Update dependencies 2024/04 (#68)

* Bump pre-commit action version

* Bump dependency versions

* Workaround upcoming pandas.df.groupby.apply limitation

* remove column copies

* update dask and lock

* put "type" back in map func

* put grouped columns back

* Relax dependency constraints, remove upper bounds

---------

Co-authored-by: Max Wolschlager <comms@meks.io>

.github/workflows/pytest.yaml

@@ -158,4 +158,4 @@ jobs:
       #-------------------------------
       #  run pre-commit/(mypy, ruff)
       #-------------------------------
-      - uses: pre-commit/action@v3.0.0
+      - uses: pre-commit/action@v3.0.1

ixmp4/data/db/iamc/timeseries/repository.py

@@ -123,10 +123,14 @@ def map_measurand(df):
                 variable_name=variable_name, unit__id=int(unit__id)
             )
             df["measurand__id"] = measurand.id
+            df["variable"] = variable_name
+            df["unit__id"] = unit__id
             return df
 
         return pd.DataFrame(
-            df.groupby(["variable", "unit__id"], group_keys=False).apply(map_measurand)
+            df.groupby(["variable", "unit__id"], group_keys=False).apply(
+                map_measurand, include_groups=False
+            )
         )
 
     def map_relationships(self, df: pd.DataFrame) -> pd.DataFrame:

ixmp4/data/db/meta/repository.py

@@ -188,9 +188,12 @@ def map_value_column(df: pd.DataFrame):
             type_ = RunMetaEntry.Type(type_str)
             col = RunMetaEntry._column_map[type_]
             df["value"] = df[col]
+            df["type"] = type_str
             return df.drop(columns=RunMetaEntry._column_map.values())
 
-        return df.groupby("type", group_keys=False).apply(map_value_column)
+        return df.groupby("type", group_keys=False).apply(
+            map_value_column, include_groups=False
+        )
 
     @check_types
     @guard("edit")