notice of new project about python-chess-annotator

[jvsteiner]

good call on linking back to the stockfish folks, have done that. I actually started this out as a terraform script, where I'd spin up some EC2 to do the analysis. Turns out all the overhead of installing the dependencies etc takes a fair amount of time, so any benefit you'd get from lower average compute rates there goes out the window. Also it's much easier to parallelize multiple games in the PGN file with Lambda. You can run a 5 min analysis on 400 games in 5 mins this way. I think I'll look into the Parameter Store idea actually - the point of this is to learn more about serverless, so why not. Any way, thanks again.

[rpdelaney]

Hey, that's pretty cool. I had an idea to do something similar but never got around to it.

Truth be told I used this project to teach myself python, but that was a long time ago, and now whenever I look at this code I throw up in my mouth a little. I'd love to rewrite it from scratch some day, but there's always something else... pgn2json sounds interesting... ;)

I wonder if another service besides Lambda would be more cost-effective?

By the way, Serverless can read values out of AWS Parameter Store. I'm working on something very similar to this in my day job, except using Lambda to deploy ECS services. By reading the environment variable in from SSM, you could manipulate the value in SSM without redeploying the app. Just a thought.

P.S. Since you're distributing a stockfish binary you might want to consider adding a link in the README back to the repo where you got the code.