See https://tira-io.github.io/ir-lab-ws-23/ for details.
Some examples:
- Topic with high nDCG@10 variange
- Topic with low nDCG@10 variance
- Topic with single relevant document retrieved by almost all systems on position 1
- Browsing some documents
- More examples to come.
Components:
- Indexing Backend (test coverage:
) - UI (test coverage:
)
Add new datasets via:
./construct_indices/remote_indexes.py
./construct_indices/construct_topics_for_ui.py
We use devcontainers for development. To start your environment, either use Github Codespaces (click on "Code" -> "Codespaces" in Github to open one) as the easiest way to get started or devpod as open source alternative (directly pointing to our Kubernetes or your local docker installation).