Add a depth-max for test logs

[benjello]

New features

• Introduce a maximal depth for compuation logs
  • Allows for limiting the depth of the computation log chain

[coveralls]

Coverage decreased (-0.003%) to 78.936% when pulling 6529d120aee94c39ceb9547bb4dc5e1e7a9b894a on depth-max into b9ead83dc62d702cc50a0593e4de438b3454a286 on master.

[bonjourmauko]

Hi @benjello and @nikhilwoodruff ! Thanks for this contribution.

There's already an aggregate option for the tracer, this is a different feature? https://openfisca.org/doc/openfisca-python-api/tracer.html#openfisca_core.tracers.Tracer.print_computation_log

[bonjourmauko]

Also @benjello could you please:

1. Provide a command to test the feature in the command line? :)
2. Tracer is unit tested, so for this PR to get accepted I suggest adding a unit test

[benjello]

Hi @benjello and @nikhilwoodruff ! Thanks for this contribution.

There's already an aggregate option for the tracer, this is a different feature? https://openfisca.org/doc/openfisca-python-api/tracer.html#openfisca_core.tracers.Tracer.print_computation_log

Yes completely different. The agreggate option is for test on population data.

[benjello]

Thank you very much @maukoquiroga !

[benjello]

Last question for you @maukoquiroga : should I squash the commits ?

[benjello]

I do however, in order to integrate it, need we work on three aspects of it:

1. Having a reproducible way of testing this. I suggest we complete https://openfisca.org/doc/openfisca-python-api/openfisca_test.html with an example of using this flag.

There is no example of output in that page so I do not know what example to use.

It is actually a problem of the online doc. It does not come with easily reproductible examples ...

[benjello]

@maukoquiroga : could we go on and merge (and open a related issue in the openfisca-doc) ?