fuzzymatch.py

fuzzymatch.py

import pandas as pd
from rapidfuzz import process, fuzz

def main():
  # Load CSV files
  salesforce_df = pd.read_csv("<...>")
  brands_df = pd.read_csv("<...>")

  # Convert brand names into a dictionary for efficient lookup
  brand_dict = dict(zip(brands_df["brand_name"], brands_df["brand_id"]))
  brand_names = list(brand_dict.keys())  # List of brand names for matching

  # Function to find best match
  def find_best_match(account_name):
      best_match, score, *_ = process.extractOne(str(account_name), brand_names, scorer=fuzz.token_sort_ratio)
      return best_match, round(score / 10, 1)

  # Apply fuzzy matching across all rows at once (vectorized)
  salesforce_df[["Best Match Brand Name", "Match Score"]] = salesforce_df["Account Name"].apply(
      lambda name: pd.Series(find_best_match(name))
  )

  # Merge results with brand IDs
  salesforce_df["Brand ID"] = salesforce_df["Best Match Brand Name"].map(brand_dict)

  # Filter for pretty good matches only
  matched_df = salesforce_df[salesforce_df["Match Score"] >= 9]

  # Select relevant columns
  matched_df = matched_df[[
      "Account ID", "Account Name", "Brand ID", "Best Match Brand Name"
  ]]

  # Rename columns
  matched_df.columns = [
      "Salesforce Account ID", "Salesforce Account Name", "Brand ID", "Brand Name"
  ]

  # Save the results to a CSV file
  matched_df.to_csv("matched_salesforce_brands.csv", index=False)

  print("✅ Matching complete! Check 'matched_salesforce_brands.csv'.")

if __name__ == "__main__":
    main()