[Bug]: Inconsistent handling of 2.0 projects

[husqwc]

scratchattach version

2.1.13

What happened?

I wanted to see if sa.get_project() would work by trying to fetch https://scratch.mit.edu/projects/918894112. It didn't work, so I assumed that it couldn't fetch 2.0 projects. I run my code, and then came back only to discover that it crashed because https://scratch.mit.edu/projects/1023073403/ was a 2.0 project, and the file format was not the one my code was expecting. Please either completely support 2.0 projects or drop support for them entirely.

Your code.

print(sa.get_project(1023073403).raw_json()) // works
print(sa.get_project(918894112).raw_json()) // doesn't work

[faretek1]

the problem is that sometimes, when a request is made to the usual project JSON endpoint and the project is from scratch 2, the scratch server may return a sb2 (zip) file, instead of json data. it looks like this:

versus

The thing that is inconsistent here is scratch's handling of your code, not scratchattach, but that doesn't mean that this cannot be solved:

This can be parsed using python's zipfile library but it will take a bit. also note that neither scratchattach.editor nor project_json_capabilities support sb2 or sb files.

Having to make a special zipfile reader within a project json retriever function is a bit of a bad solution, as it would be better to implement this within an actual project parser (sa.editor or project_json_capabilities). But this is a future problem I guess

[faretek1]

fixed in #417
note: this fix will probably not be uploaded to pypi for a while. you could either wait or trying to install the indev version of scratchattach locally