haystack-ml-stack 0.3.2__tar.gz → 0.3.4__tar.gz

{haystack_ml_stack-0.3.2 → haystack_ml_stack-0.3.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: haystack-ml-stack
-Version: 0.3.2
+Version: 0.3.4
 Summary: Functions related to Haystack ML
 Author-email: Oscar Vega <oscar@haystack.tv>
 License: MIT

{haystack_ml_stack-0.3.2 → haystack_ml_stack-0.3.4}/pyproject.toml

@@ -5,7 +5,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "haystack-ml-stack"
-version = "0.3.2"
+version = "0.3.4"
 description = "Functions related to Haystack ML"
 readme = "README.md"
 authors = [{ name = "Oscar Vega", email = "oscar@haystack.tv" }]

{haystack_ml_stack-0.3.2 → haystack_ml_stack-0.3.4}/src/haystack_ml_stack/__init__.py

@@ -11,4 +11,4 @@ from ._serializers import SerializerRegistry, FeatureRegistryId
 
 __all__ = [*__all__, "SerializerRegistry", "FeatureRegistryId"]
 
-__version__ = "0.3.2"
+__version__ = "0.3.4"

{haystack_ml_stack-0.3.2 → haystack_ml_stack-0.3.4}/src/haystack_ml_stack/app.py

@@ -6,6 +6,8 @@ from http import HTTPStatus
 from typing import Any, Dict, List, Optional
 import time
 from contextlib import asynccontextmanager, AsyncExitStack
+import traceback
+import json
 
 import aiobotocore.session
 from aiobotocore.config import AioConfig
@@ -160,6 +162,7 @@ def create_app(
             "user_cache_size": len(user_features_cache),
             "model_name": state.get("model_name"),
             "stream_features": state.get("stream_features", []),
+            "user_features": state.get("user_features", []),
         }
 
     @app.post("/score", status_code=HTTPStatus.OK)
@@ -172,10 +175,23 @@ def create_app(
 
         try:
             data = await request.json()
-        except Exception as e:
+        except json.JSONDecodeError as e:
+            body = await request.body()
+            logger.error(
+                "Received malformed json. Raw body: %s\n%s",
+                body.decode(errors="replace"),
+                traceback.format_exc(),
+            )
             raise HTTPException(
                 status_code=HTTPStatus.BAD_REQUEST, detail="Invalid JSON payload"
             ) from e
+        except Exception as e:
+            logger.error(
+                "Unexpected exception when parsing request.\n %s", traceback.format_exc()
+            )
+            raise HTTPException(
+                status_code=HTTPStatus.INTERNAL_SERVER_ERROR, detail="Unknown exception"
+            ) from e
         query_params = {}
         for k in request.query_params.keys():
             values = request.query_params.getlist(k)
@@ -253,7 +269,7 @@ def create_app(
             model_output = model["predict"](model_input, model["params"])
             predict_end = time.perf_counter_ns()
         except Exception as e:
-            logger.error("Model prediction failed: %s", e)
+            logger.error("Model prediction failed: \n%s", traceback.format_exc())
             raise HTTPException(
                 status_code=HTTPStatus.INTERNAL_SERVER_ERROR,
                 detail="Model prediction failed",

{haystack_ml_stack-0.3.2 → haystack_ml_stack-0.3.4}/src/haystack_ml_stack/utils.py

@@ -204,7 +204,7 @@ def device_watched_count_cleanups(
 
     counts_obj: StreamPWatched = stream.get(
         f"PWATCHED#24H#{device_type}", StreamPWatched()
-    )
+    ).data
     if out is None:
         out = {}
     out = _cleanup_entry_context_counts(
@@ -336,7 +336,7 @@ def user_pwatched_cleanup(
             "launch_first_in_session",
         ]
     _validate_pwatched_entry_context(entry_contexts)
-    counts_obj = user.get("PWATCHED#6M", UserPWatched())
+    counts_obj = user.get("PWATCHED#6M", UserPWatched()).data
     out = _cleanup_entry_context_counts(
         counts_obj=counts_obj,
         entry_contexts=entry_contexts,

{haystack_ml_stack-0.3.2 → haystack_ml_stack-0.3.4}/src/haystack_ml_stack.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: haystack-ml-stack
-Version: 0.3.2
+Version: 0.3.4
 Summary: Functions related to Haystack ML
 Author-email: Oscar Vega <oscar@haystack.tv>
 License: MIT

{haystack_ml_stack-0.3.2 → haystack_ml_stack-0.3.4}/tests/test_utils.py

@@ -372,6 +372,58 @@ def test_stream_pwatched_cleanup():
     assert (actual == expected).all()
 
