Bump polars version (#319)

We had tied our polars version to 1.16 because 1.17 introduced a regression. seems fixed now - so we can get back up to speed

.pre-commit-config.yaml

@@ -6,6 +6,7 @@ repos:
       - id: end-of-file-fixer
       - id: check-yaml
       - id: check-added-large-files
+        exclude: 'uv.lock'
 
   - repo: https://github.com/astral-sh/ruff-pre-commit
     # Ruff version.

pyproject.toml

@@ -31,7 +31,10 @@ keywords = [
     "tools",
 ]
 requires-python = ">=3.9"
-dependencies = ['polars==1.16', 'typing_extensions']
+dependencies = [
+    "polars>=0.20",
+    'typing_extensions',
+]
 
 [tool.setuptools.dynamic]
 version = {attr="pdstools.__version__"}
@@ -42,7 +45,14 @@ pega_io = ['aioboto3', 'polars_hash']
 api = ['httpx', 'pydantic', 'anyio']
 healthcheck = ['pdstools[adm]', 'great_tables>=0.13', 'quarto', 'papermill', 'xlsxwriter>=3.0', 'pydot']
 app = ['pdstools[healthcheck]', 'streamlit>=1.23', 'st-pages']
-onnx = ['scikit-learn==1.5.1', 'skl2onnx==1.17.0', 'onnx==1.16.1', 'onnxruntime==1.18.1', 'pdstools[api]']
+onnx = [
+    'scikit-learn==1.5.1',
+    'skl2onnx==1.17.0',
+    'onnx==1.16.1',
+    'pdstools[api]',
+    "onnxruntime==1.18.1 ; python_full_version < '3.10'",
+    "onnxruntime==1.20.1 ; python_full_version >= '3.10'",
+]
 all = ['pdstools[app, pega_io, onnx]']
 dev = ['pre-commit']
 docs = ['sphinx','furo','sphinx-autoapi','nbsphinx','sphinx-copybutton','myst-parser', 'sphinx-argparse']

python/pdstools/__init__.py

@@ -1,6 +1,6 @@
 """Pega Data Scientist Tools Python library"""
 
-__version__ = "4.1.1"
+__version__ = "4.1.2"
 
 from pathlib import Path
 

python/pdstools/adm/Plots.py

@@ -323,7 +323,7 @@ def over_time(
         )
 
         grouping_columns = list(set(by.meta.root_names()))
-        df = df.with_columns(by)
+        df = df.with_columns(by).set_sorted("SnapshotTime")
         if facet:
             grouping_columns = [by_col, facet]
         else:
@@ -367,7 +367,7 @@ def over_time(
 
         title = "over all models" if facet is None else f"per {facet}"
         fig = px.line(
-            final_df.to_pandas(use_pyarrow_extension_array=False),
+            final_df,
             x="SnapshotTime",
             y=metric,
             color=by_col,