Refactor calculate tasks (#29)

* Update convert to network task to take single dataframe * Simplify calculate measures tasks to take single network * Clean up calculate measures tasks * Remove calculate cluster tasks * Fix README capitalization
allen-cell-animated · Aug 30, 2023 · 9bf0695 · 9bf0695
1 parent 40ff0bc
commit 9bf0695
Show file tree

Hide file tree

Showing 8 changed files with 56 additions and 204 deletions.
diff --git a/README.md b/README.md
@@ -1,5 +1,5 @@
-[![Build Status](https://allen-cell-animated.github.io/abm-colony-collection/_badges/build.svg)](https://github.com/allen-cell-animated/abm-colony-collection/actions?query=workflow%3Abuild)
-[![Lint Status](https://allen-cell-animated.github.io/abm-colony-collection/_badges/lint.svg)](https://github.com/allen-cell-animated/abm-colony-collection/actions?query=workflow%3Alint)
+[![Build status](https://allen-cell-animated.github.io/abm-colony-collection/_badges/build.svg)](https://github.com/allen-cell-animated/abm-colony-collection/actions?query=workflow%3Abuild)
+[![Lint status](https://allen-cell-animated.github.io/abm-colony-collection/_badges/lint.svg)](https://github.com/allen-cell-animated/abm-colony-collection/actions?query=workflow%3Alint)
 [![Documentation](https://allen-cell-animated.github.io/abm-colony-collection/_badges/documentation.svg)](https://allen-cell-animated.github.io/abm-colony-collection/)
 [![Coverage](https://allen-cell-animated.github.io/abm-colony-collection/_badges/coverage.svg)](https://allen-cell-animated.github.io/abm-colony-collection/_coverage/)
 [![Code style](https://allen-cell-animated.github.io/abm-colony-collection/_badges/style.svg)](https://github.com/psf/black)

diff --git a/src/abm_colony_collection/__init__.py b/src/abm_colony_collection/__init__.py
@@ -1,10 +1,8 @@
-"""Tasks for analyzing colony dynamics including clustering and graph measures."""
+"""Tasks for analyzing colony dynamics including neighbors and graph measures."""
 
 from prefect import task
 
 from .calculate_centrality_measures import calculate_centrality_measures
-from .calculate_cluster_distances import calculate_cluster_distances
-from .calculate_cluster_sizes import calculate_cluster_sizes
 from .calculate_degree_measures import calculate_degree_measures
 from .calculate_distance_measures import calculate_distance_measures
 from .convert_to_network import convert_to_network
@@ -13,8 +11,6 @@
 from .make_voxels_array import make_voxels_array
 
 calculate_centrality_measures = task(calculate_centrality_measures)
-calculate_cluster_distances = task(calculate_cluster_distances)
-calculate_cluster_sizes = task(calculate_cluster_sizes)
 calculate_degree_measures = task(calculate_degree_measures)
 calculate_distance_measures = task(calculate_distance_measures)
 convert_to_network = task(convert_to_network)

diff --git a/src/abm_colony_collection/calculate_centrality_measures.py b/src/abm_colony_collection/calculate_centrality_measures.py
@@ -1,37 +1,22 @@
 import networkx as nx
-import numpy as np
 import pandas as pd
 
 
-def calculate_centrality_measures(networks: dict) -> pd.DataFrame:
-    all_measures = []
-
-    for (seed, tick), network in networks.items():
-        degree_centralities = list(nx.degree_centrality(network).values())
-        closeness_centralities = list(nx.closeness_centrality(network).values())
-        betweenness_centralities = list(nx.betweenness_centrality(network).values())
-
-        degree_centrality_mean = np.mean(degree_centralities)
-        closeness_centrality_mean = np.mean(closeness_centralities)
-        betweenness_centrality_mean = np.mean(betweenness_centralities)
-
-        degree_centrality_std = np.std(degree_centralities, ddof=1)
-        closeness_centrality_std = np.std(closeness_centralities, ddof=1)
-        betweenness_centrality_std = np.std(betweenness_centralities, ddof=1)
-
-        all_measures.append(
-            {
-                "SEED": seed,
-                "TICK": tick,
-                "DEGREE_CENTRALITY_MEAN": degree_centrality_mean,
-                "CLOSENESS_CENTRALITY_MEAN": closeness_centrality_mean,
-                "BETWEENNESS_CENTRALITY_MEAN": betweenness_centrality_mean,
-                "DEGREE_CENTRALITY_STD": degree_centrality_std,
-                "CLOSENESS_CENTRALITY_STD": closeness_centrality_std,
-                "BETWEENNESS_CENTRALITY_STD": betweenness_centrality_std,
-            }
-        )
-
-    all_measures_df = pd.DataFrame(all_measures)
-
-    return all_measures_df
+def calculate_centrality_measures(network: nx.Graph) -> pd.DataFrame:
+    # Calculate different centrality measures for network.
+    degree_centralities = nx.degree_centrality(network)
+    closeness_centralities = nx.closeness_centrality(network)
+    betweenness_centralities = nx.betweenness_centrality(network)
+
+    # Extract centrality measures for each node in network.
+    measures = [
+        {
+            "ID": node,
+            "DEGREE_CENTRALITY": degree_centralities[node],
+            "CLOSENESS_CENTRALITY": closeness_centralities[node],
+            "BETWEENNESS_CENTRALITY": betweenness_centralities[node],
+        }
+        for node in network.nodes
+    ]
+
+    return pd.DataFrame(measures)
diff --git a/src/abm_colony_collection/calculate_cluster_distances.py b/src/abm_colony_collection/calculate_cluster_distances.py
diff --git a/src/abm_colony_collection/calculate_cluster_sizes.py b/src/abm_colony_collection/calculate_cluster_sizes.py
diff --git a/src/abm_colony_collection/calculate_degree_measures.py b/src/abm_colony_collection/calculate_degree_measures.py
@@ -1,25 +1,15 @@
-import numpy as np
+import networkx as nx
 import pandas as pd
 
