ReceiptManager / receipt-parser-legacy

A supermarket receipt parser written in Python using tesseract OCR
https://tech.trivago.com/2015/10/06/python_receipt_parser/
Apache License 2.0
806 stars 198 forks source link

Docker image not working in Ubuntu 22.04 #158

Closed jmferreirab closed 2 years ago

jmferreirab commented 2 years ago

Issue: Docker image not working in Ubuntu 22.04

Expected behaviour: Running "make docker-run" completes without error and processes the sample images. Actual behaviour: Running make docker-run spits out error "make": executable file not found in $PATH: unknown.

Environment: Ubuntu 22.04 Python: 3.7.9 (from docker official images) Docker version 20.10.16, build aa7e414

Additional info: make docker-build completed without errors after ensuring dockerfile and pyproject.toml pointed to python version 3.7.9

Attaching screenshot of exact error:

image

monolidth commented 2 years ago

will check it out, thanks for reporting :)

monolidth commented 2 years ago

Well, I test it now and it works in my machine Macbook Pro 2021 M1 Chip, however: for me (and I might be wrong) it seems like a driver issue, see:

One user suggested to change virtualization method: Solved the issue after virtualization changes. Virtualization: kvm

Hope this help you, if not hit me up :)

monolidth commented 2 years ago

Small change: https://github.com/ReceiptManager/receipt-parser-legacy/commit/649c3feb37fd77b212a971471086165c6beca3c3 . I bump the python version :) Just use git pull.

jmferreirab commented 2 years ago

Thank you for the follow up @mateoradman and @monolidth . It's much appreciated.

Mateo's response makes a lot of sense. I was trying to run this in a Ubuntu from VirtualBox (hosted on Windows), but if it needs direct access to nvidia drivers, I won't be able to continue on Ubuntu as all it gets as a VM is an emulated driver according to nvidia-smi output. As such I will be closing this :)

A bit off topic, but if all you want to do is read receipts, the tesseract base cli may help. I found success with tesseract cli at version 4.0 out of the box on english receipts. (tesseract image_path text_result.txt -l eng --psm 6).