Keep parquet data in hexadecimal format

src/parser.py

@@ -63,7 +63,7 @@ def parse_log(input_path: PathLike, out_bus1: PathLike, out_bus2: PathLike) -> N
 
 def parse_csv(csv_path: PathLike) -> pl.LazyFrame:
     """
-    Ingests a parsed CSV file, unpacks hex strings into 8 integer columns,
+    Ingests a parsed CSV file, unpacks hex strings into 8 hex columns,
     generates sequential timestamps if missing, and returns a Polars LazyFrame.
     """
     lf = pl.scan_csv(csv_path, schema_overrides={"ID": pl.String, "Data": pl.String})
@@ -76,12 +76,11 @@ def parse_csv(csv_path: PathLike) -> pl.LazyFrame:
         expr = (
             pl.col("Data").str.strip_chars().str.split(" ")
             .list.get(i, null_on_oob=True)
-            .str.to_integer(base=16, strict=False)
-            .alias(f"b{i}").cast(pl.UInt8)
+            .alias(f"b{i}")
         )
         byte_exprs.append(expr)
 
-    lf = lf.with_columns(byte_exprs).drop("Data")
+    lf = lf.with_columns(byte_exprs)
 
     return lf.with_columns([
         pl.col("DLC").cast(pl.UInt8),
@@ -112,5 +111,5 @@ if __name__ == '__main__':
     elif args.command == "parquet":
         print(f"[*] Processing {args.input_csv}...")
         lf = parse_csv(args.input_csv)
-        save_parquet(lf, args.output_parquet)
+        lf.sink_parquet(args.output_parquet)
         print(f"[+] Saved parquet file to {args.output_parquet}")