pybiolib 1.2.983__tar.gz → 1.2.998__tar.gz

{pybiolib-1.2.983 → pybiolib-1.2.998}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: pybiolib
-Version: 1.2.983
+Version: 1.2.998
 Summary: BioLib Python Client
 License: MIT
 Keywords: biolib

{pybiolib-1.2.983 → pybiolib-1.2.998}/biolib/_internal/utils/multinode.py

@@ -88,20 +88,21 @@ def fasta_batch_records(records, work_per_batch_min, work_per_residue=1, verbose
     batch = []
     current_longest_seq_len = 0
     for record in records:
-        # Add to batch
-        batch.append(record)
-
         seq_len = len(record.sequence)
-        current_longest_seq_len = max(current_longest_seq_len, seq_len)
+        potential_longest_seq_len = max(current_longest_seq_len, seq_len)
 
-        # Calculate work units based on longest sequence * number of sequences
-        current_work_units = current_longest_seq_len * work_per_residue * len(batch)
+        # Calculate work units if we were to add this record
+        potential_work_units = potential_longest_seq_len * work_per_residue * (len(batch) + 1)
 
-        # If above limit, start a new batch
-        if current_work_units >= work_per_batch_min:
+        if potential_work_units >= work_per_batch_min and len(batch) > 0:
             batches.append(batch)
             batch = []
             current_longest_seq_len = 0
+            potential_longest_seq_len = seq_len
+
+        # Add to batch
+        batch.append(record)
+        current_longest_seq_len = potential_longest_seq_len
 
     # Append last batch if present
     if batch:

{pybiolib-1.2.983 → pybiolib-1.2.998}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "pybiolib"
-version = "1.2.983"
+version = "1.2.998"
 description = "BioLib Python Client"
 readme = "PYPI_README.md"
 license = "MIT"