add query iterator

[MXueguang]

in current version, we added a flag --ordered to the pypi replication to use the query order array.

$ python -m pyserini.search --topics msmarco_passage_dev_subset \
                                             --index msmarco-passage \
                                             --output run.msmarco-passage.txt \
                                             --bm25 --msmarco \
                                             --ordered

[lintool]

Can you kill the --ordered option? I can't think of a reason why you wouldn't want ordered?

But otherwise, 🚢 it!

[lintool]

Hrm, conflicts need to be resolved.

After this gets merged we need to re-do replication numbers.

Actually @MXueguang - one more thing, please document in query_iterator why we actually need to do this - a reference to the GitHub issue should do.

Happy to review it again if you want, but otherwise just merge please.

[lintool]

We'll also need to rerun multi-threaded search and make sure there are no weird interactions.

[MXueguang]

updated scores in pypi-replication.md

multithreading gives same results

[lintool]

lgtm!