pochmann/wca-api-go

WCA API written in Go.

WCA API written in Go

Might end up as API on the WCA server, but so far an experiment using the WCA export so everyone can try it.

Online demo

• http://104.236.109.5:8080/cubers/2004MAOT02
• http://104.236.109.5:8080/cubers/2008AURO01/results
• http://104.236.109.5:8080/rankings/333

How to use it

1. Set up Go and Python and download or clone this repository.
2. Run get_wca_export.py in the data folder to download and unzip the current WCA export.
3. Run generate_loader.py to analyze the WCA export structure and build wca-data.go (which defines the types and loads the data in Go).
4. Compile with go install.
5. Run the server with wca-api-go.
6. Visit http://127.0.0.1:8080/cubers/2003POCH01.

To update using a new WCA export, repeat steps 2, 4 and 5. Step 3 is only necessary when the WCA export structure changes, though then the main program might need to get changed accordingly.

More views

• Visit http://127.0.0.1:8080/rankings/333 for the top 100 in 3x3 (change the eventId for other events).
• Visit http://127.0.0.1:8080/cubers/2008AURO01/results for Sébastien Auroux's results (he has the most).
• Run render_using_api.py for a speed test, getting Sébastien's results from the API and then rendering them to HTML (you'll need Jinja2).

Speed tests

Speeds on Stefan's Pentium 997 Linux laptop (using ab -n 1000 $url):

$ . ./speedtest.sh 

http://127.0.0.1:8080/cubers/2003POCH01
Time per request:       0.299 [ms] (mean)

http://127.0.0.1:8080/rankings/333
Time per request:       2.146 [ms] (mean)

http://127.0.0.1:8080/cubers/2008AURO01/results
Time per request:       0.312 [ms] (mean)

The last one is by far the biggest, but it's fast because it's pre-json'ed and pre-gzip'ed (only for Sébastien - doing it for everyone would take a bit over a minute).