collect: fix Pool type annotation

src/pp5/collect.py

@@ -14,7 +14,7 @@
 from pathlib import Path
 from datetime import datetime
 from dataclasses import dataclass
-from multiprocessing.pool import AsyncResult
+from multiprocessing.pool import Pool, AsyncResult
 
 import numpy as np
 import pandas as pd
@@ -183,7 +183,7 @@ def collect(self) -> dict:
         LOGGER.info(f"Collection metadata:\n" f"{collection_meta_formatted}")
         return self._collection_meta
 
-    def _finalize_collection(self, pool):
+    def _finalize_collection(self, pool: Pool):
         LOGGER.info(f"Finalizing collection for {self.id}...")
         if self.out_dir is None:
             return
@@ -438,7 +438,7 @@ def _collection_functions(self):
             "Write dataset": self._write_dataset,
         }
 
-    def _collect_precs(self, pool: mp.Pool):
+    def _collect_precs(self, pool: Pool):
         meta = {}
         pdb_ids = self.query.execute()
         n_structs = len(pdb_ids)
@@ -481,7 +481,7 @@ def _collect_precs(self, pool: mp.Pool):
         )
         return meta
 
-    def _filter_collected(self, pool: mp.Pool) -> dict:
+    def _filter_collected(self, pool: Pool) -> dict:
         """
         Filters collected structures according to conditions on their metadata.
         """
@@ -534,14 +534,14 @@ def _update_rejected_counts(filter_name: str, idx: pd.Series):
             "n_collected_filtered": len(df_filtered),
         }
 
-    def _filter_metadata(self, pool: mp.Pool, df_all: pd.DataFrame) -> pd.Series:
+    def _filter_metadata(self, pool: Pool, df_all: pd.DataFrame) -> pd.Series:
         # Even though we query by resolution, the metadata resolution is different
         # than what we can query on. Metadata shows resolution after refinement,
         # while the query is using data collection resolution.
         idx_filter = df_all[COL_RESOLUTION].astype(float) <= self.resolution
         return idx_filter
 
-    def _filter_redundant_unps(self, pool: mp.Pool, df_all: pd.DataFrame) -> pd.Series:
+    def _filter_redundant_unps(self, pool: Pool, df_all: pd.DataFrame) -> pd.Series:
 
         if self.seq_similarity_thresh == 1.0:
             LOGGER.info("Skipping sequence similarity filter...")
@@ -620,7 +620,7 @@ def _filter_redundant_unps(self, pool: mp.Pool, df_all: pd.DataFrame) -> pd.Seri
         filtered_idx = df_all[COL_UNP_ID].isin(filtered_unp_ids)
         return filtered_idx
 
-    def _write_dataset(self, pool: mp.Pool) -> dict:
+    def _write_dataset(self, pool: Pool) -> dict:
         all_csvs = tuple(self.prec_csv_out_dir.glob("*.csv"))
         n_pdb_ids = len(all_csvs)
         LOGGER.info(f"Creating dataset file from {n_pdb_ids} precs...")
@@ -812,7 +812,7 @@ def _collection_functions(self):
             "Collect pgroups": self._collect_all_pgroups,
         }
 
-    def _collect_all_structures(self, pool: mp.Pool):
+    def _collect_all_structures(self, pool: Pool):
         meta = {}
 
         if self._all_file:
@@ -864,7 +864,7 @@ def _collect_all_structures(self, pool: mp.Pool):
         meta["n_all_structures"] = len(self._df_all)
         return meta
 
-    def _collect_all_refs(self, pool: mp.Pool):
+    def _collect_all_refs(self, pool: Pool):
         meta = {}
 
         if self._ref_file:
@@ -906,7 +906,7 @@ def _collect_all_refs(self, pool: mp.Pool):
         self._out_filepaths.append(filepath)
         return meta
 
-    def _collect_all_pgroups(self, pool: mp.Pool):
+    def _collect_all_pgroups(self, pool: Pool):
         meta = {}
 
         # Initialize a local BLAST DB.