Fix ordering of textlines within a textbox when boxes_flow is disabled

[jstockwin]

Pull request

Closes #411

Note: It does work to add the analyze call inside the get_key function, but that felt a bit strange since it has nothing to do with getting the key, therefore I decided to do it outside of this.

How Has This Been Tested?

Example PDF

Example code:

from pdfminer.high_level import extract_pages
from pdfminer.layout import LAParams

la_params = LAParams(boxes_flow=None)

for page in extract_pages("example.pdf", laparams=la_params, caching=False):
    print("*****OUTPUT:*****")
    for element in page:
        print(element)

Before fix:

*****OUTPUT:*****
<LTTextBoxHorizontal(-1) 58.100,343.402,106.178,406.702 'Text 3Text 2Text 1'>

After fix:

*****OUTPUT:*****
<LTTextBoxHorizontal(-1) 58.100,343.402,106.178,406.702 'Text 1\nText 2\nText 3\n'>

Checklist

• I have added tests that prove my fix is effective or that my feature
  works -> Actually the tests I added for without boxes flow kind of show the issue, you can see the output changes (less \n). The output is now consistent.
• 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

[jstockwin]

Not sure why the tests still say pending... Travis is green: https://travis-ci.org/github/pdfminer/pdfminer.six/builds/671715428

I actually had this issue on a different project the other day...

[pietermarsman]

Looks good!

Maybe you can add your test pdf (with "Test 1 Test 2 Test3") as a test case.

[jstockwin]

@pietermarsman Cool, I've added my test PDF, and added some tests for when boxes flow is disabled. I also threw in some bonus tests for the line margin parameter, since I know the line margin is 0.2 in this case.

I'll try and get round to adding some more tests, but probably not part of this PR.

[jstockwin]

Travis didn't build for my new commit... not sure how to fix that

[jstockwin]

@pietermarsman I have just rebased to resolve conflicts in the changelog, and Travis has now done its job. Any chance this can get merged?

[pietermarsman]

Thanks @jstockwin!

(Sorry for the delay)