haystack-ml-stack 0.4.5a3__tar.gz → 0.4.7__tar.gz

{haystack_ml_stack-0.4.5a3 → haystack_ml_stack-0.4.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: haystack-ml-stack
-Version: 0.4.5a3
+Version: 0.4.7
 Summary: Functions related to Haystack ML
 Author-email: Oscar Vega <oscar@haystack.tv>
 License: MIT

{haystack_ml_stack-0.4.5a3 → haystack_ml_stack-0.4.7}/pyproject.toml

@@ -5,7 +5,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "haystack-ml-stack"
-version = "0.4.5a3"
+version = "0.4.7"
 description = "Functions related to Haystack ML"
 readme = "README.md"
 authors = [{ name = "Oscar Vega", email = "oscar@haystack.tv" }]

{haystack_ml_stack-0.4.5a3 → haystack_ml_stack-0.4.7}/src/haystack_ml_stack/_serializers.py

@@ -422,6 +422,17 @@ global_playlist_stats_v1_features: list[FeatureRegistryId] = [
         feature_id="PLAYLIST_CATEGORY_STATS#1D#TV",
         version="v1",
     ),
+    # Same feature structure but with an updated filter for the source query
+    FeatureRegistryId(
+        entity_type="GLOBAL",
+        feature_id="PLAYLIST_CATEGORY_STATS_V2#1D#MOBILE",
+        version="v1",
+    ),
+    FeatureRegistryId(
+        entity_type="GLOBAL",
+        feature_id="PLAYLIST_CATEGORY_STATS_V2#1D#TV",
+        version="v1",
+    ),
 ]
 
 user_playlist_stats_v1_features: list[FeatureRegistryId] = [

haystack_ml_stack-0.4.7/src/haystack_ml_stack/_version.py

@@ -0,0 +1 @@
+__version__ = "0.4.7"

{haystack_ml_stack-0.4.5a3 → haystack_ml_stack-0.4.7}/src/haystack_ml_stack/dynamo.py

@@ -9,7 +9,7 @@ from boto3.dynamodb.types import TypeDeserializer
 
 from . import exceptions
 from ._serializers import FeatureRegistryId, SerializerRegistry
-from .utils import _complete_features_for_channels
+from .utils import _complete_features_for_channels, DEFAULT_CHANNELS
 
 logger = logging.getLogger(__name__)
 
@@ -274,7 +274,7 @@ _MOBILE_OS = {"ios", "android", "iphone", "galaxy"}
 
 
 def _get_os_cat(client_os: str) -> str:
-    normalized = client_os.lower().replace("debug","") if client_os else ""
+    normalized = client_os.lower().replace("debug", "") if client_os else ""
     return "MOBILE" if normalized in _MOBILE_OS else "TV"
 
 
@@ -503,16 +503,32 @@ def _process_channels(
         for stat in playlist_stats_user.values():
             all_channels.update(stat.data.keys())
 
-    # Get not preferred channels to be ignored
+    all_channels = all_channels | set(DEFAULT_CHANNELS)
+
+    # My headlines will stay on top so not part of the candidates
     # The group labels are also ignored, not real channels for UI
-    ignore_channels = set(["national_favorite", "local_favorite"])
+    ignore_channels = set(["my headlines", "national_favorite", "local_favorite"])
+
+    # Rule for weather
+    if user.get("geo", {}).get("country") != "US":
+        ignore_channels.add("weather")
+    # Rule for live_es
+    if "es" not in user.get("languages", []):
+        ignore_channels.add("live_es")
+
     inserted_channel_names = set()
     if channel_candidates is not None:
         preferred = set(user.get("preferredChannels", []) or [])
         for ch in channel_candidates.data:
-            if ch.category_group in ("national_favorite", "local_favorite"):
-                if ch.name not in preferred:
-                    ignore_channels.add(ch.name)
+            # Rule for authors
+            if (
+                ch.category_group in ("national_favorite", "local_favorite")
+                and ch.name not in preferred
+            ):
+                ignore_channels.add(ch.name)
+            # Rule for shows
+            if ch.category_group == "show" and ch.name.split("/")[0] not in preferred:
+                ignore_channels.add(ch.name)
 
         for ch in channel_candidates.data:
             if ch.name not in ignore_channels:
@@ -525,21 +541,7 @@ def _process_channels(
                 )
                 inserted_channel_names.add(ch.name)
 
-    DEFAULT_CHANNELS = [
-        "local news",
-        "science & technology",
-        "business & finance",
-        "entertainment news",
-        "live",
-        "live_es",
-        "weather",
-        "politics",
-        "international",
-        "top videos",
-        "editor picks",
-    ]
-
-    for name in all_channels | set(DEFAULT_CHANNELS):
+    for name in all_channels:
         if name not in ignore_channels and name not in inserted_channel_names:
             channels.append(
                 {

