This is the docker image for ATIRE conforming to the OSIRRC jig for the Open-Source IR Replicability Challenge (OSIRRC) at SIGIR 2019. This image is available on Docker Hub. The OSIRRC 2019 image library contains a log of successful executions of this image.
- Supported test collections:
robust04, andcore17. - Supported hooks:
init,index,search
The following jig command can be used to index TREC disks 4/5 for robust04:
python3 run.py prepare \
--repo osirrc2019/atire \
--tag v0.1.0 \
--collections robust04=/path/to/disk45=trectext
e.g. python3 run.py prepare --repo osirrc2019/atire --collections robust04=/Users/andrew/programming/JASSv2/docker/osirrc2019/robust04=trectext
The following jig command can be used to perform a retrieval run on the collection with the robust04 test collection.
python3 run.py search \
--repo osirrc2019/atire \
--tag v0.1.0 \
--output out/atire \
--qrels qrels/qrels.robust04.txt \
--topic topics/topics.robust04.txt \
--collection robust04 \
--top_k 100"
For example:
python3 run.py search --repo osirrc2019/atire --collection robust04 \
--topic topics/topics.robust04.txt --top_k 100 \
--output /Users/andrew/programming/osirrc2019/jass-docker/output --qrels qrels/qrels.robust04.txt
BM25, BM25+ with s-stemming an Rocchio relevance feedback. Currently hard coded to the collection.
The following numbers should be able to be re-produced using the scripts provided by the jig.
TREC 2004 Robust Track Topics.
- BM25+: k1=2.0, b=0.5, d=0.2 with Rocchio relevance feedback with d=2, t=81, then BM25+ k1=1.1, b=0.6, d=0.6. All with s-stemming
| Metric | Score |
|---|---|
| MAP | 0.2184 |
| P@30 | 0.3199 |
TREC 2017 Common Core Track Topics.
- BM25: k1=0.9, b=0.4 (Robertson et al., 1995)
| Metric | Score |
|---|---|
| MAP | 0.1436 |
| P@30 | 0.4087 |
The following is a quick breakdown of what happens in each of the scripts in this repo.
The Dockerfile installs dependencies (python3, etc.), copies scripts to the root dir, and sets the working dir to /work.
The init script is straightforward - it's simply a shell script (via the #!/usr/bin/env sh she-bang) that downloads and builds ATIRE.
The index Python script (via the #!/usr/bin/python3 she-bang) reads a JSON string (see here) containing at least one collection to index (including the name, path, and format).
The collection is indexed and placed in the current working directory (i.e., /work).
At this point, jig takes a snapshot and the indexed collections are persisted for the search hook.
The search script reads a JSON string (see here) containing the collection name (to map back to the index directory from the index hook) and topic path, among other options.
The retrieval run is performed and output is placed in /output for the jig to evaluate using trec_eval.
- S. E. Robertson, S. Walker, M. Hancock-Beaulieu, M. Gatford, and A. Payne. (1995) Okapi at TREC-4. TREC
- A. Trotman, X.-F Jia, M. Crane (2012), Towards an Efficient and Effective Search Engine, Proceedings of the SIGIR 2012 Workshop on Open Source Information Retrieval, pp. 40-47
- Y. Lv, CX. Zhai (2011) Lower-Bounding Term Frequency Normalization, Proceedings of CIKM'11, pp. 7-16
- Documentation reviewed at commit
480b2f7(2019-06-16) by Ryan Clancy.