saezlab/partipy_paper

How saezlab/partipy_paper works

ParTIpy’s repository is a research-reproduction suite for archetypal analysis (AA) and Pareto task inference (PTI), rather than the implementation of the ParTIpy library itself. Researchers run data-specific scripts to prepare single-cell or spatial datasets, compare AA initialization and optimization strategies, evaluate coreset accuracy and runtime, and measure memory use; separate example scripts demonstrate analyses. Benchmark summaries and figures report results. README descriptions establish the major experiment branches, while sampled source confirms dataset loading, preprocessing, and AA calls; unsampled internals are represented at directory level, with uncertain links left out.

Main entry points: output Results and figures · data_utils.py Dataset loaders · benchmark_algorithms AA algorithm benchmarks · benchmark_coresets Coreset experiments

Components 14

Data access

Omics datasets
Dataset loaders code/utils/data_utils.py
Paths and seeds code/utils/const.py

Experiments

AA algorithm benchmarks code/benchmark_algorithms
Coreset experiments code/benchmark_coresets
Memory benchmarks code/benchmark_memory
PCHA comparisons code/benchmark_other

Examples

Omics examples code/examples

Result summaries

Algorithm summaries code/benchmark_algorithms/bench_meta.py
Coreset summaries code/benchmark_coresets/coreset_meta.py
Coreset v2 summaries code/benchmark_coresets/coreset_meta_v2.py
Memory summaries code/benchmark_memory/memory_meta.py
Results and figures output

Other components

Researcher

Connections 21

Researcher → AA algorithm benchmarks: runs benchmarks README.md
Researcher → Coreset experiments: runs experiments README.md
Researcher → Memory benchmarks: runs benchmarks README.md
Researcher → Omics examples: runs examples README.md
Dataset loaders → Omics datasets: downloads data code/utils/data_utils.py
AA algorithm benchmarks → Dataset loaders: loads datasets code/benchmark_algorithms/lupus_bench.py
Coreset experiments → Dataset loaders: loads datasets code/benchmark_coresets/lupus_coreset.py
Memory benchmarks → Dataset loaders: loads datasets code/benchmark_memory/k562_memory_bench.py
PCHA comparisons → Dataset loaders: loads datasets code/benchmark_other/ms_pypcha.py
AA algorithm benchmarks → Paths and seeds: uses settings code/benchmark_algorithms/lupus_bench.py
Coreset experiments → Paths and seeds: uses settings code/benchmark_coresets/lupus_coreset.py
Memory benchmarks → Paths and seeds: uses settings code/benchmark_memory/k562_memory_bench.py
AA algorithm benchmarks → Results and figures: writes results
Coreset experiments → Results and figures: writes results
Memory benchmarks → Results and figures: writes results
PCHA comparisons → Results and figures: writes results
Omics examples → Dataset loaders: loads datasets
Algorithm summaries → Results and figures: summarizes results
Coreset summaries → Results and figures: summarizes results
Coreset v2 summaries → Results and figures: summarizes results
Memory summaries → Results and figures: summarizes results

Drawn by GitDiagram on November 11, 2026 from the repository's file tree, README and source files. Read as Markdown ·