API: Add generator error handling

If the generator errors, there's no proper handling to send an error
packet and close the connection.

This is especially important for unloading models if the load fails
at any stage to reclaim a user's VRAM. Raising an exception caused
the model_container object to lock and not get freed by the GC.

This made sense to propegate SSE errors across all generator functions
rather than relying on abort signals.

Signed-off-by: kingbri <bdashore3@proton.me>

utils.py

@@ -1,3 +1,29 @@
+import traceback
+from pydantic import BaseModel
+from typing import Optional
+
 # Wrapper callback for load progress
 def load_progress(module, modules):
-    yield module, modules
+    yield module, modules
+
+# Common error types
+class TabbyGeneratorErrorMessage(BaseModel):
+    message: str
+    trace: Optional[str] = None
+
+class TabbyGeneratorError(BaseModel):
+    error: TabbyGeneratorErrorMessage
+
+def get_generator_error(exception: Exception):
+    error_message = TabbyGeneratorErrorMessage(
+        message = str(exception),
+        trace = traceback.format_exc()
+    )
+
+    generator_error = TabbyGeneratorError(
+        error = error_message
+    )
+
+    # Log and send the exception
+    print(f"\n{generator_error.error.trace}")
+    return generator_error.json()