Implement numpy based prediction feature

src/toldg/train.py

@@ -1,9 +1,14 @@
-from toldg.models import Config, CsvConfig, Mapping, Transaction
+import logging
+from pathlib import Path
+
+from toldg.models import Config
+from toldg.predict import train_classifier
 from toldg.process import process_csv_files
 
 
 def train(config: Config):
-    print("[train] start")
+    """Train a transaction classifier from csv files."""
+    # Process transactions to get training data
     transactions = process_csv_files(config)
-    for t in transactions:
-        pass
+    train_classifier(transactions, config.categories, config.model)
+    logging.info("Training completed")