Update fonll docstr and json loading

[scarlehoff]

While updating this line

pineko/src/pineko/fonll.py

Line 125 in d75ebe1

Update by reading the yamldb of the original theory.

I tried to run it just because and noticed that not all fktables have the theories saved as valid json. In particular some of the HERA fktables in 701XX sometimes have single quotes instead of double so I've added a "bypass" for that as well (and a few boolean values are also wrong). Not sure why it happened.

[giacomomagni]

Thanks for spotting it. Which are the keys with the wrong syntax?
Maybe we should correct this also in pinefarm or yadism

[scarlehoff]

Which are the keys with the wrong syntax?

There're a few grids where the theory key uses ' instead of " (which is invalid json) for all keys.

For the True instead of true, I don't remember.

[felixhekhorn]

Thanks for spotting it. Which are the keys with the wrong syntax? Maybe we should correct this also in pinefarm or yadism

I was sure that pinefarm writes the theory as meta data, but I can't find it https://github.com/NNPDF/pinefarm/blob/39abf325d3ab8ec227f7c296935e55f966184434/src/pinefarm/external/interface.py#L120 what am I missing? (and it should be fixed in pinefarm)

[scarlehoff]

I was sure that pinefarm writes the theory as meta data

The theory is not even used by all runners so it might actually better that this is not the case yet... (or that only the ones that use it actually write it).

[scarlehoff]

I was sure that pinefarm writes the theory as meta data,

@felixhekhorn I was curious and checked that yadism actually writes this using json.dumps so the error was, in theory, impossible.

Indeed, the error is generated by pineko but I don't understand how and why it happens only for some datasets and not others.
If you are curious, HERACOMB_SIGMARED_B in theory 701 has this problem.

I'll open an issue about this.