Fused MCP Agents

import fused from fused.models.udf import Header from fused.models import Schema @fused.udf def yearly_crime_amount( up_to_year: int = 2021, lat: float=49.2806, lon: float=-123.1259, buffer_amount: float = 1000, ): """ This UDF takes the lat / lon + buffer amount of any area within Vancouver and returns the number of different crimes per category that happened that year """ import datetime import geopandas as gpd import pandas as pd import shapely from shapely.geometry import Point current_year = int(datetime.datetime.now().year) list_of_years_to_process = [year for year in range(up_to_year, current_year)] print(f"{list_of_years_to_process=}") yearly_type_summaries = {} yearly_crime_location = {} @fused.cache() def getting_yearly_data(year): path = f"s3://fused-users/fused/max/maxar_ted_demo/crimedata_csv_AllNeighbourhoods_{str(year)}.csv" df = pd.read_csv(path) df["geometry"] = gpd.points_from_xy(df["X"], df["Y"], crs="EPSG:32610") gdf = gpd.GeoDataFrame(df) print(gdf.shape) gdf.to_crs(4326, inplace=True) aoi_gdf = gpd.GeoDataFrame(geometry=[Point(lon, lat)], crs="EPSG:4326") # Project to a local UTM projection for accurate buffering in meters # Get UTM zone from longitude utm_zone = int(((lon + 180) / 6) % 60) + 1 hemisphere = 'north' if lat >= 0 else 'south' utm_crs = f"+proj=utm +zone={utm_zone} +{hemisphere} +ellps=WGS84 +datum=WGS84 +units=m +no_defs" gdf_utm = aoi_gdf.to_crs(utm_crs) gdf_utm['geometry'] = gdf_utm.buffer(buffer_amount) gdf_buffered = gdf_utm.to_crs("EPSG:4326") clipped_gdf = gpd.sjoin(gdf, gdf_buffered, predicate='intersects', how='inner') # Getting stats number_of_crimes = pd.DataFrame({f"number_crimes": [clipped_gdf.shape[0]]}) grouped_by_type = clipped_gdf.groupby("TYPE").size().reset_index(name="count") # print(f"{type(grouped_by_type)=}") # print(f"{grouped_by_type=}") # print(f"{grouped_by_type.columns=}") return clipped_gdf, grouped_by_type for year in list_of_years_to_process: clipped_gdf, grouped_by_type = getting_yearly_data(year) yearly_crime_location[year] = clipped_gdf yearly_type_summaries[year] = grouped_by_type df_all_years = pd.concat( [df.assign(year=year) for year, df in yearly_type_summaries.items()], ignore_index=True ) df = pd.concat([df.assign(year=year) for year, df in yearly_type_summaries.items()], ignore_index=True) print(f"{df_all_years=}") return df_all_years