Wheels on Pypi

[ricardomga]

Please consider uploading the wheels to PyPI! It would be realy helpful.

[ricardomga]

@simonflueckiger could you please help and provide wheels for linux too? Thanks.

[simonflueckiger]

@ricardomga just letting you know that I'm currently looking into this

[ricardomga]

@simonflueckiger thank you a lot!

[ricardomga]

@simonflueckiger I really need this, if there is something that I can do to speed up the process let me know.

[simonflueckiger]

@ricardomga let me know what setup you have (OS / 32bit or 64bit / python version) and I can give you a wheel for testing :)

[ricardomga]

@simonflueckiger 64bit, gnu/linux (I think the distro is not important), python 3.6

[simonflueckiger]

@ricardomga thanks, forgot to ask: are you looking for tesseract 4 or 3.05.01/02?

[ricardomga]

@simonflueckiger 4 please

[simonflueckiger]

@ricardomga can you do some testing with this:

tesserocr-2.3.0-cp36-cp36m-manylinux1_x86_64.zip

it bundles tesseract 4.0.0-beta.2, leptonica-1.76.0, libgif 5.1.4, libjpeg 9b, libpng 1.6.34, libtiff 4.0.9, zlib 1.2.11

[simonflueckiger]

I uploaded all the binaries for linux 32bit/64bit and Python 2.7/3.4/3.5/3.6/3.7 in case anyone else would like to test them:

https://github.com/simonflueckiger/tesserocr-windows_build/releases/tag/tesserocr-v2.3.0-tesseract-4.0.0-beta.2

[ricardomga]

@simonflueckiger awsome, thank you a lot!

@sirfz It would be realy nice if you upload this to Pypi, the installation would as simple as pip install tesserocr. This would really easy the deploy process in the cloud providers and to through docker containers, please consider this. @simonflueckiger what do you think about uploading this to Pypi?

[simonflueckiger]

@ricardomga did the wheels work for you?

@sirfz @ricardomga I think uploading them to PyPI is a splendid idea! Since all the binary dependencies were built on the manylinux docker images and are packaged with the wheels, they should be working out of the box on pretty much any distro released after 2007.

[ricardomga]

@simonflueckiger I am getting a segmentation fault in docker image python:3.6.5 Any sugestion?

[simonflueckiger]

@ricardomga What code are you trying to run, do you know which specific line results in the segfault? What's the exact error you are getting? Which docker image are you using? Also, what do you get for

import tesserocr

print(tesserocr.tesseract_version())
print(tesserocr.get_languages())

[ricardomga]

@simonflueckiger docker image: python:3.6.5

import tesserocr
pritnt(tesserocr.tesseract_version())

'tesseract 4.0.0-beta.1\n leptonica-1.76.0\n  libgif 5.1.4 : libjpeg 9b : libpng 1.6.34 : libtiff 4.0.9 : zlib 1.2.11 : libwebp 1.0.0'

print(tesserocr.get_languages())

this statment stays hanged

image = Image.open('sample.jpg')
print(tesserocr.image_to_text(image))

This gives the error Segmentation fault (core dumped)

[ricardomga]

@simonflueckiger were you able to replicate the problem?

[simonflueckiger]

@ricardomga yes, I am getting the exact same behavior just as you described it. Looking into it ...

[ricardomga]

@simonflueckiger thank you

[ricardomga]

@simonflueckiger sorry to bother, any news?

[ricardomga]

@simonflueckiger can you please try with tesseract 4.0.0 beta4?

[ricardomga]

@simonflueckiger can you please try with tesseract 4.0.0 RC1?

[simonflueckiger]

I discontinued the Linux builds a while ago (only Windows is supported now). @sirfz can we close this issue? If there still is demand, I could create a PyPI repo and upload the Windows wheels there.

[sirfz]

I'd be happy to include build wheels if someone wants to contribute a CI pipeline that produces relevant wheels. (added "help wanted" tag)

[betaboon]

fyi i tried my hand at this in #318

[betaboon]

@sirfz as the wheels provided by #318 are not yet available on pypi, i would argue this issue should be re-opened.

[winstxnhdw]

Hey @betaboon, I assume the wheels are released on pypi by now. Doesn't that make the README docs outdated since we no longer have to install Tesseract on our system?