index.ts

import fs from "node:fs";
import { parse } from "csv-parse";
import FuzzySet from "fuzzyset";
import { z } from "zod";

const __dirname = new URL(".", import.meta.url).pathname;

const Row = z.object({
  c1: z.string(),
  c2: z.string(),
  c3: z.string(),
  c4: z.string(),
  c5: z.string(),
});

const processFile = async () => {
  const records = [];
  const parser = fs.createReadStream(`${__dirname}/input.csv`).pipe(parse());
  for await (const [c1, c2, c3, c4, c5] of parser) {
    const record = Row.parse({
      c1,
      c2,
      c3,
      c4,
      c5,
    });
    records.push(record);
  }
  return records;
};

function toCSV(data: any[]) {
  const csvRows = [];
  // Get the headers
  const headers = Object.keys(data[0]);
  csvRows.push(headers.join(","));

  // Map the data to CSV format
  for (const row of data) {
    const values = headers.map((header) => {
      const escaped = ("" + row[header]).replace(/"/g, '\\"');
      return `"${escaped}"`;
    });
    csvRows.push(values.join(","));
  }

  const csv = csvRows.join("\n");
  fs.writeFileSync("output.csv", csv);
}

(async () => {
  const records = await processFile();

  const results = [];
  for (const record of records) {
    const result = {
      ...record,
      c1_c2: compareStrings(record.c1, record.c2),
      c1_c3: compareStrings(record.c1, record.c3),
      c1_c4: compareStrings(record.c1, record.c4),
      c1_c5: compareStrings(record.c1, record.c5),
      c2_c3: compareStrings(record.c2, record.c3),
      c2_c4: compareStrings(record.c2, record.c4),
      c2_c5: compareStrings(record.c2, record.c5),
      c3_c4: compareStrings(record.c3, record.c4),
      c3_c5: compareStrings(record.c3, record.c5),
      c4_c5: compareStrings(record.c4, record.c5),
    };
    results.push(result);
  }
  toCSV(results);
})();

function compareStrings(aString: string, bString: string) {
  const a = aString.split(",").map((s) => s.trim());
  const b = bString.split(",").map((s) => s.trim());
  const aSet = FuzzySet(a);
  const results = b
    .map((s) => {
      const res = aSet.get(s);
      return res ? res[0][0] : 0;
    })
    .sort((a, b) => b - a);
  return toPercentage(results[0]);
}

function toPercentage(value: number): string {
  return `${(value * 100).toFixed(0)}`;
}