Rank fusion using Pyserini

[Fatima-Haouari]

Kindly direct me to examples on how to use Pyserini to create fused runs using reciprocal_rank_fusion and taking the average/sum of scores.

[lintool]

We don't have the documentation written up, but consult this test case for usage: https://github.com/castorini/pyserini/blob/master/integrations/sparse/test_verify_rrf.py

A contribution to the documentation would be welcome.

[Fatima-Haouari]

Thanks a lot for you help. I managed to create my fused runs now. However I have the following issue.
The returned fused runs have the document IDs in float format (not string) which is causing a formatting issue with long documents IDs such the example below:
938526354907201539 Q0 1.2399882500827095e+18 10 11.35319995880127 interpolation
Kindly advice

[lintool]

Yea, there's some bug in the code resulting from Python duck typing. Please file an issue with a minimal example. It'd be even better if you could submit a patch that would fix the bug, complete with a test case. My guess is that the fix would be pretty simple, just an explicit cast somewhere...