👌`parse_stdout()`: add support for multiple EPW versions

[mbercx]

As new versions of EPW are released, the contents of the stdout might change. This leads to some questions:

1. Does EPW have an XML output like pw.x? Is that planned?
2. Do we want to have a compatibility policy, similar to what we have for aiida-quantumespresso?
3. How do we want to implement support for multiple EPW versions in the parse_stdout() method?

[ymzhang0]

As far as I know, EPW doesn't have any xml output now. I remember that @sponce24 mentioned EPW develops have the plan to have an xml output in the future.

I haven't used very old version of EPW. For standard output, I think 5.8 and 5.9 the formats are the same. After 5.9 till 6.0 now it changes a bit. Maybe currently we start from these two versions? For me I really like the new version 6.0 since it provides with the intermediate representation to speed up calculations for low temperature. But implementing things from the newest version to the precedents seems to be weird.

I think we can't infer the version of epw.x from inputs.code because it's not forced to report the version of code in the registration. Then the only way is to extract the version of code from the parser. The base parser will give the code_version in stdout, parsed_data, logs = self.parse_stdout_from_retrieved(logs). For the choice of parse_stdout function, either we have a if-else or we maintain some registry.

[mbercx]

Maybe currently we start from these two versions?

So we'll support EPW>=5.8, i.e. v5.8, v5.9 and v6.0? Sure, that makes sense to me. In any case we'll have to support multiple versions, so this will be a good way to figure out the least painful way of doing so. :) Note: our current test files are for EPW v5.7.

PS: What do EPW versions mean? Do they follow the SemVer MAJOR/MINOR concept? How do you decide to do a major or minor release? Maybe @sponce24 can also comment here.

Then the only way is to extract the version of code from the parser.

Sure, for parsing this is fine. If there is ever a change in the input file(s), that's a different story. But we can have that fight another day. 😉

For the choice of parse_stdout function, either we have a if-else or we maintain some registry.

This we'll have to think about. Having multiple parse_stdout functions might duplicate a lot of code. Having too many if/else statements inside the parse_stdout function might be messy. Let's start simple: do you have an example of an output that we already parse that has changed? If not: one that we want to parse but has changed in v6.0?

[ymzhang0]

OK If the current files are from v5.7 then I would assume v5.7 share the same format of output files with v5.8.

I'm not sure how EPW developers define the version of EPW. There seems rare patch version in EPW releases. Usually they have major.minor. Let's wait for @sponce24 to have some explaination.

I do have an example outputs for v6.0. I attach the file here. I can't find immediately the output for v5.8 but since it is the same as v5.7 maybe you can compare it with the one in the current test folder.

v6.0.a2f.txt

There are changes for example:
Estimated Allen-Dynes Tc = 13.939630 K for muc = 0.13000 --> Estimated Tc using Allen-Dynes modified McMillan expression = 4.1950 K.