Needs more tests

[vinayak-mehta]

Listing down some things which can be tested, additionally Codecov can be referred to look at the modules that have poor coverage.

• Whether an error/warning is raised or not (protip: check out scikit-learn tests)
  • Unknown flavor in read_pdf
  • utils.validate_input ValueError
  • File format error in PDFHandler
  • No tables found log in both parsers (should be converted to warnings)
• The CLI (currently, it's 0% covered!)
  • Testing Click applications
  • Cannot test Commdn groups with click.testing.CliRunner pallets/click#267
• Whether a type of plot is generated or not (currently, plotting.py is only 16% covered.)
  • https://github.com/matplotlib/pytest-mpl
  • http://www.davidketcheson.info/2015/01/13/using_matplotlib_image_comparison.html
  • https://github.com/pandas-dev/pandas/blob/master/pandas/tests/plotting/common.py#L528
• Test each configuration parameter (/combinations of them) for both Parsers. We could generate test PDFs from dataframes, using LaTeX, parse the PDFs using the library and then compare the output with the initial dataframes.

[vinayak-mehta]

@vaibhavmule A good start would be to install and use the library, get a feel of it and then add to the list above.

[vaibhavmule]

Sure, let me pull the repo and If I get stuck, will let you know.

[vinayak-mehta]

@vaibhavmule I've added tests for errors/warnings, moving onto CLI. Feel free to ping here if you're stuck. Will probably create a Gitter room soon.

[vinayak-mehta]

I'm doing number 4 in the list now.

[vinayak-mehta]

Opening a separate issue for testing the plotting module.

[vaibhavmule]

There is this library, https://github.com/matplotlib/pytest-mpl

[vinayak-mehta]

Yeah, I've mentioned it in the first comment on this issue. Opened #127.