trec02-ar.yaml

---
corpus: trec02-ar
corpus_path: collections/newswire/clir/trec.ar/arabic_newswire_a_ldc2001t55/transcripts/

index_path: indexes/lucene-index.trec02-ar/
collection_class: CleanTrecCollection
generator_class: DefaultLuceneDocumentGenerator
index_threads: 16
index_options: -storePositions -storeDocvectors -storeRaw -language ar
index_stats:
  documents: 383872
  documents (non-empty): 383872
  total terms: 62378088

metrics:
  - metric: MAP
    command: bin/trec_eval
    params: -m map
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: true
  - metric: P20
    command: bin/trec_eval
    params: -m P.20
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: true
  - metric: nDCG@20
    command: bin/trec_eval
    params: -m ndcg_cut.20
    separator: "\t"
    parse_index: 2
    metric_precision: 4
    can_combine: true

topic_reader: Trec
topics:
  - name: "[TREC 2002 (Monolingual Arabic)](https://github.com/castorini/anserini-tools/tree/master/topics-and-qrels/topics.trec02ar-ar.txt)"
    id: trec02
    path: topics.trec02ar-ar.txt
    qrel: qrels.trec02ar.txt

models:
  - name: bm25
    display: BM25
    params: -bm25 -language ar
    results:
      MAP:
        - 0.2932
      P20:
        - 0.3610
      nDCG@20:
        - 0.4056