Upgade numpy and python version (#573)

* Upgade numpy version

Signed-off-by: Nathalie Jonathan <[email protected]>

* Upgrade python version

Signed-off-by: Nathalie Jonathan <[email protected]>

* Update changelog

Signed-off-by: Nathalie Jonathan <[email protected]>

* Fix failing tests

Signed-off-by: Nathalie Jonathan <[email protected]>

---------

Signed-off-by: Nathalie Jonathan <[email protected]>

Author: nathaliellenaa

.ci/Dockerfile.client

@@ -1,4 +1,4 @@
-ARG PYTHON_VERSION=3.9
+ARG PYTHON_VERSION=3.12
 FROM python:${PYTHON_VERSION}
 
 WORKDIR /code/opensearch-py-ml

.ci/run-tests

@@ -4,7 +4,7 @@
 
 # Default environment variables
 export TEST_SUITE="${TEST_SUITE:=oss}"
-export PYTHON_VERSION="${PYTHON_VERSION:=3.9}"
+export PYTHON_VERSION="${PYTHON_VERSION:=3.12}"
 export PANDAS_VERSION=${PANDAS_VERSION-1.5.0}
 export PYTHON_CONNECTION_CLASS="${PYTHON_CONNECTION_CLASS:=Urllib3HttpConnection}"
 export CLUSTER="${1:-opensearch}"

CHANGELOG.md

@@ -34,6 +34,7 @@ Inspired from [Keep a Changelog](https://keepachangelog.com/en/1.0.0/)
 - Update model upload history -  opensearch-project/opensearch-semantic-highlighter-v1 (v.1.0.0)(TORCH_SCRIPT) by @nathaliellenaa ([#550](https://github.com/opensearch-project/opensearch-py-ml/pull/550))
 - Update pretrained_models_all_versions.json by @nathaliellenaa ([#560](https://github.com/opensearch-project/opensearch-py-ml/pull/560))
 - Upgrade torch version to fix security issue ([#561](https://github.com/opensearch-project/opensearch-py-ml/pull/561))
+- Upgrade numpy and python version ([#562](https://github.com/opensearch-project/opensearch-py-ml/pull/562))
 
 ### Fixed
 - Fix for uploading models with function_name instead of model_task ([#553](https://github.com/opensearch-project/opensearch-py-ml/pull/553))

DEVELOPER_GUIDE.md

@@ -42,7 +42,7 @@ If you don't already have it installed (check with `git --version`) we recommend
 
 #### Install `python`
 
-You can install any version of python starting from 3.8.
+You can install any version of python starting from 3.11.
 
 ### Fork and clone OpenSearch-py-ml
 

noxfile.py

@@ -102,7 +102,7 @@ def lint(session):
     #         session.error("\n" + "\n".join(sorted(set(errors))))
 
 
-@nox.session(python=["3.8", "3.9", "3.10"])
+@nox.session(python=["3.11", "3.12", "3.13"])
 def test(session):
     session.install(
         "-r",
@@ -137,7 +137,7 @@ def test(session):
     )
 
 
-@nox.session(python=["3.9"])
+@nox.session(python=["3.12"])
 def docs(session):
     # Run this so users get an error if they don't have Pandoc installed.
     session.install("-r", "docs/requirements-docs.txt")
@@ -151,7 +151,7 @@ def docs(session):
 # While nox is typically used for automating testing, in this case, we utilize it
 # to automate the action workflow, leveraging its ability to set up the environment
 # required for model autotracing.
-@nox.session(python=["3.9"])
+@nox.session(python=["3.12"])
 def trace(session):
     session.install(
         "-r",
@@ -168,7 +168,7 @@ def trace(session):
     )
 
 
-@nox.session(python=["3.9"])
+@nox.session(python=["3.12"])
 def sparsetrace(session):
     session.install(
         "-r",
@@ -185,7 +185,7 @@ def sparsetrace(session):
     )
 
 
-@nox.session(python=["3.9"])
+@nox.session(python=["3.12"])
 def semantic_highlighter_trace(session):
     session.install(
         "-r",

opensearch_py_ml/field_mappings.py

@@ -132,7 +132,7 @@ def nan_value(self) -> Any:
         """Returns NaN for any field except datetimes which use NaT"""
         if self.is_timestamp:
             return pd.NaT
-        return np.float64(np.NaN)
+        return np.float64(np.nan)
 
 
 class FieldMappings:

opensearch_py_ml/operations.py

@@ -311,7 +311,7 @@ def mode(
                     if isinstance(values[0], pd.Timestamp):
                         values.extend([pd.NaT] * row_diff)
                     else:
-                        values.extend([np.NaN] * row_diff)
+                        values.extend([np.nan] * row_diff)
                 pd_dict[key] = values
 
             return pd.DataFrame(pd_dict)
@@ -666,7 +666,7 @@ def _unpack_metric_aggs(
 
                         # All of the below calculations result in NaN if count<=1
                         if count <= 1:
-                            agg_value = np.NaN
+                            agg_value = np.nan
 
                         elif os_agg[1] == "std_deviation":
                             agg_value *= count / (count - 1.0)
@@ -716,16 +716,16 @@ def _unpack_metric_aggs(
                     agg_value is None or np.isnan(agg_value)
                 ):
                     if is_dataframe_agg and not numeric_only:
-                        agg_value = np.NaN
+                        agg_value = np.nan
                     elif not is_dataframe_agg and numeric_only is False:
-                        agg_value = np.NaN
+                        agg_value = np.nan
 
                 # Cardinality is always either NaN or integer.
                 elif pd_agg in ("nunique", "count"):
                     agg_value = (
                         int(agg_value)
                         if isinstance(agg_value, (int, float))
-                        else np.NaN
+                        else np.nan
                     )
 
                 # If this is a non-null timestamp field convert to a pd.Timestamp()

opensearch_py_ml/series.py

@@ -1620,7 +1620,7 @@ def nunique(self) -> pd.Series:
 
         >>> s = oml.DataFrame(OPENSEARCH_TEST_CLIENT, 'flights')['Carrier']
         >>> s.nunique()
-        4
+        np.int64(4)
         """
         results = super().nunique()
         return results.squeeze()

requirements-dev.txt

@@ -3,7 +3,7 @@
 #
 pandas>=1.5.2,<2.3,!=2.1.0
 matplotlib>=3.6.2,<4
-numpy>=1.24.0,<2
+numpy>=2.3.2
 opensearch-py>=2.2.0
 torch>=2.7.1-rc1
 onnx>=1.15.0

requirements.txt

@@ -3,7 +3,7 @@
 #
 pandas>=1.5.2,<2.3,!=2.1.0
 matplotlib>=3.6.2,<4
-numpy>=1.24.0,<2
+numpy>=2.3.2
 opensearch-py>=2.2.0
 torch>=2.7.1-rc1
 onnx>=1.15.0