Unpin igraph version

pyproject.toml

@@ -1,11 +1,11 @@
 [tool]
 [tool.poetry]
-name = "pixelgen-pixelator"
+name = "pixelator"
 version = "0.0.0"
 homepage = "https://github.com/PixelgenTechnologies/pixelator"
 repository = "https://github.com/PixelgenTechnologies/pixelator"
-documentation = ""
-description = "Software package to process sequencing FASTQ from Molecular Pixelation (MPX) assays"
+documentation = "https://software.pixelgen.com"
+description = "A commandline tool and library to process and analyze sequencing data from Molecular Pixelation (MPX) assays."
 authors = ["Pixelgen Technologies AB <[email protected]>",]
 maintainers = [
     "Alvaro Martinez Barrio <[email protected]>"
@@ -36,14 +36,14 @@ cutadapt = ">=4.2"
 umi_tools = "<=1.1.4"
 pyfastx = "*"
 yappi = "*"
-igraph = "0.10.2"
-pandas = ">=2.0.0"
+igraph = "0.10.*"
+pandas = "^2.0.0"
 numpy = "<1.24.0"
 annoy = "<=1.17.0"
 esda = "*"
 libpysal = "*"
 xopen = "*"
-anndata = "*"
+anndata = ">=0.8.0"
 numba = ">=0.56.4"
 scanpy = "*"
 leidenalg = "*"
@@ -52,7 +52,7 @@ cssselect = "*"
 typing_extensions = "*"
 scipy = "*"
 pyarrow = "*"
-semver = "*"
+semver = "^3.0.0"
 ruamel-yaml = "^0.17.21"
 pydantic = "^1.10.7"
 polars = "^0.17.12"
@@ -80,9 +80,9 @@ pytest-snapshot = "*"
 pytest-cov = "*"
 pytest-mock = "*"
 pytest-dependency = "*"
-sphinx-copybutton = "^0.5.2"
-sphinx-design = "^0.4.1"
-sphinx-inline-tabs = "^2023.4.21"
+sphinx-copybutton = ">=0.5.2"
+sphinx-design = ">=0.4.1"
+sphinx-inline-tabs = ">=2023.4.21"
 myst-parser = "^2.0.0"
 furo = "^2023.5.20"
 

src/pixelator/graph/community_detection.py

@@ -6,7 +6,7 @@
 import json
 import logging
 from pathlib import Path
-from typing import Dict, List, Optional, Tuple
+from typing import Dict, List, Optional, Set, Tuple
 
 import numpy as np
 import pandas as pd
@@ -158,7 +158,7 @@ def community_detection_crossing_edges(
     graph: Graph,
     leiden_iterations: int = 10,
     beta: float = 0,
-) -> List[Tuple[str]]:
+) -> List[Set[str]]:
     """Detect spurious edges connecting components by community detection.
 
     Use the Leiden [1]_ community detection algorithm to detect communities in a graph.
@@ -174,8 +174,8 @@ def community_detection_crossing_edges(
     :param beta: parameter to control the randomness on cluster selection when Leiden
                  merges clustering (0 - maximize objective function i.e. modularity;
                  inf - uniform distribution to merge with any other cluster)
-    :returns: a list of tuples with the edges between communities (edges ids)
-    :rtype: List[Tuple[str]]
+    :returns: a list of sets with the edges between communities (edges ids)
+    :rtype: List[Set[str]]
     :raises AssertionError: if the method is not supported
     """
     logger.debug(
@@ -203,8 +203,8 @@ def community_detection_crossing_edges(
         # get the crossing edges
         graph.es["is_crossing"] = vertex_clustering.crossing()
         edges = graph.es.select(is_crossing_eq=True)
-        # translate the edges to tuples of their corresponding vertex names
-        edges = [(e.vertex_tuple[0]["name"], e.vertex_tuple[1]["name"]) for e in edges]
+        # translate the edges to sets of their corresponding vertex names
+        edges = [{e.vertex_tuple[0]["name"], e.vertex_tuple[1]["name"]} for e in edges]
         logger.debug(
             "Community detection detected %i crossing edges in %i communities with a "
             "modularity of %f",
@@ -220,16 +220,16 @@ def community_detection_crossing_edges(
 def detect_edges_to_remove(
     edgelist: pd.DataFrame,
     leiden_iterations: int = 10,
-) -> List[Tuple[str]]:
+) -> List[Set[str]]:
     """Use Leiden algorithm to detect communities from an edgelist.
 
