Making Tracking and Hook/Test converge

[PonteIneptique]

Concerns mostly @srdee and me. @balmas a little

The way tracking.py is made would not be totally compatible with the tests I do. Most of all, it's used also to track conversion from old Perseus to new Perseus, while Hook/test ( https://github.com/Capitains/Hook/blob/dev/test/__init__.py ) test what is available and do not care about provenance.

I have been reusing stuff from MyCapytain and Tracking :) Stella, what are the point you think would need to be test in both ?

[srdee]

Sorry for being late in getting to this! If I understand the question correctly, I think the only things that would need to be tested in both are:

• EpiDoc compliance
• XML validity
• CTS metadata and refsDecl
• CTS compliance
• Ancient Greek betacode>UTF8 error (which we don't have a test for yet)

Does that sound about right? What am I missing?

[srdee]

This is my understanding (correct me if I'm wrong):

• tracking.py is mostly designed to provide information to editors, so they know what needs to be fixed and what comes from where, so the information can & should be more detailed
• Hook is designed to provide information to users, so they know what's available and whether or not they need to go become editors and fix text, at which point they can look at the more detailed information, so the information should be more concise and aggregated

if that's the case, the overlap is what I wrote above

[srdee]

@balmas what do you think? and @PonteIneptique now that the wiki's in a (hopefully) usable state, from tomorrow I'm going back and cleaning up the tracking scripts. Let me know if there's anything else you'd like to see there, apart from what's already in this issue list

[PonteIneptique]

Re : https://github.com/PerseusDL/tei-conversion-tools/issues/24#issuecomment-128148670

Hook is made for testing, the same way tracking is. The only difference is that it is made to be run automatically to provide information during pull request or commit on master branch. It will also issue individual files informations but it is not meant (Or not right now at least) for being used locally as it is a server. Logs of testing are available but the most important thing that is display is status (ie passed or failed test) and coverage (ie how many tests were successful).

The thing Hook is not for is tracking files movement between repositories, perseus copyrighted to non copyrighted repo, etc. Neither should it write tests result to the repo like tracking.py does.

Ps : By the way, you might find a good way to track citation/refsDecl here https://github.com/Capitains/Hook/blob/dev/test/test.py#L63-L101

[balmas]

@PonteIneptique is Hook also going to be able to run on branches as well as master? Could it be configured to call the tracking.py to update the status upon commit?

[PonteIneptique]

Writing something on the repository is more of milestone 2 or 3. It should be able to run other branches, but probably not the first iteration. Just PR and master.

[srdee]

@PonteIneptique thanks for the link! looks awesome