Generate butterfly data in CSV

generate_butterflies.csv

# This gist generates a specified number of CSV files 
# Each CSV file contains a varying number of rows about butterflies

import os
import random
import pandas as pd

output_dir = "your/output/directory"

# Step 1: Generate Butterfly Species Data
def generate_csv_files(output_dir, num_files=25):
   # Ensure output directory exists
   try:
       if not os.path.exists(output_dir):
           os.makedirs(output_dir)
       print(f"Output directory: {os.path.abspath(output_dir)}")
   except Exception as e:
       print(f"Failed to create output directory: {e}")
       return

   # Species details
   species_data = [
       {"species": "Papilio machaon", "native_habitat": "Europe"},
       {"species": "Danaus plexippus", "native_habitat": "North America"},
       {"species": "Pieris rapae", "native_habitat": "Asia"},
       {"species": "Vanessa atalanta", "native_habitat": "Europe"},
       {"species": "Morpho menelaus", "native_habitat": "South America"},
       {"species": "Heliconius charithonia", "native_habitat": "Central America"},
   ]

   colors = ["black", "orange", "white", "blue", "yellow"]

   # Generate CSV files
   for file_index in range(1, num_files + 1):
       num_rows = random.randint(1000, 10000)
       rows = []
       for _ in range(num_rows):
           butterfly = random.choice(species_data)
           row = {
               "species": butterfly["species"],
               "native_habitat": butterfly["native_habitat"],
               "age_days": random.randint(1, 111),
               "main_color": random.choice(colors),
           }
           rows.append(row)

       # Save file
       file_path = os.path.join(output_dir, f"butterflies_{file_index}.csv")
       try:
           if not rows:
               print(f"No data generated for file {file_index}. Skipping...")
               continue
           pd.DataFrame(rows).to_csv(file_path, index=False)
           print(f"Generated {file_path} with {num_rows} rows.")
       except Exception as e:
           print(f"Failed to write {file_path}: {e}")


# Run the generator
generate_csv_files(output_dir)