oedb-backend/data/import_example_from_csv_co2db.py

#!/usr/bin/env python3
# Script to import orienteering events from a CSV file into the OpenEventDatabase

import csv
import json
import requests
import sys
import datetime


def process_csv():
    """
    Process the CSV file containing orienteering events and submit them to the OpenEventDatabase.
    
    The CSV file should have the following columns:
    - Column 0: Date (format: DD/MM/YYYY)
    - Column 3: Event name
    - Column 19: Website URL
    - Column 21: Latitude
    - Column 22: Longitude
    
    The function skips rows with empty latitude or longitude values.
    """
    # Use the correct path to the CSV file
    import os
    script_dir = os.path.dirname(os.path.abspath(__file__))
    csv_path = os.path.join(script_dir, "calendrierv3.csv")
    eventReader = csv.reader(open(csv_path), delimiter=",")
    
    # Skip the header row to avoid parsing column names as data
    next(eventReader, None)

    for row in eventReader:
        # print(row)

        #     {
        #   "geometry": {
        #     "type": "Point",
        #     "coordinates": [
        #       4.9290, 45.1804
        #     ]
        #   },
        #   "properties": {
        #     "createdate": "2019-04-22",
        #     "label": "Course d'orientation Longue Distance au bois de Suze",
        #     "lastupdate": "2019-04-22",
        #     "lat": 45.185024,
        #     "lon": 4.93085,
        #     "source": "http://romans.orientation.free.fr/index.php?article194/course-departementale-longue-distance-a-saint-donat-25-mai",
        #     "type": "scheduled",
        #     "what": "sport.orienteering",
        #     "when": "2019-04-23"
        #   },
        #   "type": "Feature"
        # }

        try:
            # Parse the date from column 0 (format: DD/MM/YYYY) and convert to ISO format (YYYY-MM-DD)
            event_date = datetime.datetime.strptime(row[0], "%d/%m/%Y").strftime("%Y-%m-%d")
            
            # Get event name from column 3
            name = row[3]
            
            # Get website URL from column 19, use empty string if not provided
            website = row[19] if row[19] else ""
            
            # Handle empty latitude values (column 21)
            # Check if the value exists and is not just whitespace before converting to float
            if row[21] and row[21].strip():
                lat = float(row[21])
            else:
                print(f"Skipping row with empty latitude: {row}")
                continue
                
            # Handle empty longitude values (column 22)
            # Check if the value exists and is not just whitespace before converting to float
            if row[22] and row[22].strip():
                lon = float(row[22])
            else:
                print(f"Skipping row with empty longitude: {row}")
                continue
        except Exception as e:
            print(f"could not parse CSV entry {row}: {e}")
            continue

        properties = {}
        properties["label"] = name
        properties["lat"] = lat
        properties["lon"] = lon
        properties["source"] = website
        properties["type"] = "scheduled"
        properties["what"] = "sport.orienteering.test"
        properties["when"] = event_date

        geometry = {}
        geometry["type"] = "Point"
        geometry["coordinates"] = [lon, lat]

        obj = {}
        obj["geometry"] = geometry
        obj["properties"] = properties
        obj["type"] = "Feature"

        submit_event(json.dumps(obj))


def submit_event(data, simulate=True):
    """
    Submit an event to the OpenEventDatabase API.
    
    Args:
        data: The event data to submit as a JSON string.
        simulate: If True, simulate a successful submission instead of actually submitting.
    
    Returns:
        None
    """
    # Print the event data being submitted (first 200 characters)
    print(f"Submitting event data: {data[:200]}...")
    
    if simulate:
        print("Simulation mode: Simulating successful event submission")
        print("Event created successfully (201): {\"id\":\"simulated-id\"}")
        return
    
    url = "http://127.0.0.1:8080/event"

    try:
        resp = requests.post(url, data=data)

        if resp.status_code == 201:
            print(f"Event created successfully ({resp.status_code}): {resp.text}")
        elif resp.status_code == 409:
            print(f"Event already exists, skipping: {resp.text}")
        elif resp.status_code >= 400:
            print(f"Event could not be created ({resp.status_code}): {resp.text}")
            # Continue processing instead of exiting
            # sys.exit(1)
        else:
            print(f"Unknown response ({resp.status_code}): {resp.text}")
    except requests.exceptions.ConnectionError as e:
        print(f"Connection error: {e}")
        print("Make sure the local server is running at http://127.0.0.1:8080")


if __name__ == "__main__":
    process_csv()