无法在 heroku 中配置 pytesseract

Question

I try to deploy pytesseract app in heroku after doing much researchs online.

I added TESSDATA_PREFIX=./.apt/usr/share/tesseract-ocr/4.00/tessdata in Heroku Config vars

I have https://github.com/heroku/heroku-buildpack-apt in my heroku buildpack.

I have Aptfile containing:

tesseract-ocr
tesseract-ocr-eng

I have

pytesseract.tesseract_cmd = '/app/.apt/usr/bin/tesseract'

in my code.

I am deploying flask API to heroku, so my Procfile is: web: gunicorn app:app

The error from heroku logs:

2022-11-16T04:22:39.262113+00:00 app[web.1]: text = pytesseract.image_to_string(img, config="--psm 6")
2022-11-16T04:22:39.262115+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.10/site-packages/pytesseract/pytesseract.py", line 423, in image_to_string  
2022-11-16T04:22:39.262116+00:00 app[web.1]: return {
2022-11-16T04:22:39.262117+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.10/site-packages/pytesseract/pytesseract.py", line 426, in <lambda>
2022-11-16T04:22:39.262117+00:00 app[web.1]: Output.STRING: lambda: run_and_get_output(*args),
2022-11-16T04:22:39.262117+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.10/site-packages/pytesseract/pytesseract.py", line 288, in run_and_get_output
2022-11-16T04:22:39.262118+00:00 app[web.1]: run_tesseract(**kwargs)
2022-11-16T04:22:39.262118+00:00 app[web.1]: File "/app/.heroku/python/lib/python3.10/site-packages/pytesseract/pytesseract.py", line 264, in run_tesseract    
2022-11-16T04:22:39.262119+00:00 app[web.1]: raise TesseractError(proc.returncode, get_errors(error_string))
2022-11-16T04:22:39.262121+00:00 app[web.1]: pytesseract.pytesseract.TesseractError: (127, '/app/.apt/usr/bin/tesseract: error while loading shared libraries: libarchive.so.13: cannot open shared object file: No such file or directory')

Anything I missed or how should I solve this?

Answer 1

The error message indicates a missing library:

error while loading shared libraries: libarchive.so.13: cannot open shared object file: No such file or directory

The apt buildpack doesn't do dependency resolution, so you may have to explicitly include transitive dependencies.

You can search https://packages.ubuntu.com to see which packages contain missing files. Make sure to match the Ubuntu LTS major version to the Heroku stack you are using, eg for Heroku 22 you'll want to look at packages for Ubuntu 22.04 LTS (Jammy).

In this case, the libarchive13 package contains libarchive.so.13 . Add that to your Aptfile , commit, and redeploy. If you find other missing dependencies, repeat the process.