-
Notifications
You must be signed in to change notification settings - Fork 1
Common Issues
This page covers the most common early problems. Use it as a fast triage page, then move into the main repo docs for deeper debugging.
If you are running from source:
- confirm the virtual environment was created cleanly
- upgrade
pip - reinstall
requirements.txt - run the source launcher again
Useful commands:
py -3 -m venv .venv
.\.venv\Scripts\python.exe -m pip install --upgrade pip
.\.venv\Scripts\python.exe -m pip install -r requirements.txt
.\scripts\Run-App.ps1If you are running a packaged build:
- make sure you extracted the whole release folder
- do not launch only the
.exewithout its bundled files - unzip to a normal writable location
This is usually a local environment or packaged-runtime problem, not an estimate parsing problem.
Read:
Common reasons:
- the PDF is scanned and OCR quality is poor
- the packet includes guide pages or non-estimate pages
- the carrier layout is not yet well-covered
- text extraction succeeded technically but produced broken or fragmented text
Check:
- the in-app warnings
- the generated
*.canonical.json - the readable
*.esx.xml
Scanned PDFs are supported on a best-effort basis through local OCR.
What to expect:
- better than raw text extraction on image-heavy packets
- still sensitive to skew, blur, page quality, and unusual layouts
- not all scans will produce line-item-quality output
First determine whether the problem started in parsing or export.
Look at:
*.canonical.json*.esx.xml
If the canonical file already looks wrong, focus on parsing.
If the canonical file looks reasonable but the XML/package looks wrong, focus on export or mapping.
Source-mode default log:
logs\pdf_to_esx_agent.log
Packaged-build default log:
%LOCALAPPDATA%\PDF-TO-ESX-Agent\logs\pdf_to_esx_agent.log
Main repository | Deep docs in docs_repo | Issues
docs_repo/ in the main repository is the authoritative deep technical source of truth.