Merge remote-tracking branch 'origin/main' into validate_outputs

config/base.toml

@@ -1,3 +1,14 @@
 [parameters]
 seed = 0
 region = "leeds"
+number_of_households = 10000
+zone_id = "OA21CD"
+travel_times = true # Only set to true if you have travel time matrix at the level specified in boundary_geography
+boundary_geography = "OA"
+
+
+[work_assignment]
+use_percentages = true
+weight_max_dev = 0.2
+weight_total_dev = 0.8
+max_zones = 8

config/base_500.toml

@@ -0,0 +1,14 @@
+[parameters]
+seed = 0
+region = "leeds"
+number_of_households = 500
+zone_id = "OA21CD"
+travel_times = true # Only set to true if you have travel time matrix at the level specified in boundary_geography
+boundary_geography = "OA"
+
+
+[work_assignment]
+use_percentages = true
+weight_max_dev = 0.2
+weight_total_dev = 0.8
+max_zones = 8

config/base_5000.toml

@@ -0,0 +1,13 @@
+[parameters]
+seed = 0
+region = "leeds"
+number_of_households = 5000
+zone_id = "OA21CD"
+travel_times = true # Only set to true if you have travel time matrix at the level specified in boundary_geography
+boundary_geography = "OA"
+
+[work_assignment]
+use_percentages = true
+weight_max_dev = 0.2
+weight_total_dev = 0.8
+max_zones = 8

config/base_all.toml

@@ -0,0 +1,13 @@
+[parameters]
+seed = 0
+region = "leeds"
+zone_id = "OA21CD"
+travel_times = true # Only set to true if you have travel time matrix at the level specified in boundary_geography
+boundary_geography = "OA"
+
+
+[work_assignment]
+use_percentages = false
+weight_max_dev = 0.0
+weight_total_dev = 1.0
+max_zones = 4

pyproject.toml

@@ -42,6 +42,7 @@ click = "^8.1.7"
 tomlkit = "^0.13.0"
 cml-pam = "0.3.2"
 gdal = "<=3.8.4"
+pandera = "^0.20.4"
 
 [tool.poetry.dev-dependencies]
 pytest = ">= 6"

scripts/0_preprocess_inputs.py

@@ -0,0 +1,78 @@
+import geopandas as gpd
+import pandas as pd
+from uatk_spc import Reader
+
+import acbm
+from acbm.cli import acbm_cli
+from acbm.config import load_config
+from acbm.logger_config import preprocessing_logger as logger
+from acbm.preprocessing import edit_boundary_resolution
+
+
+@acbm_cli
+def main(config_file):
+    config = load_config(config_file)
+    config.init_rng()
+    region = config.region
+    # Pick a region with SPC output saved
+    spc_path = acbm.root_path / "data/external/spc_output/raw/"
+
+    # ----- BOUNDARIES
+    logger.info("Preprocessing Boundary Layer")
+
+    ## Read in the boundary layer for the whole of England
+
+    logger.info("1. Reading in the boundary layer for the whole of England")
+
+    boundaries = gpd.read_file(
+        acbm.root_path / "data/external/boundaries/oa_england.geojson"
+    )
+
+    boundaries = boundaries.to_crs(epsg=4326)
+
+    ## --- Dissolve boundaries if resolution is MSOA
+
+    boundary_geography = config.parameters.boundary_geography  # can only be OA or MSOA
+    logger.info(f"2. Dissolving boundaries to {boundary_geography} level")
+
+    boundaries = edit_boundary_resolution(
+        study_area=boundaries, geography=boundary_geography, zone_id=config.zone_id
+    )
+
+    ## --- Filter to study area
+    # we filter using msoa21cd values, which exist regardless of the boundary resolution
+
+    logger.info("3. Filtering boundaries to specified study area")
+
+    # Step 1: Get zones from SPC (these will be 2011 MSOAs)
+    spc = Reader(spc_path, region, backend="pandas")
+    zones_in_region = list(spc.info_per_msoa.keys())
+
+    # Step 2: Filter boundaries to identified zones
+
+    # a) get MSOA11CD to MSOA21CD lookup
+    msoa_lookup = pd.read_csv(
+        acbm.root_path
+        / "data/external/MSOA_2011_MSOA_2021_Lookup_for_England_and_Wales.csv"
+    )
+    # Filter msoa_lookup to include only rows where MSOA11CD is in zones_in_region
+    msoa_lookup_filtered = msoa_lookup[msoa_lookup["MSOA11CD"].isin(zones_in_region)]
+    # Extract the corresponding MSOA21CD values
+    msoa21cd_values = msoa_lookup_filtered["MSOA21CD"].tolist()
+
+    # b) filter boundaries to include only rows where MSOA21CD is in msoa21cd_values
+    boundaries_filtered = boundaries[boundaries["MSOA21CD"].isin(msoa21cd_values)]
+
+    ## Save the output as parquet
+    logger.info(
+        f"4. Saving the boundaries to {acbm.root_path / 'data/external/boundaries/'} path"
+    )
+
+    boundaries_filtered.to_file(
+        acbm.root_path / "data/external/boundaries/study_area_zones.geojson",
+        driver="GeoJSON",
+    )
+
+
+if __name__ == "__main__":
+    main()

scripts/1_prep_synthpop.py

@@ -1,28 +1,18 @@
-import click
-import numpy as np
-import tomlkit
 from uatk_spc.builder import Builder
 
 import acbm
+from acbm.cli import acbm_cli
+from acbm.config import load_config
 
 
-@click.command()
-# TODO: add override for case when seed provided from CLI
-# @click.option("--seed", default=1, help="Seed for random state", type=int)
-@click.option("--config", prompt="Filepath relative to repo root of config", type=str)
-def main(config):
-    # Read config
-    with open(acbm.root_path / config, "rb") as f:
-        config_dict = tomlkit.load(f)
-    seed = config_dict["parameters"]["seed"]
-    region = config_dict["parameters"]["region"]
-
-    # Seed RNG
-    np.random.seed(seed)
+@acbm_cli
+def main(config_file):
+    config = load_config(config_file)
+    config.init_rng()
+    region = config.region
 
     # Pick a region with SPC output saved
     path = acbm.root_path / "data/external/spc_output/raw/"
-    region = "leeds"
 
     # Add people and households
     spc_people_hh = (