refactor create_routes to use fct_monthly_routes

Author: tiffanychu90

open_data/create_routes_data2.py

@@ -0,0 +1,91 @@
+"""
+Create `ca_transit_routes` to publish to Geoportal.
+"""
+
+import geopandas as gpd
+import google.auth
+import pandas as pd
+from calitp_data_analysis import utils
+from create_stops_data2 import prep_crosswalk
+from update_vars import OPEN_DATA_GCS, analysis_month
+
+credentials, _ = google.auth.default()
+
+
+def prep_route_shapes(gdf: gpd.GeoDataFrame) -> gpd.GeoDataFrame:
+    """
+    Aggregate day_type (weekday/Sat/Sun) to all shape-route
+    combinations that month.
+    """
+    route_group_cols = ["name", "shape_id", "route_name"]
+    # Group across day_types
+    shape_geom = gdf[route_group_cols + ["geometry"]].drop_duplicates()
+
+    gdf2 = (
+        gdf.groupby(route_group_cols + ["month_first_day"])
+        .agg(
+            {
+                "n_trips": "sum",
+            }
+        )
+        .reset_index()
+    )
+
+    gdf3 = pd.merge(shape_geom, gdf2, on=route_group_cols, how="inner")
+
+    # should route_name be split to show route_id and route_name?
+    return gdf3
+
+
+def rename_route_columns(gdf: gpd.GeoDataFrame) -> gpd.GeoDataFrame:
+    """
+    Put all the renaming here.
+    """
+    keep_cols = [
+        # from shapes
+        "name",
+        "route_name",
+        "shape_id",
+        "route_type",
+        "n_trips",
+        "geometry",
+        # from trips
+        "route_ids",
+        # calculate
+        "on_shn"  # dummy,
+        "shn_districts",
+        "pct_route_on_hwy_all_districts",
+        # from bridge
+        "analysis_name",
+        # schedule_source_record_id
+    ]
+
+    gdf = gdf[keep_cols].rename(columns={})
+
+    return gdf
+
+
+def publish_routes(analysis_month: str):
+    routes = gpd.read_parquet(
+        f"{OPEN_DATA_GCS}routes_{analysis_month}.parquet",
+        storage_options={"token": credentials.token},
+    ).pipe(prep_route_shapes)
+
+    crosswalk = pd.read_parquet(f"{OPEN_DATA_GCS}bridge_gtfs_analysis_name_x_ntd.parquet").pipe(prep_crosswalk)
+
+    routes2 = pd.merge(
+        routes, crosswalk.rename(columns={"schedule_gtfs_dataset_name": "name"}), on=["name"], how="inner"
+    )
+
+    # TODO1 route_id unparsed from route_name?
+    # TODO2 (add SHN derived columns): use existing function and work it into this
+    # TODO3 (standardize columns for Geoportal): pipe through rename_routecolumns
+
+    return routes2
+
+
+if __name__ == "__main__":
+
+    routes = publish_routes(analysis_month)
+
+    utils.geoparquet_gcs_export(routes, OPEN_DATA_GCS, f"export/ca_transit_routes_{analysis_month}")