 
+def test_device_stream_pwatched_cleanup():
+    stream_pwatched_data = {
+        "version": 1,
+        "data": {
+            "autoplay": {"attempts": 1, "watched": 1},
+            "ch_swtch": {"attempts": 2, "watched": 0},
+        },
+    }
+    stream_pwatched = features_pb2_v1.StreamPWatched()
+    ProtoParseDict(js_dict=stream_pwatched_data, message=stream_pwatched)
+    stream = {"PWATCHED#24H#TV": stream_pwatched}
+    out = {}
+    utils.device_watched_count_cleanups(
+        stream=stream,
+        entry_contexts=["autoplay", "ch_swtch", "sel_thumb"],
+        device_type="TV",
+        out=out,
+    )
+    expected = pd.Series(
+        {
+            "STREAM_AUTOPLAY_TV_24H_TOTAL_ATTEMPTS": 1,
+            "STREAM_AUTOPLAY_TV_24H_TOTAL_WATCHED": 1,
+            "STREAM_CH_SWTCH_TV_24H_TOTAL_ATTEMPTS": 2,
+            "STREAM_CH_SWTCH_TV_24H_TOTAL_WATCHED": 0,
+            "STREAM_SEL_THUMB_TV_24H_TOTAL_ATTEMPTS": 0,
+            "STREAM_SEL_THUMB_TV_24H_TOTAL_WATCHED": 0,
+        }
+    )
+    actual = pd.Series(out).loc[expected.index]
+    assert (actual == expected).all()
+    stream = {"PWATCHED#24H#MOBILE": stream_pwatched}
+    out = {}
+    utils.device_watched_count_cleanups(
+        stream=stream,
+        entry_contexts=["autoplay", "ch_swtch", "sel_thumb"],
+        device_type="MOBILE",
+        out=out,
+    )
+    expected = pd.Series(
+        {
+            "STREAM_AUTOPLAY_MOBILE_24H_TOTAL_ATTEMPTS": 1,
+            "STREAM_AUTOPLAY_MOBILE_24H_TOTAL_WATCHED": 1,
+            "STREAM_CH_SWTCH_MOBILE_24H_TOTAL_ATTEMPTS": 2,
+            "STREAM_CH_SWTCH_MOBILE_24H_TOTAL_WATCHED": 0,
+            "STREAM_SEL_THUMB_MOBILE_24H_TOTAL_ATTEMPTS": 0,
+            "STREAM_SEL_THUMB_MOBILE_24H_TOTAL_WATCHED": 0,
+        }
+    )
+    actual = pd.Series(out).loc[expected.index]
+    assert (actual == expected).all()
+
+
 def test_stream_global_pselect_cleanup():
     stream_pselect_data = {
         "version": 1,
@@ -508,3 +560,78 @@ def test_stream_similarity_top_category_functions():
     assert all(
         actual_key == expected_key for actual_key, expected_key in zip(actual, expected)
     )
+
+
+def test_user_pwatched_cleanup():
+    user_pwatched_data = {
+        "version": 1,
+        "data": {
+            "sel_thumb": {"attempts": 1, "watched": 1},
+            "ch_swtch": {"attempts": 2, "watched": 0},
+        },
+    }
+    user_pwatched_msg = features_pb2_v1.UserPWatched()
+    ProtoParseDict(js_dict=user_pwatched_data, message=user_pwatched_msg)
+    user = {"PWATCHED#6M": user_pwatched_msg}
+    out = {}
+    utils.user_pwatched_cleanup(
+        user=user, entry_contexts=["autoplay", "sel_thumb", "ch_swtch"], out=out
+    )
+    expected = pd.Series(
+        {
+            "USER_AUTOPLAY_6M_TOTAL_ATTEMPTS": 0,
+            "USER_AUTOPLAY_6M_TOTAL_WATCHED": 0,
+            "USER_SEL_THUMB_6M_TOTAL_ATTEMPTS": 1,
+            "USER_SEL_THUMB_6M_TOTAL_WATCHED": 1,
+            "USER_CH_SWTCH_6M_TOTAL_ATTEMPTS": 2,
+            "USER_CH_SWTCH_6M_TOTAL_WATCHED": 0,
+        }
+    )
+    actual = pd.Series(out).loc[expected.index]
+    assert (expected == actual).all()
+
+
+def test_user_pselect_cleanup():
+    user_pselect_data = {
+        "version": 1,
+        "data": {
+            "all_browsed": {
+                "first_pos": {
+                    "total_selects": 0,
+                    "total_selects_and_watched": 0,
+                    "total_browsed": 1,
+                },
+                "rest_pos": {
+                    "total_selects": 2,
+                    "total_selects_and_watched": 2,
+                    "total_browsed": 1,
+                },
+            },
+            "up_to_4_browsed": {
+                "first_pos": {
+                    "total_selects": 0,
+                    "total_selects_and_watched": 0,
+                    "total_browsed": 1,
+                },
+                "rest_pos": {
+                    "total_selects": 2,
+                    "total_selects_and_watched": 2,
+                    "total_browsed": 0,
+                },
+            },
+        },
+    }
+    user_pselect_msg = features_pb2_v1.UserPSelect()
+    ProtoParseDict(js_dict=user_pselect_data, message=user_pselect_msg)
+    user = {"PSELECT#6M": user_pselect_msg}
+    out = {}
+    utils.user_pselect_cleanup(user=user, position_debiasing="up_to_4_browsed", out=out)
+    expected = pd.Series(
+        {
+            "USER_6M_TOTAL_BROWSED_UP_TO_4_BROWSED": 1,
+            "USER_6M_TOTAL_SELECTS_UP_TO_4_BROWSED": 2,
+            "USER_6M_TOTAL_SELECTS_AND_WATCHED_UP_TO_4_BROWSED": 2,
+        }
+    )
+    actual = pd.Series(out).loc[expected.index]
+    assert (actual == expected).all()