{haystack_ml_stack-0.4.5a3 → haystack_ml_stack-0.4.7}/src/haystack_ml_stack/generated/v1/features_pb2.py

@@ -2,7 +2,7 @@
 # Generated by the protocol buffer compiler.  DO NOT EDIT!
 # NO CHECKED-IN PROTOBUF GENCODE
 # source: features.proto
-# Protobuf Python Version: 6.33.2
+# Protobuf Python Version: 7.34.1
 """Generated protocol buffer code."""
 from google.protobuf import descriptor as _descriptor
 from google.protobuf import descriptor_pool as _descriptor_pool
@@ -11,9 +11,9 @@ from google.protobuf import symbol_database as _symbol_database
 from google.protobuf.internal import builder as _builder
 _runtime_version.ValidateProtobufRuntimeVersion(
     _runtime_version.Domain.PUBLIC,
-    6,
-    33,
-    2,
+    7,
+    34,
+    1,
     '',
     'features.proto'
 )

{haystack_ml_stack-0.4.5a3 → haystack_ml_stack-0.4.7}/src/haystack_ml_stack/generated/v1/features_pb2.pyi

@@ -7,7 +7,7 @@ from typing import ClassVar as _ClassVar, Optional as _Optional, Union as _Union
 DESCRIPTOR: _descriptor.FileDescriptor
 
 class EntryContextCounts(_message.Message):
-    __slots__ = ()
+    __slots__ = ("attempts", "watched")
     ATTEMPTS_FIELD_NUMBER: _ClassVar[int]
     WATCHED_FIELD_NUMBER: _ClassVar[int]
     attempts: int
@@ -15,7 +15,7 @@ class EntryContextCounts(_message.Message):
     def __init__(self, attempts: _Optional[int] = ..., watched: _Optional[int] = ...) -> None: ...
 
 class SelectCounts(_message.Message):
-    __slots__ = ()
+    __slots__ = ("total_selects", "total_selects_and_watched", "total_browsed")
     TOTAL_SELECTS_FIELD_NUMBER: _ClassVar[int]
     TOTAL_SELECTS_AND_WATCHED_FIELD_NUMBER: _ClassVar[int]
     TOTAL_BROWSED_FIELD_NUMBER: _ClassVar[int]
@@ -25,7 +25,7 @@ class SelectCounts(_message.Message):
     def __init__(self, total_selects: _Optional[int] = ..., total_selects_and_watched: _Optional[int] = ..., total_browsed: _Optional[int] = ...) -> None: ...
 
 class EntryContextPWatched(_message.Message):
-    __slots__ = ()
+    __slots__ = ("autoplay", "sel_thumb", "choose_next", "ch_swtch", "launch_first_in_session")
     AUTOPLAY_FIELD_NUMBER: _ClassVar[int]
     SEL_THUMB_FIELD_NUMBER: _ClassVar[int]
     CHOOSE_NEXT_FIELD_NUMBER: _ClassVar[int]
@@ -39,7 +39,7 @@ class EntryContextPWatched(_message.Message):
     def __init__(self, autoplay: _Optional[_Union[EntryContextCounts, _Mapping]] = ..., sel_thumb: _Optional[_Union[EntryContextCounts, _Mapping]] = ..., choose_next: _Optional[_Union[EntryContextCounts, _Mapping]] = ..., ch_swtch: _Optional[_Union[EntryContextCounts, _Mapping]] = ..., launch_first_in_session: _Optional[_Union[EntryContextCounts, _Mapping]] = ...) -> None: ...
 
 class PositionPSelect(_message.Message):
-    __slots__ = ()
+    __slots__ = ("first_pos", "second_pos", "third_pos", "rest_pos")
     FIRST_POS_FIELD_NUMBER: _ClassVar[int]
     SECOND_POS_FIELD_NUMBER: _ClassVar[int]
     THIRD_POS_FIELD_NUMBER: _ClassVar[int]
@@ -51,7 +51,7 @@ class PositionPSelect(_message.Message):
     def __init__(self, first_pos: _Optional[_Union[SelectCounts, _Mapping]] = ..., second_pos: _Optional[_Union[SelectCounts, _Mapping]] = ..., third_pos: _Optional[_Union[SelectCounts, _Mapping]] = ..., rest_pos: _Optional[_Union[SelectCounts, _Mapping]] = ...) -> None: ...
 
