lvu / pref_analyzer

Preferans position analyzer

Preferans position analyzer

Two implementations of recursive analyzer for preferans game positions.

I wrote the Python one initially, but it was too slow, so I've reimplemented it in Scala. But I didn't want to throw the Python implementation away, so it's here too. Scala implementation is about 50 times faster.

Input format

The input is a position description in JSON format; two samples are provided in files kow_miser.json and train.json; the first one is the famous Sofia Kowalewskaya misère.

The format should be rather straightforward; some top-level fields to note:

• misere: Boolean, defaults to false;
• turn: East, West or South, the player making the first move;
• trump: Unicode symbol of trump suit; no trumps by default.

Invocation

To launch the Python implementation, run

    python pref_analyzer.py < position.json

for Scala implementation:

    sbt run < position.json

or

    sbt package
    java -cp target/scala-2.12/pref_analyzer_2.12-0.1.0-SNAPSHOT.jar:$(cat target/streams/compile/dependencyClasspath/_global/streams/export) lvu.pref.PrefAnalyzer < position.json

(modify according to your Scala version etc).