Improve performance of VS Test results file parsing

[tlecomte]

On a VS test result file containing many tests, the current code is pretty slow. This was observed when migrating a codebase from NUnit3 result files to trx.

Here we improve the performance by caching the parsing of the results:

• a lookup of tests per feature
• a dictionary of test results per test id

On a sample file with about 1000 tests, the performance improves from 40.2 seconds to 2.5 seconds.

This was tested with:

.\Pickles.exe -f xxx -o xxx --trfmt=vstest --enableComments=false -df=json -lr=xxx.trx

The resulting pickledFeatures.json file is identical to the one before these changes (except the GeneratedOn field).

[tlecomte]

Hello @dirkrombauts ! This is similar to #503 from 2 years ago, but for VS tests this time 😄

I've seen your message in the readme about your search for a new maintainer. I hope you'll be willing to accept this PR anyway!

[dirkrombauts]

@tlecomte thank you for your contribution!

I have moved away from Pickles. A while ago I wrote a message about it being time to hand over the torch.

You are still actively using Pickles, and just now wrote a contribution. Do you want to take over the project? I will give you full access to the repository and everything, and sign over the website domain to you.

[tlecomte]

Thanks @dirkrombauts for the offer. Unfortunately I don't think I have enough time to step in. Really sorry !

[dirkrombauts]

Too bad. I will merge the code into develop. I currently have no plan on when to release a new version.

[dirkrombauts]

Released in version v2.21.0.