     This method uses the community detection Leiden algorithm to detect
     communities in the whole graph corresponding to the edge list given as input.
     Edges connecting the communities are computed and returned.
     :param edgelist: The edge list used to create the graph
     :param leiden_iterations: the number of iterations for the leiden algorithm
-    :return: A list of edges (tuple) that are connecting communities
-    :rtype: List[Tuple[str]]
+    :return: A list of edges (sets) that are connecting communities
+    :rtype: List[Set[str]]
     """
     logger.debug(
         "Detecting edges to remove using the leiden algorithm"
@@ -320,7 +320,7 @@ def recover_technical_multiplets(
     """
 
     def vertex_name_pairs_to_upis(
-        edge_tuples: List[Tuple[str]],
+        edge_tuples: List[Set[str]],
     ) -> List[str]:
         """Translate each pair of vertices into full UPI info.
 

src/pixelator/graph/graph.py

@@ -62,7 +62,12 @@ def from_edgelist(
             edgelist.shape[0],
         )
 
-        graph = igraph.Graph.DataFrame(edgelist, directed=False, use_vids=False)
+        vertices = pd.DataFrame(
+            set(edgelist["upia"].unique()).union(set(edgelist["upib"].unique()))
+        )
+        graph = igraph.Graph.DataFrame(
+            edgelist, vertices=vertices, directed=False, use_vids=False
+        )
         if "sequence" in edgelist.columns:
             all_sequences = edgelist["sequence"].unique()
             logger.debug(

src/pixelator/graph/utils.py

@@ -140,7 +140,9 @@ def create_node_markers_counts(
     if "markers" not in graph.vs.attributes():
         raise AssertionError("Could not find 'markers' in vertex attributes")
     markers = list(sorted(graph.vs[0]["markers"].keys()))
-    node_marker_counts = pd.DataFrame.from_records(graph.vs["markers"], columns=markers)
+    node_marker_counts = pd.DataFrame.from_records(
+        graph.vs["markers"], columns=markers, index=graph.vs["name"]
+    )
     node_marker_counts = node_marker_counts.reindex(
         sorted(node_marker_counts.columns), axis=1
     )

tests/analysis/colocalization/test_prepare.py

@@ -57,8 +57,8 @@ def test_prepare_from_graph_and_edgelist_eq_for_no_neigbours(edgelist):
         # We drop the indexes since whey will be named differently depending on the
         # method.
         assert_frame_equal(
-            graph_result.reset_index(drop=True),
-            edgelist_result.reset_index(drop=True),
+            graph_result.sort_index().reset_index(drop=True),
+            edgelist_result.sort_index().reset_index(drop=True),
             check_dtype=False,
             check_names=False,
             check_column_type=False,

tests/data/param_files/test_data_pe_T1.annotate.meta.json

@@ -6,8 +6,6 @@
             "--min-size": 1,
             "--max-size": 10000,
             "--dynamic-filter": null,
-            "--cell-type-assignments": false,
-            "--majority-vote": false,
             "--aggregate-calling": true,
             "--output": "."
         }

tests/data/param_files/test_data_pe_T2.annotate.meta.json

@@ -6,8 +6,6 @@
             "--min-size": 1,
             "--max-size": 10000,
             "--dynamic-filter": null,
-            "--cell-type-assignments": false,
-            "--majority-vote": false,
             "--aggregate-calling": true,
             "--output": "."
         }

tests/graph/test_graph.py

@@ -161,7 +161,7 @@ def test_community_detection_crossing_edges(graph_with_communities):
         graph=graph_with_communities,
         leiden_iterations=2,
     )
-    assert result == [("CTCGTACCTGGGACTGATACT", "TGTAAGTCAGTTGCAGGTTGG")]
+    assert result == [{"CTCGTACCTGGGACTGATACT", "TGTAAGTCAGTTGCAGGTTGG"}]
 
 
 def test_community_detection_crossing_edges_no_communities(graph_without_communities):
@@ -176,4 +176,4 @@ def test_community_detection_crossing_edges_no_communities(graph_without_communi
 def test_detect_edges_to_remove(edgelist_with_communities):
     """Test discovery of edges to remove from edgelist."""
     result = detect_edges_to_remove(edgelist_with_communities, leiden_iterations=2)
-    assert result == [("CTCGTACCTGGGACTGATACT", "TGTAAGTCAGTTGCAGGTTGG")]
+    assert result == [{"CTCGTACCTGGGACTGATACT", "TGTAAGTCAGTTGCAGGTTGG"}]

tests/graph/test_graph_utils.py

@@ -41,10 +41,7 @@ def pentagram_graph_fixture():
         (4, 1),
         (4, 2),
     ]
-    edgelist = pd.DataFrame(edges)
-    # add dummy upia/upib data since the creation methods requires it
-    edgelist["upia"] = "a"
-    edgelist["upib"] = "a"
+    edgelist = pd.DataFrame(edges, columns=["upia", "upib"])
     g = Graph.from_edgelist(
         edgelist=edgelist,
         add_marker_counts=False,