Add a check for consistency

app.py

@@ -107,12 +107,31 @@ def inference(file_name, _clear_punctuation, _show_chars, _batch_mode):
 
     df = pl.read_ndjson(file_name)
 
+    required_columns = [
+        "filename",
+        "inference_start",
+        "inference_end",
+        "inference_total",
+        "duration",
+        "reference",
+        "prediction",
+    ]
+    required_columns_batch = [
+        "inference_start",
+        "inference_end",
+        "inference_total",
+        "filenames",
+        "durations",
+        "references",
+        "predictions",
+    ]
+
     inference_seconds = df["inference_total"].sum()
 
     if _batch_mode:
-        if "durations" not in df.columns:
+        if not all(col in df.columns for col in required_columns_batch):
             raise gr.Error(
-                "Please use a JSONL file with 'durations' column for batch mode."
+                f"Please provide a JSONL file with the following columns: {required_columns_batch}"
             )
 
         duration_seconds = 0
@@ -138,6 +157,11 @@ def inference(file_name, _clear_punctuation, _show_chars, _batch_mode):
         for reference in references_batch:
             references.extend(reference)
     else:
+        if not all(col in df.columns for col in required_columns):
+            raise gr.Error(
+                f"Please provide a JSONL file with the following columns: {required_columns}"
+            )
+
         duration_seconds = df["duration"].sum()
 
         rtf = inference_seconds / duration_seconds