search

#!/usr/bin/env bash

cd /work/anserini

# MS MARCO data directory
DATA_DIR=./msmarco_data

# Only interested in queries from the qrels file
python3 ./src/main/python/msmarco/filter_queries.py \
    --qrels=${DATA_DIR}/qrels.dev.small.tsv \
    --queries=${DATA_DIR}/queries.dev.tsv \
    --output_queries=${DATA_DIR}/queries.dev.small.tsv

# Do retrieval
python3 ./src/main/python/msmarco/retrieve.py \
    --index ${DATA_DIR}/lucene-index-msmarco \
    --qid_queries ${DATA_DIR}/queries.dev.small.tsv \
    --output ${DATA_DIR}/run.dev.small.tsv --hits 1000

# MS MARCO eval on run
python3 ./src/main/python/msmarco/msmarco_eval.py \
    ${DATA_DIR}/qrels.dev.small.tsv \
    ${DATA_DIR}/run.dev.small.tsv

# Convert run to TREC format
python3 ./src/main/python/msmarco/convert_msmarco_to_trec_run.py \
    --input_run ${DATA_DIR}/run.dev.small.tsv \
    --output_run ${DATA_DIR}/run.dev.small.trec

# Copy to output dir for trec_eval
cp ${DATA_DIR}/run.dev.small.trec /output/