 class BrowsedDebiasedPositionPSelects(_message.Message):
-    __slots__ = ()
+    __slots__ = ("up_to_4_browsed", "all_browsed")
     UP_TO_4_BROWSED_FIELD_NUMBER: _ClassVar[int]
     ALL_BROWSED_FIELD_NUMBER: _ClassVar[int]
     up_to_4_browsed: PositionPSelect
@@ -59,7 +59,7 @@ class BrowsedDebiasedPositionPSelects(_message.Message):
     def __init__(self, up_to_4_browsed: _Optional[_Union[PositionPSelect, _Mapping]] = ..., all_browsed: _Optional[_Union[PositionPSelect, _Mapping]] = ...) -> None: ...
 
 class PlaylistStatsForGlobal(_message.Message):
-    __slots__ = ()
+    __slots__ = ("watched_count", "not_watched_count", "capped_watched_secs", "capped_not_watched_secs", "watched_secs", "not_watched_secs")
     WATCHED_COUNT_FIELD_NUMBER: _ClassVar[int]
     NOT_WATCHED_COUNT_FIELD_NUMBER: _ClassVar[int]
     CAPPED_WATCHED_SECS_FIELD_NUMBER: _ClassVar[int]
@@ -75,7 +75,7 @@ class PlaylistStatsForGlobal(_message.Message):
     def __init__(self, watched_count: _Optional[int] = ..., not_watched_count: _Optional[int] = ..., capped_watched_secs: _Optional[float] = ..., capped_not_watched_secs: _Optional[float] = ..., watched_secs: _Optional[float] = ..., not_watched_secs: _Optional[float] = ...) -> None: ...
 
 class PlaylistStatsForUser(_message.Message):
-    __slots__ = ()
+    __slots__ = ("total_days", "start_days", "active_days", "total_watched", "capped_total_watched")
     TOTAL_DAYS_FIELD_NUMBER: _ClassVar[int]
     START_DAYS_FIELD_NUMBER: _ClassVar[int]
     ACTIVE_DAYS_FIELD_NUMBER: _ClassVar[int]
@@ -89,7 +89,7 @@ class PlaylistStatsForUser(_message.Message):
     def __init__(self, total_days: _Optional[int] = ..., start_days: _Optional[int] = ..., active_days: _Optional[int] = ..., total_watched: _Optional[float] = ..., capped_total_watched: _Optional[float] = ...) -> None: ...
 
 class Channel(_message.Message):
-    __slots__ = ()
+    __slots__ = ("name", "category_group", "start_date")
     NAME_FIELD_NUMBER: _ClassVar[int]
     CATEGORY_GROUP_FIELD_NUMBER: _ClassVar[int]
     START_DATE_FIELD_NUMBER: _ClassVar[int]
@@ -99,7 +99,7 @@ class Channel(_message.Message):
     def __init__(self, name: _Optional[str] = ..., category_group: _Optional[str] = ..., start_date: _Optional[int] = ...) -> None: ...
 
 class StreamPSelect(_message.Message):
-    __slots__ = ()
+    __slots__ = ("version", "data")
     VERSION_FIELD_NUMBER: _ClassVar[int]
     DATA_FIELD_NUMBER: _ClassVar[int]
     version: int
@@ -107,7 +107,7 @@ class StreamPSelect(_message.Message):
     def __init__(self, version: _Optional[int] = ..., data: _Optional[_Union[BrowsedDebiasedPositionPSelects, _Mapping]] = ...) -> None: ...
 
 class StreamPWatched(_message.Message):
-    __slots__ = ()
+    __slots__ = ("version", "data")
     VERSION_FIELD_NUMBER: _ClassVar[int]
     DATA_FIELD_NUMBER: _ClassVar[int]
     version: int
@@ -115,7 +115,7 @@ class StreamPWatched(_message.Message):
     def __init__(self, version: _Optional[int] = ..., data: _Optional[_Union[EntryContextPWatched, _Mapping]] = ...) -> None: ...
 
 class UserPWatched(_message.Message):
-    __slots__ = ()
+    __slots__ = ("version", "data")
     VERSION_FIELD_NUMBER: _ClassVar[int]
     DATA_FIELD_NUMBER: _ClassVar[int]
     version: int
@@ -123,9 +123,9 @@ class UserPWatched(_message.Message):
     def __init__(self, version: _Optional[int] = ..., data: _Optional[_Union[EntryContextPWatched, _Mapping]] = ...) -> None: ...
 
 class UserPersonalizingPWatched(_message.Message):
-    __slots__ = ()
+    __slots__ = ("version", "data")
     class DataEntry(_message.Message):
-        __slots__ = ()
+        __slots__ = ("key", "value")
         KEY_FIELD_NUMBER: _ClassVar[int]
         VALUE_FIELD_NUMBER: _ClassVar[int]
         key: str
@@ -138,7 +138,7 @@ class UserPersonalizingPWatched(_message.Message):
     def __init__(self, version: _Optional[int] = ..., data: _Optional[_Mapping[str, EntryContextPWatched]] = ...) -> None: ...
 
