Enable JSON response from the validator

[ml-evs]

This PR adds a -j/--json flag to the validator that lowers verbosity to "-1" and only prints a JSON representation of the results to stdout.

[codecov]

Codecov Report

Merging #502 into master will increase coverage by 0.00%.
The diff coverage is 97.82%.

@@           Coverage Diff           @@
##           master     #502   +/-   ##
=======================================
  Coverage   91.58%   91.59%           
=======================================
  Files          61       61           
  Lines        3103     3118   +15     
=======================================
+ Hits         2842     2856   +14     
- Misses        261      262    +1     

Flag	Coverage Δ
#project	91.59% <97.82%> (+<0.01%)	⬆️
#validator	63.72% <69.56%> (+0.11%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
optimade/validator/__init__.py	9.09% <0.00%> (-0.29%)	⬇️
optimade/validator/utils.py	90.29% <100.00%> (+0.14%)	⬆️
optimade/validator/validator.py	82.67% <100.00%> (+0.49%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e05e0b8...698a424. Read the comment docs.

[CasperWA]

This is basically a sale for me - I just review as a "Comment" to make sure my few comments are indeed considered :)

Mainly, if you create a class for the results attribute instead of a simple dict you could still use the old writing way of retrieving the various results, i.e., using . (dot) instead of key access.

[ml-evs]

Now we have the same problem as with the other PR, mkdocs can't handle dataclasses or NamedTuple :(

[CasperWA]

Now we have the same problem as with the other PR, mkdocs can't handle dataclasses or NamedTuple :(

I wonder if there is a comment syntax (like # pylint: disable=some-rule) to exclude a class from MkDocs?
Otherwise, we should contribute to MkDocs, or rather its doc-string parser 😕

[CasperWA]

Now we have the same problem as with the other PR, mkdocs can't handle dataclasses or NamedTuple :(

I wonder if there is a comment syntax (like # pylint: disable=some-rule) to exclude a class from MkDocs?
Otherwise, we should contribute to MkDocs, or rather its doc-string parser confused

Looking into the error from the CI run, it seems it is not MkDocs, but rather mkdocstrings.

[ml-evs]

Now we have the same problem as with the other PR, mkdocs can't handle dataclasses or NamedTuple :(

I wonder if there is a comment syntax (like # pylint: disable=some-rule) to exclude a class from MkDocs?
Otherwise, we should contribute to MkDocs, or rather its doc-string parser confused

Looking into the error from the CI run, it seems it is not MkDocs, but rather mkdocstrings.

Yeah sorry, that's what I mean every time I say mkdocs... this one seems weird, mkdocstrings is meant to support dataclasses and the error is to do with parsing the markdown api_reference/validator/validator.md 😕

[CasperWA]

Now we have the same problem as with the other PR, mkdocs can't handle dataclasses or NamedTuple :(

I wonder if there is a comment syntax (like # pylint: disable=some-rule) to exclude a class from MkDocs?
Otherwise, we should contribute to MkDocs, or rather its doc-string parser confused

Looking into the error from the CI run, it seems it is not MkDocs, but rather mkdocstrings.

Yeah sorry, that's what I mean every time I say mkdocs... this one seems weird, mkdocstrings is meant to support dataclasses and the error is to do with parsing the markdown api_reference/validator/validator.md confused

So what it's doing (I'm guessing) is following the reference (to optimade/validator/validator.py) and using pytkdocs to extract the docstrings, then it's converting this to XML, which is where it fails, because the generated XML file is badly made.

[ml-evs]

So what it's doing (I'm guessing) is following the reference (to optimade/validator/validator.py) and using pytkdocs to extract the docstrings, then it's converting this to XML, which is where it fails, because the generated XML file is badly made.

So it "works" if I remove the explicit dataclass.field fields, though it still returns warnings for __eq__, __init__ and __repr__ (as is the case with NamedTuple in the other PR) :(

[CasperWA]

Very kewl @ml-evs

Approved, and a single question :)

[CasperWA]

OooOooohh ✨

[CasperWA]

Just a question: Is there a reason you're not using pydantic here instead?

[ml-evs]

Uhh, guess this is slightly more lightweight but you could use either

[CasperWA]

Was just wondering if this would solve the documentation issue as well...