Remove json tricks (#908)

examples/pytorch/object-detector/predictor.py

@@ -32,11 +32,11 @@ def predict(self, payload):
         with torch.no_grad():
             pred = self.model(img_tensor)
 
-        predicted_class = [self.coco_labels[i] for i in list(pred[0]["labels"].cpu().numpy())]
+        predicted_class = [self.coco_labels[i] for i in pred[0]["labels"].cpu().tolist()]
         predicted_boxes = [
-            [(i[0], i[1]), (i[2], i[3])] for i in list(pred[0]["boxes"].detach().cpu().numpy())
+            [(i[0], i[1]), (i[2], i[3])] for i in pred[0]["boxes"].detach().cpu().tolist()
         ]
-        predicted_score = list(pred[0]["scores"].detach().cpu().numpy())
+        predicted_score = pred[0]["scores"].detach().cpu().tolist()
         predicted_t = [predicted_score.index(x) for x in predicted_score if x > threshold]
         if len(predicted_t) == 0:
             return [], []

pkg/workloads/cortex/lib/requirements.txt

@@ -6,4 +6,3 @@ pyyaml==5.3
 requests==2.22.0
 
 datadog==0.33.0
-json_tricks==3.13.5

pkg/workloads/cortex/lib/util.py

@@ -20,28 +20,10 @@
 import msgpack
 import pathlib
 import inspect
-import json_tricks
 from inspect import Parameter
 from copy import deepcopy
 
 
-def json_tricks_encoder(*args, **kwargs):
-    kwargs["primitives"] = True
-    kwargs["obj_encoders"] = json_tricks.nonp.DEFAULT_ENCODERS
-    params = list(inspect.signature(json_tricks.TricksEncoder).parameters.values())
-    params += list(inspect.signature(json.JSONEncoder).parameters.values())
-    expected_keys = set()
-    for param in params:
-        if param.kind == Parameter.POSITIONAL_OR_KEYWORD or param.kind == Parameter.KEYWORD_ONLY:
-            expected_keys.add(param.name)
-
-    for key in list(kwargs.keys()):
-        if key not in expected_keys:
-            kwargs.pop(key)
-
-    return json_tricks.TricksEncoder(*args, **kwargs)
-
-
 def extract_zip(zip_path, dest_dir=None, delete_zip_file=False):
     if dest_dir is None:
         dest_dir = os.path.dirname(zip_path)

pkg/workloads/cortex/serve/serve.py

@@ -35,7 +35,7 @@
 from cortex.lib.type import API
 from cortex.lib.log import cx_logger, debug_obj
 from cortex.lib.storage import S3
-
+from cortex.lib.exceptions import UserRuntimeException
 
 if os.environ["CORTEX_VERSION"] != consts.CORTEX_VERSION:
     errMsg = f"your Cortex operator version ({os.environ['CORTEX_VERSION']}) doesn't match your predictor image version ({consts.CORTEX_VERSION}); please update your cluster by following the instructions at https://www.cortex.dev/cluster-management/update, or update your predictor image by modifying the appropriate `image_*` field(s) in your cluster configuration file (e.g. cluster.yaml) and running `cortex cluster update --config cluster.yaml`"
@@ -153,12 +153,16 @@ def predict(request: Any = Body(..., media_type="application/json"), debug=False
 
     debug_obj("payload", request, debug)
     prediction = predictor_impl.predict(request)
-    debug_obj("prediction", prediction, debug)
 
     try:
         json_string = json.dumps(prediction)
-    except:
-        json_string = util.json_tricks_encoder().encode(prediction)
+    except Exception as e:
+        raise UserRuntimeException(
+            f"the return value of predict() or one of its nested values is not JSON serializable",
+            str(e),
+        ) from e
+
+    debug_obj("prediction", json_string, debug)
 
     response = Response(content=json_string, media_type="application/json")