Merge branch for alternative replacement for datasets previously crea…

…ted with partial

Propose alternative replacements for partial.

docs/conf.py

@@ -47,3 +47,8 @@
 
 add_module_names = False
 modindex_common_prefix = ["egon.data.", "egon.data.datasets."]
+
+autodoc_type_aliases = {
+    "Dependencies": "egon.data.datasets.Dependencies",
+    "Tasks": "egon.data.datasets.Tasks"
+}

src/egon/data/datasets/__init__.py

@@ -160,6 +160,10 @@ def __init__(self, graph: TaskGraph):
             )
 
 
+#: A dataset can depend on other datasets or the tasks of other datasets.
+Dependencies = Iterable[Union["Dataset", Task]]
+
+
 @dataclass
 class Dataset:
     #: The name of the Dataset
@@ -176,7 +180,7 @@ class Dataset:
     #: downstream of any of the listed dependencies. In case of bare
     #: :class:`Task`, a direct link will be created whereas for a
     #: :class:`Dataset` the link will be made to all of its last tasks.
-    dependencies: Iterable[Union[Dataset, Task]] = ()
+    dependencies: Dependencies = ()
     #: The tasks of this :class:`Dataset`. A :class:`TaskGraph` will
     #: automatically be converted to :class:`Tasks_`.
     tasks: Tasks = ()

src/egon/data/datasets/electricity_demand_timeseries/hh_buildings.py

@@ -3,7 +3,7 @@
 assigned to OSM-buildings.
 
 """
-from dataclasses import dataclass
+
 import random
 
 from geoalchemy2 import Geometry
@@ -14,7 +14,7 @@
 import pandas as pd
 
 from egon.data import db
-from egon.data.datasets import Dataset, Tasks
+from egon.data.datasets import Dataset
 from egon.data.datasets.electricity_demand_timeseries.hh_profiles import (
     HouseholdElectricityProfilesInCensusCells,
     get_iee_hh_demand_profiles_raw,
@@ -38,7 +38,9 @@ class HouseholdElectricityProfilesOfBuildings(Base):
 
     Mapping of demand timeseries and buildings and cell_id. This table is created within
     :py:func:`hh_buildings.map_houseprofiles_to_buildings()`.
+
     """
+
     __tablename__ = "egon_household_electricity_profile_of_buildings"
     __table_args__ = {"schema": "demand"}
 
@@ -753,7 +755,6 @@ def map_houseprofiles_to_buildings():
         )
 
 
-@dataclass
 class setup(Dataset):
     """
     Household electricity demand time series for scenarios in 2035 and 2050
@@ -850,30 +851,38 @@ class setup(Dataset):
 
     .. code-block:: SQL
 
-        SELECT t1.cell_id, building_count, hh_count, hh_types
-            FROM (
-                SELECT
-                    cell_id,
-                    Count(distinct(building_id)) as building_count,
-                    count(profile_id) as hh_count
-                FROM demand.egon_household_electricity_profile_of_buildings
-                GROUP BY cell_id
-            ) as t1
+        SELECT t1.cell_id, building_count, hh_count, hh_types FROM (
+            SELECT
+                cell_id,
+                COUNT(DISTINCT(building_id)) AS building_count,
+                COUNT(profile_id) AS hh_count
+            FROM demand.egon_household_electricity_profile_of_buildings
+            GROUP BY cell_id
+        ) AS t1
         FULL OUTER JOIN (
             SELECT
                 cell_id,
                 array_agg(
-                    array[cast(hh_10types as char), hh_type]
-                ) as hh_types
+                    array[CAST(hh_10types AS char), hh_type]
+                ) AS hh_types
             FROM society.egon_destatis_zensus_household_per_ha_refined
             GROUP BY cell_id
-        ) as t2
+        ) AS t2
         ON t1.cell_id = t2.cell_id
 
     """
+
     #:
     name: str = "Demand_Building_Assignment"
     #:
     version: str = "0.0.5"
     #:
-    tasks: Tasks = (map_houseprofiles_to_buildings, get_building_peak_loads)
+    tasks = (map_houseprofiles_to_buildings, get_building_peak_loads)
+
+    def __init__(self, dependencies):
+        super().__init__(
+            name=self.name,
+            version=self.version,
+            dependencies=dependencies,
+            tasks=self.tasks,
+        )

src/egon/data/datasets/mastr.py

@@ -3,12 +3,11 @@
 
 """
 
