replay-rec 0.20.3__tar.gz → 0.20.3rc0__tar.gz

{replay_rec-0.20.3 → replay_rec-0.20.3rc0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: replay-rec
-Version: 0.20.3
+Version: 0.20.3rc0
 Summary: RecSys Library
 License-Expression: Apache-2.0
 License-File: LICENSE
@@ -14,29 +14,23 @@ Classifier: Intended Audience :: Developers
 Classifier: Intended Audience :: Science/Research
 Classifier: Natural Language :: English
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
-Provides-Extra: spark
-Provides-Extra: torch
-Provides-Extra: torch-cpu
-Requires-Dist: lightning (<2.6.0) ; extra == "torch" or extra == "torch-cpu"
-Requires-Dist: lightning ; extra == "torch"
-Requires-Dist: lightning ; extra == "torch-cpu"
+Requires-Dist: d3rlpy (>=2.8.1,<2.9)
+Requires-Dist: implicit (>=0.7.2,<0.8)
+Requires-Dist: lightautoml (>=0.4.1,<0.5)
+Requires-Dist: lightning (>=2.0.2,<=2.4.0)
+Requires-Dist: numba (>=0.50,<1)
 Requires-Dist: numpy (>=1.20.0,<2)
 Requires-Dist: pandas (>=1.3.5,<2.4.0)
 Requires-Dist: polars (<2.0)
-Requires-Dist: psutil (<=7.0.0) ; extra == "spark"
-Requires-Dist: psutil ; extra == "spark"
+Requires-Dist: psutil (<=7.0.0)
 Requires-Dist: pyarrow (<22.0)
-Requires-Dist: pyspark (>=3.0,<3.5) ; extra == "spark"
-Requires-Dist: pyspark ; extra == "spark"
-Requires-Dist: pytorch-optimizer (>=3.8.0,<3.9.0) ; extra == "torch" or extra == "torch-cpu"
-Requires-Dist: pytorch-optimizer ; extra == "torch"
-Requires-Dist: pytorch-optimizer ; extra == "torch-cpu"
+Requires-Dist: pyspark (>=3.0,<3.5)
+Requires-Dist: pytorch-optimizer (>=3.8.0,<4)
+Requires-Dist: sb-obp (>=0.5.10,<0.6)
 Requires-Dist: scikit-learn (>=1.6.1,<1.7.0)
 Requires-Dist: scipy (>=1.8.1,<2.0.0)
 Requires-Dist: setuptools
-Requires-Dist: torch (>=1.8,<2.9.0) ; extra == "torch" or extra == "torch-cpu"
-Requires-Dist: torch ; extra == "torch"
-Requires-Dist: torch ; extra == "torch-cpu"
+Requires-Dist: torch (>=1.8,<2.9.0)
 Requires-Dist: tqdm (>=4.67,<5)
 Project-URL: Homepage, https://sb-ai-lab.github.io/RePlay/
 Project-URL: Repository, https://github.com/sb-ai-lab/RePlay

{replay_rec-0.20.3 → replay_rec-0.20.3rc0}/pyproject.toml

@@ -40,19 +40,19 @@ dependencies = [
     "scikit-learn (>=1.6.1,<1.7.0)",
     "pyarrow (<22.0)",
     "tqdm (>=4.67,<5)",
-    "pyspark (>=3.0,<3.5); extra == 'spark'",
-    "psutil (<=7.0.0); extra == 'spark'",
-    "torch (>=1.8, <2.9.0); extra == 'torch' or extra == 'torch-cpu'",
-    "pytorch-optimizer (>=3.8.0,<3.9.0); extra == 'torch' or extra == 'torch-cpu'",
-    "lightning (<2.6.0); extra == 'torch' or extra == 'torch-cpu'",
+    "torch (>=1.8,<2.9.0)",
+    "lightning (>=2.0.2,<=2.4.0)",
+    "pytorch-optimizer (>=3.8.0,<4)",
+    "lightautoml (>=0.4.1,<0.5)",
+    "numba (>=0.50,<1)",
+    "sb-obp (>=0.5.10,<0.6)",
+    "d3rlpy (>=2.8.1,<2.9)",
+    "implicit (>=0.7.2,<0.8)",
+    "pyspark (>=3.0,<3.5)",
+    "psutil (<=7.0.0)",
 ]
 dynamic = ["dependencies"]
-version = "0.20.3"
-
-[project.optional-dependencies]
-spark = ["pyspark", "psutil"]
-torch = ["torch", "pytorch-optimizer", "lightning"]
-torch-cpu = ["torch", "pytorch-optimizer", "lightning"]
+version = "0.20.3.preview"
 
