added openpyxl as extra dependency for filter_files_dependencies

nnpdf_data/nnpdf_data/athena_utils.py

@@ -0,0 +1,24 @@
+import pandas as pd
+from pathlib import Path
+
+def read_excel(path_xlsx: Path, beams: tuple) -> pd.DataFrame:
+    """Parse the xlsx file containing all the information regarding
+    the projections and returns the ones corresponding to the chosen
+    beam energies.
+
+    Parameters
+    ----------
+    path_xlsx : Path
+        path to the xlsx file
+    beams: tuple
+        tuple specifying the beam energies of the lepton & proton
+
+    Returns
+    -------
+    pd.DataFrame
+        returns a panda table corresponding to the chosen beams
+    """
+    xdf = pd.read_excel(path_xlsx)
+    el, ep = beams
+    df_beam = xdf[(xdf["El"] == el) & (xdf["Eh"] == ep)]
+    return df_beam[df_beam["y"] <= 1]

nnpdf_data/pyproject.toml

@@ -28,13 +28,16 @@ exclude = [
 # Data files
 include = [
     # The version file is ignored by git so it needs to be explicitly included
-    "nnpdf_data/_version.py"
+    "nnpdf_data/_version.py",
 ]
 
 [tool.poetry.dependencies]
 python = "^3.9"
 "ruamel.yaml" = "<0.18"
 
+[tool.poetry.extras]
+filter_files_dependencies = ["openpyxl",]
+
 [tool.poetry-dynamic-versioning]
 enable = true
 vcs = "git"