gpath2vec 3.0.0__tar.gz → 3.0.1__tar.gz

{gpath2vec-3.0.0 → gpath2vec-3.0.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: gpath2vec
-Version: 3.0.0
+Version: 3.0.1
 Summary: gene-set to biological pathway embeddings with enrichment analysis
 Home-page: https://github.com/teslajoy/gpath2vec
 Author: Nasim Sanati

{gpath2vec-3.0.0 → gpath2vec-3.0.1}/gpath2vec/__init__.py

@@ -7,4 +7,9 @@ from . import ea
 from . import utils
 from . import compare
 
-__version__ = "1.0.0"
+from importlib.metadata import version, PackageNotFoundError
+
+try:
+    __version__ = version("gpath2vec")
+except PackageNotFoundError:
+    __version__ = "unknown"

{gpath2vec-3.0.0 → gpath2vec-3.0.1}/gpath2vec/cli.py

@@ -22,6 +22,21 @@ from gpath2vec.aucell import compute_aucell, topk_per_niche
 METHODS = ["metapath2vec", "svd", "spectral", "line", "vae"]
 
 
+def _cluster_weights(ea_df, weight):
+    """cluster -> {stId: edge weight} over significant pathways."""
+    # vectorized; avoids materializing ea_df as a python list of dicts
+    sig_df = ea_df[ea_df["sig_pathway"].astype(bool)]
+    if weight == "fdr":
+        sig_df = sig_df.assign(w=1 - sig_df["fdr_bh"])
+    else:
+        sig_df = sig_df.assign(w=sig_df["oddsratio"])
+    # no leading underscore: itertuples renames such fields positionally
+    clusters = {}
+    for r in sig_df[["cluster", "stId", "w"]].itertuples(index=False):
+        clusters.setdefault(str(r.cluster), {})[str(r.stId)] = float(r.w)
+    return clusters
+
+
 def _run_embedder(method, graph, ea_mat, study_id, dimensions, epochs, lr,
                   window, seed=1234):
     """run the selected embedding method, return embeddings dict."""
@@ -294,15 +309,7 @@ def run_pipeline(genes, gene_sets, output_dir, study_id, level, gene_filter,
 
     # network
     click.echo("step 2: network")
-    # vectorized; avoids materializing ea_df as a python list of dicts
-    sig_df = ea_df[ea_df["sig_pathway"].astype(bool)]
-    if weight == "fdr":
-        sig_df = sig_df.assign(_w=1 - sig_df["fdr_bh"])
-    else:
-        sig_df = sig_df.assign(_w=sig_df["oddsratio"])
-    clusters = {}
-    for r in sig_df[["cluster", "stId", "_w"]].itertuples(index=False):
-        clusters.setdefault(str(r.cluster), {})[str(r.stId)] = float(r._w)
+    clusters = _cluster_weights(ea_df, weight)
 
     # same contract as aggregate_min_fdr: one entry per pathway with min FDR
     pw_min = ea_df.groupby("stId", as_index=False)["fdr_bh"].min()

{gpath2vec-3.0.0 → gpath2vec-3.0.1}/gpath2vec.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: gpath2vec
-Version: 3.0.0
+Version: 3.0.1
 Summary: gene-set to biological pathway embeddings with enrichment analysis
 Home-page: https://github.com/teslajoy/gpath2vec
 Author: Nasim Sanati

{gpath2vec-3.0.0 → gpath2vec-3.0.1}/gpath2vec.egg-info/SOURCES.txt

@@ -16,5 +16,6 @@ gpath2vec.egg-info/entry_points.txt
 gpath2vec.egg-info/requires.txt
 gpath2vec.egg-info/top_level.txt
 tests/test_aucell_topk.py
+tests/test_cli_cluster_weights.py
 tests/test_ea_aggregation.py
 tests/test_metapath_topology.py

{gpath2vec-3.0.0 → gpath2vec-3.0.1}/setup.py

@@ -1,6 +1,6 @@
 from setuptools import setup, find_packages
 
-__version__ = "3.0.0"
+__version__ = "3.0.1"
 
 setup(
     name="gpath2vec",

gpath2vec-3.0.1/tests/test_cli_cluster_weights.py

@@ -0,0 +1,41 @@
+"""unit tests for end2end step 2 cluster weights."""
+
+import pandas as pd
+import pytest
+
+import gpath2vec
+from gpath2vec.cli import _cluster_weights
+
+
+def _ea_df():
+    return pd.DataFrame({
+        "cluster": ["A", "A", "B"],
+        "stId": ["R-HSA-X", "R-HSA-Y", "R-HSA-X"],
+        "fdr_bh": [0.01, 0.5, 0.2],
+        "oddsratio": [4.0, 1.1, 2.5],
+        "sig_pathway": [True, False, True],
+    })
+
+
+def test_fdr_weight_over_significant_only():
+    out = _cluster_weights(_ea_df(), "fdr")
+    assert set(out) == {"A", "B"}
+    assert set(out["A"]) == {"R-HSA-X"}
+    assert out["A"]["R-HSA-X"] == pytest.approx(0.99)
+    assert out["B"]["R-HSA-X"] == pytest.approx(0.8)
+
+
+def test_oddsratio_weight():
+    out = _cluster_weights(_ea_df(), "oddsratio")
+    assert out["A"]["R-HSA-X"] == pytest.approx(4.0)
+    assert out["B"]["R-HSA-X"] == pytest.approx(2.5)
+
+
+def test_no_significant_pathways_is_empty():
+    df = _ea_df().assign(sig_pathway=False)
+    assert _cluster_weights(df, "fdr") == {}
+
+
+def test_version_matches_distribution_metadata():
+    from importlib.metadata import version
+    assert gpath2vec.__version__ == version("gpath2vec")