Update documentation for boxes_flow, allow None

[jstockwin]

Pull request

Closes #395

• Allows passing boxes_flow as None to disable the advanced layout analysis

• Updates the documentation to this effect (if you want me to call it something other than "advanced layout analysis - let me know!)

• I've added a private __validate method to LAParams to check that boxes flow is either None, or between -1 and +1. Happy to remove this if you think it's excessive..

• I've NOT added any other checks about any other params to __validate. I'm happy to do this, but there don't seem to be any other specified ranges in the docs, so I'd just be checking the types, really.

How Has This Been Tested?

I checked before the change that passing None breaks, and after that it works (using extract_pages). The speed changes a lot without the extra layout analysis, so I could tell it was working. tox passes.

Checklist

• I have added tests that prove my fix is effective or that my feature
  works
• I have added docstrings to newly created methods and classes
• I have optimized the code at least one time after creating the initial
  version
• I have updated the README.md or I am verified that this
  is not necessary
• I have updated the readthedocs documentation or I
  verified that this is not necessary
• I have added a consice human-readable description of the change to
  CHANGELOG.md

[pietermarsman]

Thanks for this beautifull small improvement. It is these things that will make it a pleasure to work with pdfminer.six.

I've added a bunch of minor requests.

[jstockwin]

@pietermarsman Thanks for the review - all makes sense.
I've rebased the first commit to resolve conflicts in CONTRIBUTING.MD, and then added a new commit which addresses your comments.
Please take another look when you get a chance.

[pietermarsman]

I'll merge it tonight.

If you have time: i've 2 tiny suggestions

[jstockwin]

I've just noticed there are a few other places that the documentation needs updating - don't merge until I've fixed that :)

[jstockwin]

Okay, now this should be good 👍

[pietermarsman]

Wow: one of the tests failed because apparently tox doesn't work with python 3.4 anymore. In this PR we can solve it by specifying an older version of tox.

I'll see if we can drop support for python 3.4 in the future.

[jstockwin]

Agree, except I think this project depends on tox_travis which just installed tox>=2.

This installed the latest version of tox, which is v3.14.6. Tox deprecated Python 3.4 in v3.14.1, but the deprecation says it won't actually get removed until the next major release. (tox changelog)

I don't think specifying a specific version of tox_travis will fix this, as it's the version of tox that matters... I guess I'll try pinning tox to a specific version and hopefully that'll help? Really I think these are issues for tox_travis/tox, though.

That said, I agree dropping python 3.4 support is a good idea.

[jstockwin]

Okay, that worked - tests are now green ✔️