Add Python script to parse replays using sc2reader to easily import it on the website

[BurnySc2]

See title.

Python script: reads a SC2 replays which reads the actions (make scv, drop mule, put scv into gas) and converts it to JSON format which can be easily read by typescript on the website.

[jtsauls]

Is the .json format the same as the one exported by "Copy SC2 Bot Instructions"?

Using the spawningtool code mentioned in #1 , this seems relatively straight forward and I could take a stab at it.

I am only confused what "frame" is in that .json. It's around 22-23 x elapsed seconds? Whats the formula for that?

@BurnySc2 Thoughts?

[BurnySc2]

The formula is time_in_seconds * 22.4 == frames, based on 16 frames per second of the SC2 engine, and because they used the 'faster' mode which multiplies it by 1.4 -> 16 * 1.4 = 22.4.

My goal was to host a separate site (with backend this time) where people could upload their replays and it would spit out a sc2-planner link. I have already started working on it a month ago but have less time recently.
I have made the repo public if you want to take a look at it, I'm using ggtracker/sc2reader
https://github.com/BurnySc2/sc2-replay-analyser

Now is just the frontend + backend interaction missing for the website (which I would write in React + Typescript + Sanic backend framework) - while the backend uses the sc2-replay-analyser library to parse/analyse replays.

If you purely use the BO instructions from spawning tool, there will be a lot of actions missing - e.g. chronoboost or dropping mules or injects, and sometimes there are weird artifacts in their output (e.g. line with leading comma, sometimes x2).

[jtsauls]

Yeah, that would be a really nice solution with a separate website. Looks like you kind have done all the heavy lifting--I was basically just thinking to write the function sc2planner.create_link. Using sc2reader is definitely the right way to go.

I'll poke around--but don't want to oversell my skills or time. I'm not sure how helpful I can be as I don't have any website experience.

Parsing of the text output from Spawning Tool would definitely be challenging and require assumptions.