ranx

library0.3.21pypypi✓ verified 86d ago

ranx is a blazing-fast Python library for ranking evaluation, comparison, and fusion. Version 0.3.21 supports Python >=3.8. It provides efficient computation of IR metrics (MAP, NDCG, etc.), statistical significance tests, and result fusion. Release cadence is irregular.

ai-ml search

pip install ranx

github ↗homepage ↗

INSTALL

IMPORT

SIG · RANX

ranx

ai-mlpythonv0.3.21

harness data pending

Install & Compatibility

Where this runs

No compatibility data collected yet for this library.

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

compare

✓ from ranx import compare

correct import

fuse

✓ from ranx import fuse

correct import

evaluate

✓ from ranx import evaluate

correct import

Create Qrels and Runs, evaluate metrics, compare runs with statistical tests, and fuse runs.

from ranx import evaluate, compare, fuse from ranx import Run, Qrels # Create Qrels and Run objects from dicts qrels_dict = {"q1": {"d1": 1, "d2": 0}, "q2": {"d3": 1}} run_dict = {"q1": {"d1": 0.9, "d2": 0.8}, "q2": {"d3": 0.95}} qrels = Qrels(qrels_dict) run = Run(run_dict) # Evaluate single run results = evaluate(qrels, run, "ndcg@5,map@5") print(results) # Compare two runs run2_dict = {"q1": {"d1": 0.85, "d2": 0.75}, "q2": {"d3": 0.9}} run2 = Run(run2_dict) report = compare(qrels, [run, run2], metrics=["ndcg@5", "map@5"]) print(report) # Fuse runs combined = fuse([run, run2], method="rbc") print(combined)

Debug

Known issues

breakingIn ranx 0.2.x, the API used `Qrels` and `Run` from `ranx` but some functions expected pandas DataFrames. In 0.3.x, all inputs are dict-based or custom objects. Old code using DataFrame directly may break.

fix

Convert DataFrames to dicts or use `Qrels.from_df()` and `Run.from_df()` methods.

affects: <0.3.0

gotchaMetrics string must not contain spaces. e.g., 'ndcg@5, map@5' will cause a parsing error. Use 'ndcg@5,map@5'.

fix

Remove spaces: 'ndcg@5,map@5'

affects: All

gotcharanx uses 0-based relevance scores internally, but Qrels values should be non-negative integers. Using floats may cause unexpected behavior.

fix

Ensure relevance scores are integers (0, 1, 2, ...).

affects: All

Errors

Common errors & fixes

ModuleNotFoundError: No module named 'ranx'

ranx not installed or installed in a different environment.

fix

Run `pip install ranx` in your current Python environment.

KeyError: 'qrels'

Passing a dict directly to evaluate() without wrapping in Qrels.

fix

Wrap qrels dict in `Qrels(qrels_dict)` before passing to evaluate.

Upgrade

Version history

0.3.21latest on PyPI · released Aug 7, 2025

Audit

Dependencies

numpyrequiredNumerical computations and array operations.

scipyrequiredStatistical tests (e.g., t-test, Wilcoxon).

pandasoptionalData handling for results and runs.

Agent activity

39 hits · last 30 days

node

Resources

githubhttps://github.com/AmenRa/ranx ↗

homepagehttps://github.com/AmenRa/ranx ↗