 class UserPSelect(_message.Message):
-    __slots__ = ()
+    __slots__ = ("version", "data")
     VERSION_FIELD_NUMBER: _ClassVar[int]
     DATA_FIELD_NUMBER: _ClassVar[int]
     version: int
@@ -146,9 +146,9 @@ class UserPSelect(_message.Message):
     def __init__(self, version: _Optional[int] = ..., data: _Optional[_Union[BrowsedDebiasedPositionPSelects, _Mapping]] = ...) -> None: ...
 
 class UserPersonalizingPSelect(_message.Message):
-    __slots__ = ()
+    __slots__ = ("version", "data")
     class DataEntry(_message.Message):
-        __slots__ = ()
+        __slots__ = ("key", "value")
         KEY_FIELD_NUMBER: _ClassVar[int]
         VALUE_FIELD_NUMBER: _ClassVar[int]
         key: str
@@ -161,9 +161,9 @@ class UserPersonalizingPSelect(_message.Message):
     def __init__(self, version: _Optional[int] = ..., data: _Optional[_Mapping[str, BrowsedDebiasedPositionPSelects]] = ...) -> None: ...
 
 class StreamSimilarityScores(_message.Message):
-    __slots__ = ()
+    __slots__ = ("version", "data")
     class DataEntry(_message.Message):
-        __slots__ = ()
+        __slots__ = ("key", "value")
         KEY_FIELD_NUMBER: _ClassVar[int]
         VALUE_FIELD_NUMBER: _ClassVar[int]
         key: str
@@ -176,9 +176,9 @@ class StreamSimilarityScores(_message.Message):
     def __init__(self, version: _Optional[int] = ..., data: _Optional[_Mapping[str, float]] = ...) -> None: ...
 
 class GlobalPlaylistStats(_message.Message):
-    __slots__ = ()
+    __slots__ = ("version", "data")
     class DataEntry(_message.Message):
-        __slots__ = ()
+        __slots__ = ("key", "value")
         KEY_FIELD_NUMBER: _ClassVar[int]
         VALUE_FIELD_NUMBER: _ClassVar[int]
         key: str
@@ -191,9 +191,9 @@ class GlobalPlaylistStats(_message.Message):
     def __init__(self, version: _Optional[int] = ..., data: _Optional[_Mapping[str, PlaylistStatsForGlobal]] = ...) -> None: ...
 
 class UserPlaylistStats(_message.Message):
-    __slots__ = ()
+    __slots__ = ("version", "data")
     class DataEntry(_message.Message):
-        __slots__ = ()
+        __slots__ = ("key", "value")
         KEY_FIELD_NUMBER: _ClassVar[int]
         VALUE_FIELD_NUMBER: _ClassVar[int]
         key: str
@@ -206,7 +206,7 @@ class UserPlaylistStats(_message.Message):
     def __init__(self, version: _Optional[int] = ..., data: _Optional[_Mapping[str, PlaylistStatsForUser]] = ...) -> None: ...
 
 class GlobalChannels(_message.Message):
-    __slots__ = ()
+    __slots__ = ("version", "data")
     VERSION_FIELD_NUMBER: _ClassVar[int]
     DATA_FIELD_NUMBER: _ClassVar[int]
     version: int

{haystack_ml_stack-0.4.5a3 → haystack_ml_stack-0.4.7}/src/haystack_ml_stack/utils.py

@@ -679,6 +679,21 @@ def _validate_pwatched_entry_context(entry_contexts: list[str]):
         raise ValueError(f"Invalid entry contexts found: {invalid_contexts}")
 
 
+DEFAULT_CHANNELS = [
+    "local news",
+    "science & technology",
+    "business & finance",
+    "entertainment news",
+    "live",
+    "live_es",
+    "weather",
+    "politics",
+    "international",
+    "top videos",
+    "editor picks",
+]
+
+
 def _complete_features_for_channels(
     channels: list[dict],
     user_features: dict[str, UserPlaylistStats],

{haystack_ml_stack-0.4.5a3 → haystack_ml_stack-0.4.7}/src/haystack_ml_stack.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: haystack-ml-stack
-Version: 0.4.5a3
+Version: 0.4.7
 Summary: Functions related to Haystack ML
 Author-email: Oscar Vega <oscar@haystack.tv>
 License: MIT

haystack_ml_stack-0.4.5a3/src/haystack_ml_stack/_version.py

@@ -1 +0,0 @@
-__version__ = "0.4.5a3"