Working through the Foundations of Retrieval onboarding path, lessons 1-2. Both docs/start-here.md and docs/experiments-msmarco-passage.md instruct:
git submodule update --init --recursive
and then call python tools/scripts/msmarco/convert_collection_to_jsonl.py (also filter_queries.py, msmarco_passage_eval.py, convert_msmarco_to_trec_*.py).
The tools/ submodule was removed in #3382 , so there's no submodule to init and tools/scripts/ doesn't exist. I worked around it by git clone https://github.com/castorini/eval.git tools, which makes every tools/scripts/msmarco/... path resolve as written.
Happy to send a PR once you confirm the preferred fix (clone eval into tools/ in the docs vs. vendoring the scripts vs. something else).
Working through the Foundations of Retrieval onboarding path, lessons 1-2. Both docs/start-here.md and docs/experiments-msmarco-passage.md instruct:
git submodule update --init --recursive
and then call python tools/scripts/msmarco/convert_collection_to_jsonl.py (also filter_queries.py, msmarco_passage_eval.py, convert_msmarco_to_trec_*.py).
The tools/ submodule was removed in #3382 , so there's no submodule to init and tools/scripts/ doesn't exist. I worked around it by git clone https://github.com/castorini/eval.git tools, which makes every tools/scripts/msmarco/... path resolve as written.
Happy to send a PR once you confirm the preferred fix (clone eval into tools/ in the docs vs. vendoring the scripts vs. something else).