 [project.urls]
 homepage = "https://sb-ai-lab.github.io/RePlay/"
@@ -66,13 +66,6 @@ target-version = ["py39", "py310", "py311", "py312"]
 packages = [{include = "replay"}]
 exclude = [
     "replay/conftest.py",
-    "replay/experimental",
-]
-
-[tool.poetry.dependencies]
-torch = [
-    {markers = "extra == 'torch-cpu' and extra !='torch'", source = "torch-cpu-mirror"},
-    {markers = "extra == 'torch' and extra !='torch-cpu'", source = "PyPI"},
 ]
 
 [tool.poetry.group.dev.dependencies]
@@ -95,11 +88,6 @@ docutils = "0.16"
 data-science-types = "0.2.23"
 filelock = "~3.14.0"
 
-[[tool.poetry.source]]
-name = "torch-cpu-mirror"
-url = "https://download.pytorch.org/whl/cpu"
-priority = "explicit"
-
 [tool.poetry-dynamic-versioning]
 enable = false
 format-jinja = """0.20.3{{ env['PACKAGE_SUFFIX'] }}"""

{replay_rec-0.20.3 → replay_rec-0.20.3rc0}/replay/__init__.py

@@ -4,4 +4,4 @@
 # functionality removed in Python 3.12 is used in downstream packages (like lightfm)
 import setuptools as _
 
-__version__ = "0.20.3"
+__version__ = "0.20.3.preview"

replay_rec-0.20.3rc0/replay/experimental/metrics/__init__.py

@@ -0,0 +1,62 @@
+"""
+Most metrics require dataframe with recommendations
+and dataframe with ground truth values —
+which objects each user interacted with.
+
+- recommendations (Union[pandas.DataFrame, spark.DataFrame]):
+    predictions of a recommender system,
+    DataFrame with columns ``[user_id, item_id, relevance]``
+- ground_truth (Union[pandas.DataFrame, spark.DataFrame]):
+    test data, DataFrame with columns
+    ``[user_id, item_id, timestamp, relevance]``
+
+Metric is calculated for all users, presented in ``ground_truth``
+for accurate metric calculation in case when the recommender system generated
+recommendation not for all users.  It is assumed, that all users,
+we want to calculate metric for, have positive interactions.
+
+But if we have users, who observed the recommendations, but have not responded,
+those users will be ignored and metric will be overestimated.
+For such case we propose additional optional parameter ``ground_truth_users``,
+the dataframe with all users, which should be considered during the metric calculation.
+
+- ground_truth_users (Optional[Union[pandas.DataFrame, spark.DataFrame]]):
+    full list of users to calculate metric for, DataFrame with ``user_id`` column
+
+Every metric is calculated using top ``K`` items for each user.
+It is also possible to calculate metrics
+using multiple values for ``K`` simultaneously.
+In this case the result will be a dictionary and not a number.
+
+Make sure your recommendations do not contain user-item duplicates
+as duplicates could lead to the wrong calculation results.
+
+- k (Union[Iterable[int], int]):
+    a single number or a list, specifying the
+    truncation length for recommendation list for each user
+
+By default, metrics are averaged by users,
+but you can alternatively use method ``metric.median``.
+Also, you can get the lower bound
+of ``conf_interval`` for a given ``alpha``.
+
+Diversity metrics require extra parameters on initialization stage,
+but do not use ``ground_truth`` parameter.
+
+For each metric, a formula for its calculation is given, because this is
+important for the correct comparison of algorithms, as mentioned in our
+`article <https://arxiv.org/abs/2206.12858>`_.
+"""
+
+from replay.experimental.metrics.base_metric import Metric, NCISMetric
+from replay.experimental.metrics.coverage import Coverage
+from replay.experimental.metrics.hitrate import HitRate
+from replay.experimental.metrics.map import MAP
+from replay.experimental.metrics.mrr import MRR
+from replay.experimental.metrics.ncis_precision import NCISPrecision
+from replay.experimental.metrics.ndcg import NDCG
+from replay.experimental.metrics.precision import Precision
+from replay.experimental.metrics.recall import Recall
+from replay.experimental.metrics.rocauc import RocAuc
+from replay.experimental.metrics.surprisal import Surprisal
+from replay.experimental.metrics.unexpectedness import Unexpectedness