Update app.py

app.py

@@ -35,14 +35,19 @@ def fetch_metadata(api_key, workspace, project, version):
     return total_images, classes
 
 
-def aggregate_datasets(api_key, urls):
+def aggregate_datasets(api_key, entries):
     """
-    Given API key and list of dataset URLs, returns total images and aggregated class counts.
+    Given API key and list of (url, file_name, line_no) tuples,
+    returns total images and aggregated class counts.
+    Raises ValueError with file and line for invalid URLs.
     """
     total_images = 0
     aggregated = {}
-    for url in urls:
-        ws, proj, ver = parse_roboflow_url(url)
+    for url, file_name, line_no in entries:
+        try:
+            ws, proj, ver = parse_roboflow_url(url)
+        except ValueError:
+            raise ValueError(f"Invalid URL '{url}' in file '{file_name}', line {line_no}")
         imgs, cls_counts = fetch_metadata(api_key, ws, proj, ver)
         total_images += imgs
         for cls, cnt in cls_counts.items():
@@ -65,34 +70,38 @@ def make_bar_chart(counts):
 
 def load_datasets(api_key, file_objs):
     """
-    Read multiple .txt uploads, parse URLs, and aggregate metadata.
+    Read multiple .txt uploads, parse URLs with file/line info,
+    and aggregate metadata. Reports precise errors for invalid URLs.
     Returns total images, dataframe data, plot, and JSON mapping.
     """
-    urls = []
-    # Read each uploaded file from various possible types
+    entries = []
+    seen = set()
+
     for file_obj in file_objs:
-        # Attempt to read bytes
+        # Determine file name
+        file_name = getattr(file_obj, 'name', None) or file_obj.get('name', 'unknown')
+        # Read raw bytes or data
         try:
-            content_bytes = file_obj.read()
+            raw = file_obj.read()
         except Exception:
-            # Try dict-like interface
-            try:
-                content_bytes = file_obj['data']
-            except Exception:
-                # Assume file_obj is file path
-                with open(file_obj, 'rb') as f:
-                    content_bytes = f.read()
-        # Decode if bytes
-        if isinstance(content_bytes, (bytes, bytearray)):
-            content = content_bytes.decode('utf-8')
+            raw = file_obj.get('data') if isinstance(file_obj, dict) else None
+        if raw is None and isinstance(file_obj, str):
+            with open(file_obj, 'rb') as f:
+                raw = f.read()
+        if isinstance(raw, (bytes, bytearray)):
+            content = raw.decode('utf-8')
         else:
-            content = content_bytes
-        # Split lines into URLs
-        lines = [line.strip() for line in content.splitlines() if line.strip()]
-        urls.extend(lines)
-    # Remove duplicate URLs while preserving order
-    unique_urls = list(dict.fromkeys(urls))
-    total_images, aggregated = aggregate_datasets(api_key, unique_urls)
+            content = raw
+        lines = content.splitlines()
+        for i, line in enumerate(lines, start=1):
+            url = line.strip()
+            if not url:
+                continue
+            if url not in seen:
+                seen.add(url)
+                entries.append((url, file_name, i))
+
+    total_images, aggregated = aggregate_datasets(api_key, entries)
     df_data = [[cls, cnt] for cls, cnt in aggregated.items()]
     fig = make_bar_chart(aggregated)
     return f"{total_images}", df_data, fig, json.dumps(aggregated, indent=2)
@@ -131,4 +140,4 @@ def build_ui():
 
 if __name__ == "__main__":
     demo = build_ui()
-    demo.launch(share=True)
+    demo.launch()