Remove support for Python 3.9, add Python 3.13 (#40)

* Remove support for py39, add py313 * Actually remove 3.9 from Actions * Try upgrading pyproj * Update uv lock * Upgrade pyogrio * Update scipy * Exclude python3.13 combo * Add benchmark notebook link * Update generated PDF --------- Co-authored-by: robbibt <[email protected]>
GeoscienceAustralia · Jan 17, 2025 · d7a385a · d7a385a
1 parent bb56b91
commit d7a385a
Show file tree

Hide file tree

Showing 5 changed files with 1,128 additions and 1,328 deletions.
diff --git a/.github/workflows/main.yml b/.github/workflows/main.yml
@@ -38,8 +38,11 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: ["3.9", "3.10", "3.11", "3.12"] # , "3.13" # waiting on pyproj
+        python-version: ["3.10", "3.11", "3.12", "3.13"]
         resolution: ["highest", "lowest-direct"]
+        exclude: # Temporarily excluded due to 3.13 only supported by Numpy >=2.0
+          - python-version: "3.13"
+            resolution: "lowest-direct"
       fail-fast: false
       max-parallel: 10
     defaults:

diff --git a/paper/paper.md b/paper/paper.md
@@ -74,7 +74,7 @@ The [`eo_tides.model`](https://geoscienceaustralia.github.io/eo-tides/api/#eo_ti
 
 To support integration with satellite EO data, the `model_tides` function from `eo_tides.model` wraps `pyTMD` functionality to return predicted tides in a standardised `pandas.DataFrame` format containing information about the tide model, location and time period of each modelled tide. This allows large analyses to be broken into smaller discrete chunks that can be processed in parallel before being combined as a final step. Parallelisation in `eo-tides` is automatically optimised based on the number of available workers and the number of requested tide models and tide modelling locations. This built-in parallelisation can significantly improve tide modelling performance, especially when run on a large multi-core machine (\autoref{tab:benchmark}).
 
-Table: A benchmark comparison of tide modelling performance with parallelisation on vs. off. This comparison was performed on an 8-core and 32-core Linux machine, for a typical large-scale analysis involving a month of hourly tides modelled at 10,000 modelling locations using three tide models (FES2022, TPXO10, GOT5.6). \label{tab:benchmark}
+Table: A [benchmark comparison](https://github.com/GeoscienceAustralia/eo-tides/blob/main/paper/benchmarking.ipynb) of tide modelling performance with parallelisation on vs. off. This comparison was performed on an 8-core and 32-core Linux machine, for a typical large-scale analysis involving a month of hourly tides modelled at 10,000 modelling locations using three tide models (FES2022, TPXO10, GOT5.6). \label{tab:benchmark}
 
 | Cores | Parallelisation   | No parallelisation | Speedup |
 | ----- | ----------------- | ------------------ | ------- |

diff --git a/paper/paper.pdf b/paper/paper.pdf
diff --git a/pyproject.toml b/pyproject.toml
@@ -34,7 +34,7 @@ classifiers = [
     "Programming Language :: Python :: 3.12",
     "Programming Language :: Python :: 3.13",
 ]
-requires-python = ">=3.9,<4.0"
+requires-python = ">=3.10,<4.0"
 dependencies = [
     "colorama>=0.4.3",
     "geopandas>=0.10.0",
@@ -43,11 +43,11 @@ dependencies = [
     "odc-geo>=0.4.7",
     "pandas>=2.2.0",
     "psutil>=5.8.0",
-    "pyogrio>=0.7.0",
-    "pyproj>=3.6.1",
+    "pyogrio>=0.10.0",
+    "pyproj>=3.7.0",
     "pyTMD==2.2.0",
     "scikit-learn>=1.4.0",
-    "scipy>=1.11.2",
+    "scipy>=1.14.1",
     "shapely>=2.0.6",
     "tqdm>=4.55.0",
     "xarray>=2022.3.0",