Tutorial of how to use the iSearch datasets with ir_datasets

Obtain the iSearch datasets:

• v1.0 https://sites.google.com/view/isearch-testcollection/
• v2.0 https://zenodo.org/records/20143498

Install the ir-datasets-isearch package (details can be found in the repository).

Place the datasets in the ir_datasets base directory. The directory tree should be as follows:

.
├── iSearch-v1.0_documents.tgz
├── iSearch-v1.0_topics+assessments
│ ├── assessments
│ │ ├── graded-qrels.all-types.txt
│ │ ├── graded-qrels.BK.txt
│ │ ├── graded-qrels.PF.txt
│ │ └── graded-qrels.PN.txt
│ └── topics
│ ├── topics.trecweb
│ ├── topics.xml
│ ├── xml2trec.sh
│ └── xml2trecweb.sh
├── iSearch-v2.0_assessments
│ └── assessments
│ ├── graded-qrels.NN.txt
│ └── graded-qrels.NO.txt
└── iSearch-v2.0_documents.tgz

Execute the code snippets in this notebook.

Note

Here you can find the landing page collecting all resources and materials: https://github.com/irgroup/isearch-pp