Running in python

[hildogjr]

Is v2 already in point of test?
I want to run it in Python to integrate in other tool. Should be interesting be available on PyPI.

Also, as a KiCad user. I believe could it be interesting have a way to parse the footprint names provided by KiCad and extract the really package name.

[kasbah]

Just to leave a note here (we were previously discussing this somewhere else when you opened this issue I think). Work on a Python version has stalled. I still think it's a great idea but yeah, no one is working on it at the moment.

[set-soft]

Hi @kasbah and @hildogjr !

• I made a port to Lark, the lark grammar can be found here.
• The class to convert the parsed tree into a dict compatible with the JavaScript is here.
• I also added some basic inductor parsing.
• The code is passing all the tests I found in the JS version, but I didn't implement the CPL stuff.
• A very important difference is that I'm not relying on catching parser errors, retry and keep the last match. I incorporated the flexibility of unknown stuff (which I call crap) in the grammar itself, after all Early algorithm can handle ambiguity. The crap is collected and you can check what was discarded. A drawback: is really slow.
• I also tried to port the ne files to Lark. It can read Nearly files and generate a Python code that emulates the JS. It took some work because it only supports JS 5 and not 6. But I discarded this approach because it pulls various dependencies.
• I kept the MIT license just in case @kasbah is interested in adding it to this project.
• Note that Lark is pure Python without any extra dependency, has a Debian package, and is usually installed in desktop systems.

[kasbah]

Very nice!