installation.rst: remove need to install tesseract

[deeplow]

As stated in the document already, Tesseract OCR is already bundled into the application.

[github-actions]

CLA Assistant Lite bot All contributors have signed the CLA ✍️ ✅

[deeplow]

I have read the CLA Document and I hereby sign the CLA

[deeplow]

On the Dangerzone project our original read of the documentation made us assume that we'd have to install Tesseract OCR on the host, which can be quite challenging to do in a cross-platform way.

However, later we found that this was not the case and it made an incredible difference (thanks PyMuPDF team!). So hopefully this commit helps future users of the library realize that they don't have to install Tesseract.

[jamie-lemon]

Thanks for pointing this stuff out.

I think for this one it is better if bullet point 1, remains and it says:

1. Ensure that Tesseract is installed

Because we can't assume it is installed (even though it probably is vi the MuPDF installation), possibly someone could have removed it - who knows?

Also further up the section it says in the text: "PyMuPDF will already contain all the logic to support OCR functions. But it additionally does need Tesseract’s language support data, so installation of Tesseract-OCR is still required."

Perhaps that needs to say: "PyMuPDF will already contain all the logic to support OCR functions. But it additionally does need Tesseract’s language support data, which should already be installed on your system."

[deeplow]

Thanks for pointing this stuff out.

You're welcome.

I think for this one it is better if bullet point 1, remains and it says:

1. Ensure that Tesseract is installed

Because we can't assume it is installed (even though it probably is vi the MuPDF installation), possibly someone could have removed it - who knows?

I don't know about this. I'd be inclined to recommend against, because if I were reading that I'd assume that I'd have to install it. If it stays like that then there's little point in this PR. However you prefer, but I'd be inclined to remove it.

[jamie-lemon]

Fair enough, we can top the 1st bullet point. If you update this PR to do this part further up:
"PyMuPDF will already contain all the logic to support OCR functions. But it additionally does need Tesseract’s language support data, so installation of Tesseract-OCR is still required."
->
"PyMuPDF will already contain all the logic to support OCR functions. But it additionally does need Tesseract’s language support data, which should already be installed on your system."

Then we are good. :)

[deeplow]

Great point! I can do something like that but I think we need to tweak that phrase a bit:

PyMuPDF will already contain all the logic to support OCR functions. But it additionally does need Tesseract’s language support data, which should already be installed on your system.

This technically is not the case , I think. The tesseract data is the sole thing that's missing. How about something like this:

PyMuPDF will already contain all the logic to support OCR functions. But it additionally does need Tesseract’s language support data, which you can download here.

How does this sound?

[jamie-lemon]

Sure - we try to avoid links which are not descriptive - so here is out !

Let's say:

PyMuPDF will already contain all the logic to support OCR functions. But it additionally does need Tesseract’s language support data.

[deeplow]

Done 👌

[jamie-lemon]

Nearly - because this is RST markup and not straight MD links are a bit different - you need to change it to:

I have to do a screen shot here to prevent this comment form parsing the backticks as code :)

See: https://sublime-and-sphinx-guide.readthedocs.io/en/latest/references.html#links-to-external-web-pages

[deeplow]

Whoops. Sorry about that. Fixed.

[jamie-lemon]

LGTM