From 65bd9480e70fa7b751975c83fad8b40a40f74c65 Mon Sep 17 00:00:00 2001 From: Sebastian Lenzlinger <74497638+sebaschi@users.noreply.github.com> Date: Sun, 3 Dec 2023 12:35:38 +0100 Subject: [PATCH] Load accident file from internet if not present. --- src/integrate.py | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/src/integrate.py b/src/integrate.py index c2ccfe1..bdc39d5 100644 --- a/src/integrate.py +++ b/src/integrate.py @@ -65,8 +65,9 @@ def process_miv_data(): return cleaned_miv_df -def process_accident_data(): - +def process_accident_data(file_present: bool): + if not file_present: + du.process_urls(data_dir, accident_file_url) acc_df_unified = du.load_dataframes_from_geojson_files(data_dir, accident_file_u_string) acc_cols_to_keep = ['AccidentUID', 'AccidentHour', 'AccidentYear', 'AccidentWeekDay_en', 'AccidentType', 'AccidentSeverityCategory', 'AccidentInvolvingPedestrian', 'AccidentInvolvingBicycle', @@ -77,6 +78,6 @@ def process_accident_data(): if __name__ == '__main__': - acc_df = process_accident_data() + acc_df = process_accident_data(False) print(acc_df.dtypes) print(acc_df.head(100))