catch request errors

app.py

@@ -71,7 +71,6 @@ GEN_CHAINS = [
 
 INFERENCE_PARAMS = {
     'max_length':450,
-#    'clean_up_tokenization_spaces': False,
     'use_cache':False
 }
 
@@ -91,11 +90,21 @@ def build_inference_api():
     headers = {}# {"Authorization": f"Bearer {st.secrets['api_token']}"}
 
     def query(inputs: str, parameters):
-        payload = {"inputs": inputs}
-        payload.update(parameters)
+        payload = {
+            "inputs": inputs,
+            "parameters": parameters,
+            "options": {"wait_for_model": True},
+        }
         data = json.dumps(payload)
         response = requests.request("POST", API_URL, headers=headers, data=data)
-        return json.loads(response.content.decode("utf-8"))
+        content = response.content.decode("utf-8")
+        try:
+            # as json 
+            result_json = json.loads(content) 
+        except Exception:
+            result_json = {"error": content}
+
+        return result_json
 
     return query
 
@@ -276,16 +285,24 @@ def run_model(mode_set, user_input):
             inputs = inquire_prompt
             attempts = 0
             out = None
-            while out==None and attempts<MAX_API_CALLS: 
+            while not out and attempts<MAX_API_CALLS:
                 attempts += 1
                 try:
-                    # api call 
-                    out = inference(inputs,INFERENCE_PARAMS) 
+                    # api call
+                    out = inference(inputs, params)
+                    if not isinstance(out, list):
+                        raise ValueError('Response is not a list.')
                 except Exception:
-                    if attempts<MAX_API_CALLS:
-                        st.warning(f"HF Inference API call (attempt {attempts} of {MAX_API_CALLS}) has failed. Trying again...")
+                    if attempts < MAX_API_CALLS:
+                        st.warning(
+                            f"HF Inference API call (attempt {attempts} of {MAX_API_CALLS}) has failed. Response: {out}. Trying again..."
+                        )
+                        out = None
                     else:
-                        raise
+                        st.warning(
+                            f"HF Inference API call (attempt {attempts} of {MAX_API_CALLS}) has failed. Response: {out}. Stopping."
+                        )
+                        return None
             out = out[0]['generated_text']
             # cleanup formalization
             if to_key in ['premises_formalized','conclusion_formalized']:
@@ -390,6 +407,9 @@ def main():
             with st.spinner("Processing (may take a couple of minutes) ..."):
                 output = run_model(modes_s,user_input)
 
+                if output is None:
+                    return None
+
                 # get latest generated reasons, conclusions, argdown 
                 argdown_raw = [out['output'] for out in output if out['mode']['to']=='argdown_reconstruction']
                 argdown_raw = argdown_raw[-1] if len(argdown_raw)>0 else None