Fix EIC code aggregation and integrate JRC-PPDB-OPEN data source

docs/release-notes.md

@@ -8,7 +8,7 @@ SPDX-License-Identifier: MIT
 
 ## Upcoming Version
 
-*
+* Add the [JRC Open Power Plants Database](https://doi.org/10.5281/zenodo.3574566) (`pm.data.JRC_PPDB_OPEN()`). Its coordinates fill the ENTSOE data via EIC codes, which helps to match ENTSOE with other datasets. In the matched output, other sources' coordinates take priority over these. ([PR #306](https://github.com/PyPSA/powerplantmatching/pull/306))
 
 ## [v0.9.0](https://github.com/PyPSA/powerplantmatching/releases/tag/v0.9.0) (3rd October 2026)
 

powerplantmatching/data.py

@@ -484,6 +484,60 @@ def set_large_spanish_stores_to_reservoirs(df):
     return df
 
 
+def JRC_PPDB_OPEN(raw=False, update=False, config=None):
+    """
+    Importer for the JRC Open Power Plants Database (JRC-PPDB-OPEN,
+    Kanellopoulos et al., 2019, doi:10.5281/zenodo.3574566).
+
+    The database extends the ENTSO-E production units with coordinates,
+    which are used to fill the coordinates of the ENTSOE data. Generation
+    units are aggregated to production units, which have the same EIC codes
+    as the ENTSOE data.
+
+    Parameters
+    ----------
+    raw : bool, default False
+        Whether to return the original dataset
+    update : bool, default False
+        Whether to update the data from the URL
+    config : dict, default None
+        Custom configuration
+    """
+    config = get_config() if config is None else config
+
+    fn = get_raw_file("JRC_PPDB_OPEN", update, config)
+
+    with ZipFile(fn) as zf, zf.open("JRC_OPEN_UNITS.csv") as f:
+        df = pd.read_csv(f)
+
+    if raw:
+        return df
+
+    status_list = config["JRC_PPDB_OPEN"].get("status", ["COMMISSIONED"])  # noqa: F841
+
+    return (
+        df.assign(status_g=df.status_g.str.upper())
+        .query("status_g in @status_list")
+        .groupby("eic_p")
+        .agg(
+            Name=("name_p", "first"),
+            Fueltype=("type_g", "first"),
+            Country=("country", "first"),
+            Capacity=("capacity_g", "sum"),
+            DateIn=("year_commissioned", "min"),
+            lat=("lat", "mean"),
+            lon=("lon", "mean"),
+        )
+        .rename_axis("projectID")
+        .reset_index()
+        .assign(EIC=lambda df: df.projectID, Technology=np.nan, Set=np.nan)
+        .pipe(gather_specifications, config=config)
+        .pipe(clean_name)
+        .pipe(set_column_name, "JRC_PPDB_OPEN")
+        .pipe(config_filter, config)
+    )
+
+
 @deprecated(
     deprecated_in="0.5.0",
     details="Use the JRC data instead",
@@ -688,7 +742,7 @@ def ENTSOE(
 ):
     """
     Importer for the list of installed generators provided by the ENTSO-E
-    Transparency Project. Geographical information is not given.
+    Transparency Project. Coordinates are taken from JRC-PPDB-OPEN via EIC codes.
     If update=True, the dataset is parsed through a request to
     'https://transparency.entsoe.eu/generation/r2/\
     installedCapacityPerProductionUnit/show',
@@ -778,6 +832,7 @@ def retrieve_data(token):
 
     fn = _package_data("entsoe_country_codes.csv")
     COUNTRY_MAP = pd.read_csv(fn, index_col=0).rename(index=str).Country
+    coords = JRC_PPDB_OPEN(config=config).set_index("projectID")
 
     return (
         df.rename_axis(index="projectID")
@@ -791,8 +846,8 @@ def retrieve_data(token):
             Capacity=lambda df: pd.to_numeric(df.Capacity),
             Technology=np.nan,
             Set=np.nan,
-            lat=np.nan,
-            lon=np.nan,
+            lat=lambda df: df.projectID.map(coords.lat),
+            lon=lambda df: df.projectID.map(coords.lon),
         )
         .powerplant.convert_alpha2_to_country()
         # .pipe(fill_geoposition, **fill_geoposition_kwargs)

powerplantmatching/matching.py

@@ -297,6 +297,12 @@ def reduce_matched_dataframe(df, show_orig_names=False, config=None):
     )
     props_for_groups = pd.Series(props_for_groups)[cols].to_dict()
 
+    # ENTSOE coordinates are filled from JRC-PPDB-OPEN, use them only as fallback
+    if "ENTSOE" in sources:
+        df = df.copy()
+        others = df.lat.drop(columns="ENTSOE").notna().any(axis=1)
+        df.loc[others, [("lat", "ENTSOE"), ("lon", "ENTSOE")]] = np.nan
+
     # set low priority on Fueltype 'Other' and Set 'PP'
     # turn it since aggregating only possible for axis=0
     sdf = (

powerplantmatching/package_data/config.yaml

@@ -97,6 +97,10 @@ JRC:
   reliability_score: 5
   fn: jrc-hydro-power-plant-database.csv
   url: https://raw.githubusercontent.com/energy-modelling-toolkit/hydro-power-database/3e8a837/data/jrc-hydro-power-plant-database.csv
+JRC_PPDB_OPEN:
+  status: ["COMMISSIONED", "RESERVE", "MOTHBALLED", "CONSTRUCTION"]
+  fn: JRC-PPDB-OPEN.ver1.0.zip
+  url: https://zenodo.org/records/3574566/files/JRC-PPDB-OPEN.ver1.0.zip
 GEO:
   net_capacity: false
   reliability_score: 2

test/test_data.py

@@ -36,6 +36,10 @@ def test_data_request_processed(source):
     assert df.columns.to_list() == config["target_columns"]
 
 
+def test_ENTSOE_coordinates():
+    assert data.ENTSOE().lat.notna().mean() > 0.5
+
+
 def test_OPSD_VRE():
     df = pm.data.OPSD_VRE()
     assert not df.empty

test/test_matching.py

@@ -0,0 +1,29 @@
+# SPDX-FileCopyrightText: Contributors to powerplantmatching <https://github.com/pypsa/powerplantmatching>
+#
+# SPDX-License-Identifier: MIT
+
+import numpy as np
+import pandas as pd
+import pytest
+
+import powerplantmatching as pm
+from powerplantmatching.matching import reduce_matched_dataframe
+
+
+@pytest.mark.parametrize("gpd_lat, expected", [(50.0, 50.0), (np.nan, 49.0)])
+def test_entsoe_coordinates_are_fallback(gpd_lat, expected):
+    plant = {"Name": "Plant", "Fueltype": "Hydro", "Country": "Austria", "lon": 10.0}
+    df = pd.concat(
+        {
+            "ENTSOE": pd.DataFrame([{**plant, "Capacity": 100.0, "lat": 49.0}]),
+            "GPD": pd.DataFrame([{**plant, "Capacity": 90.0, "lat": gpd_lat}]),
+        },
+        axis=1,
+    ).swaplevel(axis=1)
+    cols = pd.MultiIndex.from_product(
+        [pm.get_config()["target_columns"], ["ENTSOE", "GPD"]]
+    )
+    df = df.reindex(columns=cols)
+    reduced = reduce_matched_dataframe(df)
+    assert reduced.loc[0, "lat"] == expected
+    assert reduced.loc[0, "Capacity"] == 100.0