Download_Images.py

#
# MASTER SCRIPT TO CLEAN UP ADDRESS FILE AND DOWNLOAD IMAGES VIA GOOGLE API
# for each adress in the address list we download the google street view images
#
import sys

sys.path.insert(0, "./Utils")
import argparse
import os
from Location import (
    get_parent_dir,
    ignore_non_int,
    seperate_house_numbers,
    tidy_split,
    find_similar_clusters,
    download_images,
)
import pandas as pd

pd.options.mode.chained_assignment = None  # default='warn'

result_folder = os.path.join(get_parent_dir(n=1), "Data", "Street_View_Images")
result_file = os.path.join(result_folder, "results.csv")
address_file = os.path.join(
    os.path.dirname(os.path.abspath(__file__)), "Locations", "SF.csv"
)
test_address_file = os.path.join(
    os.path.dirname(os.path.abspath(__file__)), "Locations", "Test_SF.csv"
)
FLAGS = None

if __name__ == "__main__":
    # surpress any inhereted default values
    parser = argparse.ArgumentParser(argument_default=argparse.SUPPRESS)
    """
    Command line options
    """
    parser.add_argument(
        "--address_file",
        type=str,
        default=address_file,
        help="absolute path to address_file. Please make sure the table structure is correct.",
    )

    parser.add_argument(
        "--result_folder",
        type=str,
        default=result_folder,
        help="absolute path to output directory",
    )
    parser.add_argument(
        "--result_file",
        type=str,
        default=result_file,
        help="absolute path to result file",
    )

    parser.add_argument(
        "--city",
        type=str,
        default="SAN FRANCISCO",
        help="The city of the provided addresses",
    )

    parser.add_argument(
        "--G_API_key",
        type=str,
        default="",
        help="Google maps API key. Key is required to download images. Go to https://console.cloud.google.com/ to obtain a key for up to 30,000 free requests.",
    )

    parser.add_argument(
        "--test",
        default=False,
        action="store_true",
        help="Test routine: run on addresses in ../Locations/Test_SF.csv",
    )

    parser.add_argument(
        "--pitch",
        type=int,
        default=10,
        help="Pitch angle of Google Street View Images.",
    )

    FLAGS = parser.parse_args()

    if FLAGS.test:
        df_vul = seperate_house_numbers(
            pd.read_csv(test_address_file), column="PROPERTY", city=FLAGS.city
        )
    else:
        df_vul = seperate_house_numbers(
            pd.read_csv(address_file), column="PROPERTY", city=FLAGS.city
        )
    df_vul = tidy_split(df_vul, "NUMBER", sep="+")
    df_vul = ignore_non_int(df_vul)
    df_vul.dropna(axis=0, subset=["TIER"], inplace=True)
    df_vul_clusters = find_similar_clusters(df_vul, column_name="STATUS", cutoff=0.97)

    ## Download buildings by category
    if not FLAGS.G_API_key:
        print(
            "No Google API key provided. Downloading images will only be simulated.",
            "\n",
        )
    results_df = pd.DataFrame(
        columns=["image", "full_path", "category", "tier", "class"]
    )
    for cluster in df_vul_clusters:
        print("Currently Downloading Images for", cluster[0], "Category")
        results_df = download_images(
            df_vul[df_vul["STATUS"].isin(cluster)],
            results_df,
            dir_path=result_folder,
            class_name="vulnerable",
            category=cluster[0],
            api_key=FLAGS.G_API_key,
            save_tier=False,
        )
    results_df.to_csv(result_file, index=False)