Conditional and table formatting support for html

[mikkokoo]

This is:

• a bugfix
• a new feature
• refactoring
• additional unit tests

Checklist:

• Changes are covered by unit tests
  • Changes are covered by existing unit tests
  • New unit tests have been added
• Code style is respected
• Commit message explains why the change is made (see https://github.com/erlang/otp/wiki/Writing-good-commit-messages)
• CHANGELOG.md contains a short summary of the change and a link to the pull request if applicable
• Documentation is updated as necessary

Why this change is needed?

Conditional and table formatting support for html writer were needed to provide better html output for processed files.

Fix #1375.

[oleibman]

Here is something like what I would like to see for HtmlTableFormatWithConditionalTest.
HtmlTableFormatWithConditionalTest.zip

[mikkokoo]

I did some profiling and found out that the getHashCode functions in Style/ get called a lot. They make the conditional formatting pretty slow. I managed to cache the hash code so that the execution time dropped by 75%.

@oleibman Should I commit that to this PR (I think yes) or make another pull request after this one?

10 files changed, 242 insertions(+), 61 deletions(-)

[oleibman]

Please make a separate PR for your HashCode improvements.

[oleibman]

This is still an overwhelmingly large number of tests. Can you model it after the others and just test a few "interesting" cells in each row, and only for the data you're looking for (background color and value).

[mikkokoo]

It is the test of different color scale formulas. There has to be quite a few data points to have meaningful results as the style is the only measurable data. They are all interesting. I'd say this is the bare minimum.

[oleibman]

Again, this is just too many tests for me to absorb. Please just choose a representative sample. Maybe a couple from column 1 (all of this column seems to match whatever test it needs), the matching cell from column 2, one or two non-matching cells from column 2, and one or two cells from column 3 (these might be a little trickier because conditional formatting doesn't apply to them). It is not that the tests I'm asking you to remove aren't worthwhile - they most decidedly are. For that reason, I would suggest commenting them out rather than deleting them. I think the "missing" tests are better handled by also creating new samples, so that we can confirm visually that the html matches the original xlsx. I will add a comment to your ticket later on how that might be done.

[mikkokoo]

There is a different code path for each test and they use a different conditional case. They are lightning fast as they only check from the same loaded data. I don't understand the rationale of how removing some of them makes the regression testing better.

Making the xlsx to samples makes sense. But I'd still keep these.

[oleibman]

As discussed above, I would like you to use your xlsx files for both unit tests and samples. Here is how you might go about it.
htmlcondtosamples.zip

[oleibman]

Thank you for all the extra work. I need to do a final review, but should be able to merge this in the next day or two.

[oleibman]

Thank you for your contribution. A complicated but much needed addition. I shall look forward to seeing what you found out about HashCode.