Upload app.py with huggingface_hub

app.py

@@ -109,28 +109,44 @@ def generate_streaming(messages, max_new_tokens, max_denoising_steps, enable_thi
     yield build_highlight(final_text, ""), final_text, final_text
 
 
+def _file_path(item):
+    """Extract a local file path from a Gradio content part / file dict."""
+    for key in ("file", "path", "url"):
+        val = item.get(key)
+        if isinstance(val, dict):
+            val = val.get("path") or val.get("url")
+        if val:
+            return val
+    return None
+
+
 def to_model_messages(history):
-    """Convert Gradio messages history into processor chat format with images."""
+    """Convert Gradio (messages format) history into processor chat format with images."""
     messages = []
     for msg in history:
         role = msg["role"]
         content = msg["content"]
+        parts = []
         if isinstance(content, str):
-            parts = [{"type": "text", "text": content}]
-        elif isinstance(content, (list, tuple)):
-            # Gradio file tuple (path, alt) or list of content dicts
-            parts = []
-            if isinstance(content, tuple):
-                parts.append({"type": "image", "url": content[0]})
-            else:
-                for item in content:
-                    if isinstance(item, dict) and item.get("type") == "text":
-                        parts.append({"type": "text", "text": item["text"]})
-                    elif isinstance(item, dict) and "path" in item:
-                        parts.append({"type": "image", "url": item["path"]})
-        else:
-            parts = [{"type": "text", "text": str(content)}]
-        messages.append({"role": role, "content": parts})
+            parts.append({"type": "text", "text": content})
+        elif isinstance(content, tuple):  # legacy (path, alt) file tuple
+            parts.append({"type": "image", "url": content[0]})
+        elif isinstance(content, dict):  # single file part, e.g. {"path": ...}
+            p = _file_path(content)
+            if p:
+                parts.append({"type": "image", "url": p})
+        elif isinstance(content, list):
+            for item in content:
+                if not isinstance(item, dict):
+                    parts.append({"type": "text", "text": str(item)})
+                elif item.get("type") == "text" or "text" in item:
+                    parts.append({"type": "text", "text": item.get("text", "")})
+                else:
+                    p = _file_path(item)
+                    if p:
+                        parts.append({"type": "image", "url": p})
+        if parts:
+            messages.append({"role": role, "content": parts})
     return messages