-from dataclasses import dataclass
 from pathlib import Path
 from urllib.request import urlretrieve
 import os
 
-from egon.data.datasets import Dataset, Tasks
+from egon.data.datasets import Dataset
 import egon.data.config
 
 WORKING_DIR_MASTR_OLD = Path(".", "bnetza_mastr", "dump_2021-05-03")
@@ -50,7 +49,6 @@ def download(dataset_name, download_dir):
     download(dataset_name="mastr_new", download_dir=WORKING_DIR_MASTR_NEW)
 
 
-@dataclass
 class mastr_data_setup(Dataset):
     """
     Download Marktstammdatenregister (MaStR) datasets unit registry.
@@ -59,10 +57,10 @@ class mastr_data_setup(Dataset):
 
     Dump 2021-05-03
       * Source: https://sandbox.zenodo.org/record/808086
-      * Used technologies: PV plants, wind turbines, biomass, hydro plants, combustion,
-        nuclear, gsgk, storage
-      * Data is further processed in dataset :py:class:`PowerPlants
-        <egon.data.datasets.power_plants.PowerPlants>`
+      * Used technologies: PV plants, wind turbines, biomass, hydro plants,
+        combustion, nuclear, gsgk, storage
+      * Data is further processed in the :py:class:`PowerPlants
+        <egon.data.datasets.power_plants.PowerPlants>` dataset
 
     Dump 2022-11-17
       * Source: https://sandbox.zenodo.org/record/1132839
@@ -81,4 +79,12 @@ class mastr_data_setup(Dataset):
     #:
     version: str = "0.0.1"
     #:
-    tasks: Tasks = (download_mastr_data,)
+    tasks = (download_mastr_data,)
+
+    def __init__(self, dependencies):
+        super().__init__(
+            name=self.name,
+            version=self.version,
+            dependencies=dependencies,
+            tasks=self.tasks,
+        )

src/egon/data/datasets/mv_grid_districts.py

@@ -6,8 +6,6 @@
 
 """
 
-from dataclasses import dataclass
-
 from geoalchemy2.types import Geometry
 from sqlalchemy import (
     ARRAY,
@@ -23,7 +21,7 @@
 from sqlalchemy.ext.declarative import declarative_base
 
 from egon.data import db
-from egon.data.datasets import Dataset, Tasks
+from egon.data.datasets import Dataset
 from egon.data.datasets.osmtgmod.substation import EgonHvmvSubstation
 from egon.data.datasets.substation_voronoi import EgonHvmvSubstationVoronoi
 from egon.data.db import session_scope
@@ -796,7 +794,6 @@ def define_mv_grid_districts():
     MvGridDistrictsDissolved.__table__.drop(bind=engine, checkfirst=True)
 
 
-@dataclass
 class mv_grid_districts_setup(Dataset):
     """
     Maps MV grid districts to federal states and writes it to database.
@@ -882,5 +879,11 @@ class mv_grid_districts_setup(Dataset):
     name: str = "MvGridDistricts"
     #:
     version: str = "0.0.2"
-    #:
-    tasks: Tasks = define_mv_grid_districts
+
+    def __init__(self, dependencies):
+        super().__init__(
+            name=self.name,
+            version=self.version,
+            dependencies=dependencies,
+            tasks=define_mv_grid_districts,
+        )

src/egon/data/datasets/re_potential_areas/__init__.py

@@ -2,7 +2,6 @@
 potential areas for wind onshore and ground-mounted PV.
 """
 
-from dataclasses import dataclass
 from pathlib import Path
 
 from geoalchemy2 import Geometry
@@ -11,7 +10,7 @@
 import geopandas as gpd
 
 from egon.data import db
-from egon.data.datasets import Dataset, Tasks
+from egon.data.datasets import Dataset
 import egon.data.config
 
 Base = declarative_base()
@@ -121,7 +120,6 @@ def insert_data():
         )
 
 
-@dataclass
 class re_potential_area_setup(Dataset):
     """
     Downloads potential areas for PV and wind power plants from data bundle and
@@ -142,4 +140,12 @@ class re_potential_area_setup(Dataset):
     #:
     version: str = "0.0.1"
     #:
-    tasks: Tasks = (create_tables, insert_data)
+    tasks = (create_tables, insert_data)
+
+    def __init__(self, dependencies):
+        super().__init__(
+            name=self.name,
+            version=self.version,
+            dependencies=dependencies,
+            tasks=self.tasks,
+        )