 
-def calculate_degree_measures(networks: dict) -> pd.DataFrame:
-    all_measures = []
+def calculate_degree_measures(network: nx.Graph) -> pd.DataFrame:
+    # Extract degree for each node in network.
+    measures = [
+        {
+            "ID": node,
+            "DEGREE": degree,
+        }
+        for node, degree in network.degree()
+    ]
 
-    for (seed, tick), network in networks.items():
-        degrees = sorted([d for n, d in network.degree()], reverse=True)
-        degree_mean = np.mean(degrees)
-        degree_std = np.std(degrees, ddof=1)
-
-        all_measures.append(
-            {
-                "SEED": seed,
-                "TICK": tick,
-                "DEGREES": degrees,
-                "DEGREE_MEAN": degree_mean,
-                "DEGREE_STD": degree_std,
-            }
-        )
-
-    all_measures_df = pd.DataFrame(all_measures)
-
-    return all_measures_df
+    return pd.DataFrame(measures)
diff --git a/src/abm_colony_collection/calculate_distance_measures.py b/src/abm_colony_collection/calculate_distance_measures.py
@@ -1,47 +1,23 @@
+from typing import Union
+
 import networkx as nx
-import numpy as np
 import pandas as pd
 
 
-def calculate_distance_measures(networks: dict) -> pd.DataFrame:
-    all_measures = []
-
-    for (seed, tick), network in networks.items():
-        if not nx.is_connected(network):
-            subnetworks = [
-                network.subgraph(component) for component in nx.connected_components(network)
-            ]
-
-            radius = np.mean([nx.radius(subnetwork) for subnetwork in subnetworks])
-            diameter = np.mean([nx.diameter(subnetwork) for subnetwork in subnetworks])
-
-            eccentricities = [nx.eccentricity(subnetwork) for subnetwork in subnetworks]
-            eccentricity = np.mean([np.mean(list(ecc.values())) for ecc in eccentricities])
+def calculate_distance_measures(network: nx.Graph) -> pd.DataFrame:
+    measures: list[dict[str, Union[int, float]]] = []
 
-            shortest_paths = [
-                nx.average_shortest_path_length(subnetwork) for subnetwork in subnetworks
-            ]
-            shortest_path = np.mean(shortest_paths)
-        else:
-            radius = nx.radius(network)
-            diameter = nx.diameter(network)
+    for component in nx.connected_components(network):
+        # Calculate eccentricity for connected subnetwork.
+        eccentricity = nx.eccentricity(network.subgraph(component))
 
-            ecc = nx.eccentricity(network)
-            eccentricity = np.mean(list(ecc.values()))
-
-            shortest_path = nx.average_shortest_path_length(network)
-
-        all_measures.append(
+        # Extract distance measures for each node in subnetwork.
+        measures = measures + [
             {
-                "SEED": seed,
-                "TICK": tick,
-                "RADIUS": radius,
-                "DIAMETER": diameter,
-                "ECCENTRICITY": eccentricity,
-                "SHORTEST_PATH": shortest_path,
+                "ID": node,
+                "ECCENTRICITY": eccentricity[node],
             }
-        )
-
-    all_measures_df = pd.DataFrame(all_measures)
+            for node in component
+        ]
 
-    return all_measures_df
+    return pd.DataFrame(measures)
diff --git a/src/abm_colony_collection/convert_to_network.py b/src/abm_colony_collection/convert_to_network.py
@@ -3,20 +3,15 @@
 
 
 def convert_to_network(neighbors: pd.DataFrame) -> nx.Graph:
-    networks = {}
+    nodes = list(neighbors["ID"].values)
+    edges = [
+        (node_id, neighbor_id)
+        for node_id, neighbor_ids in zip(neighbors["ID"], neighbors["NEIGHBORS"])
+        for neighbor_id in neighbor_ids
+    ]
 
-    for tick, group in neighbors.groupby("TICK"):
-        nodes = list(group["ID"].values)
-        edges = [
-            (node_id, neighbor_id)
-            for node_id, neighbor_ids in zip(group["ID"], group["NEIGHBORS"])
-            for neighbor_id in neighbor_ids
-        ]
+    network = nx.Graph()
+    network.add_nodes_from(nodes)
+    network.add_edges_from(edges)
 
-        network = nx.Graph()
-        network.add_nodes_from(nodes)
-        network.add_edges_from(edges)
-
-        networks[tick] = network
-
-    return networks
+    return network