fix: pass file_path not DataFrame to detect_or_get_provider

Author: beatfactor

ingest/adapter.py

@@ -462,7 +462,9 @@ def _peek_headers(file_path: Path) -> list[str]:
         return []
 
 
-def enrich_with_provider(df: pd.DataFrame, provider_name: str) -> pd.DataFrame:
+def enrich_with_provider(
+    df: pd.DataFrame, provider_name: str, file_path: "Path | None" = None
+) -> pd.DataFrame:
     """Optionally enrich a DataFrame using an oceanstream provider.
 
     Parameters
@@ -471,6 +473,9 @@ def enrich_with_provider(df: pd.DataFrame, provider_name: str) -> pd.DataFrame:
         Data to enrich.
     provider_name : str
         Provider name or ``"auto"`` for auto-detection.
+    file_path : Path or None
+        Source file path, used for provider auto-detection when *provider_name*
+        is ``"auto"``.
 
     Returns
     -------
@@ -485,7 +490,8 @@ def enrich_with_provider(df: pd.DataFrame, provider_name: str) -> pd.DataFrame:
         from oceanstream.providers.factory import detect_or_get_provider
 
         provider = detect_or_get_provider(
-            provider_name if provider_name != "auto" else None, df
+            provider_name if provider_name != "auto" else None,
+            file_path=file_path,
         )
         if provider is not None:
             df = provider.enrich_dataframe(df)

process/pipeline.py

@@ -127,9 +127,11 @@ def _extract_archive(archive_path: Path) -> tuple[list[Path], str]:
     return extracted, tmpdir
 
 
-def _enrich_with_provider(df: pd.DataFrame, provider_name: str) -> pd.DataFrame:
+def _enrich_with_provider(
+    df: pd.DataFrame, provider_name: str, file_path: "Path | None" = None
+) -> pd.DataFrame:
     """Optionally enrich a DataFrame using an oceanstream provider."""
-    return enrich_with_provider(df, provider_name)
+    return enrich_with_provider(df, provider_name, file_path=file_path)
 
 
 def _make_output_path(config: "EdgeConfig", source_name: str, suffix: str = ".parquet") -> str:
@@ -211,7 +213,7 @@ async def process_file(
             return result
 
         # Provider enrichment
-        df = _enrich_with_provider(df, config.provider)
+        df = _enrich_with_provider(df, config.provider, file_path=path)
 
         # Deduplicate (skip for ADCP — multiple depth cells per timestamp)
         if "time" in df.columns and file_type != "adcp":