurely exhausted.
- Async Safety: Improved the reliability of background generation tasks. Your flashcard generation is now immune to temporary file deletions that could happen during high-traffic periods.
- Multi-Read Support: The system now holds a stable, memory-resident copy of the PDF for the duration of the generation lifecycle, ensuring every highlight and underline is captured perfectly.
🧹 System Optimization
- Tracking Cleanup: Removed legacy tracking scripts (Hotjar/Contentsquare) to further speed up the application’s initial load time and enhance user privacy.
- Stable Metadata Extraction: Refined the default deck naming logic to better handle special characters and OS-specific filename encodings.
🚀 Why This Matters?
Version 1.5.1 is about trust. You should never have to wonder “Will this PDF work?” This update ensures that whether it’s a 1-page summary or a 50-page complex academic paper like comnet-03.pdf, Cubrain will handle it with precision.