feat: allow parameterize laparams

[frascuchon]

This PR resolves issue #168

[jsvine]

Hi @frascuchon, and many thanks for your proposal! I think it’s a good idea to allow for laparams customization. I don’t, however, see where in this PR that’s implemented. Am I missing something?

Part of my difficulty in reading this PR is that there seem to be many code-formatting changes introduced at the same time. While I don’t disagree with most of them, they add a lot of noise to the PR, making it difficult to identify the parts that have meaningfully changed. Can you resubmit the PR without stylistic/formatting changes? (I’d prefer to handle the formatting topic in a separate PR; the approach makes sense to me, but I think it should be consistent throughout the whole project, rather than confined to one file.)

p.s.,. Smart to use the conversion keys as the attributes filter! Great catch.

[jsvine]

Ah, I now see what I was missing: It is already possible to pass custom LAParams to pdfplumber; but when you do, it throws an error (as you note in #168). Your solution is to use the CONVERSIONS dict instead of the IGNORE list as the initial filter in Page.process_object(obj). That does seem to address the error — thanks! — but I want to take a slightly closer look at what was causing the error in the first place. I’ll update this thread when I do.

[jsvine]

Reviewing the cause of that error, it seems relatively benign (the addition of an index property to char objects), and your fix seems like a good one. I have incorporated it in df00787 and credited you. (Thanks!) The fix is now available in v0.5.16.

Closing this PR, but if you want to discuss code-formatting, feel free to open an issue on the topic.