determine_availability_matrix_MD_UA: enable parallelism & remove plots (#1170)

* determine_availability_matrix_MD_UA: enable parallelism through temp files and remove plots

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

doc/release_notes.rst

@@ -31,14 +31,16 @@ Upcoming Release
 
 * Bugfix: Correctly read in threshold capacity below which to remove components from previous planning horizons in :mod:`add_brownfield`.
 
-* For countries not contained in the NUTS3-specific datasets (i.e. MD and UA), the mapping of GDP per capita and population per bus region used to spatially distribute electricity demand is now endogenised in a new rule :mod:`build_gdp_ppp_non_nuts3`. https://github.com/PyPSA/pypsa-eur/pull/1146 
+* For countries not contained in the NUTS3-specific datasets (i.e. MD and UA), the mapping of GDP per capita and population per bus region used to spatially distribute electricity demand is now endogenised in a new rule :mod:`build_gdp_ppp_non_nuts3`. https://github.com/PyPSA/pypsa-eur/pull/1146
 
-* The databundle has been updated to release v0.3.0, which includes raw GDP and population data for countries outside the NUTS system (UA, MD). https://github.com/PyPSA/pypsa-eur/pull/1146 
+* The databundle has been updated to release v0.3.0, which includes raw GDP and population data for countries outside the NUTS system (UA, MD). https://github.com/PyPSA/pypsa-eur/pull/1146
 
-* Updated filtering in :mod:`determine_availability_matrix_MD_UA.py` to improve speed. https://github.com/PyPSA/pypsa-eur/pull/1146 
+* Updated filtering in :mod:`determine_availability_matrix_MD_UA.py` to improve speed. https://github.com/PyPSA/pypsa-eur/pull/1146
 
 * Bugfix: Impose minimum value of zero for district heating progress between current and future market share in :mod:`build_district_heat_share`.
 
+* Enable parallelism in :mod:`determine_availability_matrix_MD_UA.py` and remove plots. This requires the use of temporary files.
+
 PyPSA-Eur 0.11.0 (25th May 2024)
 =====================================
 

rules/build_electricity.smk

@@ -202,7 +202,6 @@ rule determine_availability_matrix_MD_UA:
         + ".nc",
     output:
         availability_matrix=resources("availability_matrix_MD-UA_{technology}.nc"),
-        availability_map=resources("availability_matrix_MD-UA_{technology}.png"),
     log:
         logs("determine_availability_matrix_MD_UA_{technology}.log"),
     threads: config["atlite"].get("nprocesses", 4)

scripts/build_gdp_pop_non_nuts3.py

@@ -3,10 +3,11 @@
 #
 # SPDX-License-Identifier: MIT
 """
-Maps the per-capita GDP and population values to non-NUTS3 regions. The script
+Maps the per-capita GDP and population values to non-NUTS3 regions.
-takes as input the country code, a GeoDataFrame containing the regions, and the
+
-file paths to the datasets containing the GDP and POP values for non-NUTS3
+The script takes as input the country code, a GeoDataFrame containing
-countries.
+the regions, and the file paths to the datasets containing the GDP and
+POP values for non-NUTS3 countries.
 """
 
 import logging

scripts/determine_availability_matrix_MD_UA.py

@@ -8,16 +8,15 @@ Create land elibility analysis for Ukraine and Moldova with different datasets.
 
 import functools
 import logging
+import os
 import time
+from tempfile import NamedTemporaryFile
 
 import atlite
 import fiona
 import geopandas as gpd
-import matplotlib.pyplot as plt
 import numpy as np
 from _helpers import configure_logging, set_scenario_config
-from atlite.gis import shape_availability
-from rasterio.plot import show
 
 logger = logging.getLogger(__name__)
 
@@ -40,7 +39,7 @@ if __name__ == "__main__":
     configure_logging(snakemake)
     set_scenario_config(snakemake)
 
-    nprocesses = None  # snakemake.config["atlite"].get("nprocesses")
+    nprocesses = int(snakemake.threads)
     noprogress = not snakemake.config["atlite"].get("show_progress", True)
     config = snakemake.config["renewable"][snakemake.wildcards.technology]
 
@@ -95,8 +94,15 @@ if __name__ == "__main__":
             bbox=regions.geometry,
             layer=layer,
         ).to_crs(3035)
+
+        # temporary file needed for parallelization
+        with NamedTemporaryFile(suffix=".geojson", delete=False) as f:
+            plg_tmp_fn = f.name
         if not wdpa.empty:
-            excluder.add_geometry(wdpa.geometry)
+            wdpa[["geometry"]].to_file(plg_tmp_fn)
+            while not os.path.exists(plg_tmp_fn):
+                time.sleep(1)
+            excluder.add_geometry(plg_tmp_fn)
 
         layer = get_wdpa_layer_name(wdpa_fn, "points")
         wdpa_pts = gpd.read_file(
@@ -109,8 +115,15 @@ if __name__ == "__main__":
         wdpa_pts = wdpa_pts.set_geometry(
             wdpa_pts["geometry"].buffer(wdpa_pts["buffer_radius"])
         )
+
+        # temporary file needed for parallelization
+        with NamedTemporaryFile(suffix=".geojson", delete=False) as f:
+            pts_tmp_fn = f.name
         if not wdpa_pts.empty:
-            excluder.add_geometry(wdpa_pts.geometry)
+            wdpa_pts[["geometry"]].to_file(pts_tmp_fn)
+            while not os.path.exists(pts_tmp_fn):
+                time.sleep(1)
+            excluder.add_geometry(pts_tmp_fn)
 
     if "max_depth" in config:
         # lambda not supported for atlite + multiprocessing
@@ -146,13 +159,9 @@ if __name__ == "__main__":
     else:
         availability = cutout.availabilitymatrix(regions, excluder, **kwargs)
 
-    regions_geometry = regions.to_crs(3035).geometry
+    for fn in [pts_tmp_fn, plg_tmp_fn]:
-    band, transform = shape_availability(regions_geometry, excluder)
+        if os.path.exists(fn):
-    fig, ax = plt.subplots(figsize=(4, 8))
+            os.remove(fn)
-    gpd.GeoSeries(regions_geometry.union_all()).plot(ax=ax, color="none")
-    show(band, transform=transform, cmap="Greens", ax=ax)
-    plt.axis("off")
-    plt.savefig(snakemake.output.availability_map, bbox_inches="tight", dpi=500)
 
     availability = availability.sel(bus=buses)