Improve filename consistency in the sources (#189)

* Improve source filename consistency

* Fix minor formatting

* Add release note

* code: apply pre-commit changes

* Update outputs using latest data from master

* Update release_notes.rst

* ruff fix

---------

Co-authored-by: Thomas Gilon <[email protected]>
Co-authored-by: Fabrizio Finozzi <[email protected]>
Co-authored-by: Lukas Trippe <[email protected]>

Author: 4 people

docs/release_notes.rst

@@ -24,6 +24,8 @@ Upcoming Release
 
 * Align `snakemake` version and the related `mock_snakemake` to PyPSA-Eur (https://github.com/PyPSA/technology-data/pull/177)
 
+* Improve filename consistency in the sources (https://github.com/PyPSA/technology-data/pull/178)
+
 * Include further unit tests for compile_cost_assumptions_usa.py (https://github.com/PyPSA/technology-data/pull/182)
 
 * Updates the documentation with compile_cost_assumptions_usa.py (https://github.com/PyPSA/technology-data/pull/186)

outputs/US/costs_2020.csv

@@ -5,6 +5,7 @@
 # coding: utf-8
 
 import logging
+import os
 import re
 import sys
 from pathlib import Path
@@ -64,6 +65,12 @@ def __dir__(self):
         return dict_keys + obj_attrs
 
 
+def get_relative_fn(fn):
+    if isinstance(fn, str):
+        fn = Path(fn).resolve()
+    return fn.relative_to(os.path.commonpath([fn, os.getcwd()]))
+
+
 def mock_snakemake(
     rulename,
     root_dir=None,

outputs/US/costs_2025.csv

@@ -33,7 +33,12 @@
 
 import numpy as np
 import pandas as pd
-from _helpers import adjust_for_inflation, configure_logging, mock_snakemake
+from _helpers import (
+    adjust_for_inflation,
+    configure_logging,
+    get_relative_fn,
+    mock_snakemake,
+)
 from currency_converter import ECB_URL, CurrencyConverter
 from scipy import interpolate
 
@@ -428,7 +433,7 @@ def get_dea_maritime_data(
         df.loc[df_i, "unit"] = df.loc[df_i, "unit"].str.replace("GJ", "MWh")
 
         # add source + cost year
-        df["source"] = f"Danish Energy Agency, {fn}"
+        df["source"] = f"Danish Energy Agency, {get_relative_fn(fn)}"
         # cost year is 2023 p.10
         df["currency_year"] = 2023
         # add sheet name
@@ -552,7 +557,7 @@ def get_dea_vehicle_data(
         df.loc["Upfront vehicle cost", "unit"] += "/vehicle"
 
         # add source + cost year
-        df["source"] = f"Danish Energy Agency, {fn}"
+        df["source"] = f"Danish Energy Agency, {get_relative_fn(fn)}"
         # cost year is 2022 p.12
         df["currency_year"] = 2022
         # add sheet name
@@ -982,7 +987,7 @@ def get_data_DEA(
     # if year-specific data is missing and not fixed by interpolation fill forward with same values
     df_final = df_final.ffill(axis=1)
 
-    df_final["source"] = source_dict["DEA"] + ", " + excel_file.replace("inputs/", "")
+    df_final["source"] = f"{source_dict['DEA']}, {get_relative_fn(excel_file)}"
     if (
         tech_name in cost_year_2020
         and ("for_carbon_capture_transport_storage" not in excel_file)