baron-sso/adminfront/src/features/tenants/utils/tenantCsvImport.ts

import type { TenantSummary } from "../../../lib/adminApi";

export type TenantCSVRow = {
  rowNumber: number;
  tenantId: string;
  name: string;
  type: string;
  parentTenantId: string;
  slug: string;
  memo: string;
  emailDomain: string;
};

export type TenantImportCandidate = {
  tenantId: string;
  name: string;
  slug: string;
  score: number;
  reason: "exact_name" | "exact_slug" | "similar_name";
};

export type TenantImportPreviewRow = {
  row: TenantCSVRow;
  candidates: TenantImportCandidate[];
  defaultTenantId: string;
};

const importHeaders = [
  "tenant_id",
  "name",
  "type",
  "parent_tenant_id",
  "slug",
  "memo",
  "email_domain",
];

const headerAliases: Record<string, keyof TenantCSVRow> = {
  id: "tenantId",
  tenantid: "tenantId",
  tenant_id: "tenantId",
  name: "name",
  type: "type",
  parentid: "parentTenantId",
  parent_id: "parentTenantId",
  parenttenantid: "parentTenantId",
  parent_tenant_id: "parentTenantId",
  slug: "slug",
  memo: "memo",
  description: "memo",
  "email-domain": "emailDomain",
  emaildomain: "emailDomain",
  email_domain: "emailDomain",
  domain: "emailDomain",
  domains: "emailDomain",
};

export function parseTenantCSV(text: string): TenantCSVRow[] {
  const records = parseCSV(text.replace(/^\uFEFF/, ""));
  if (records.length === 0) return [];

  const header = new Map<keyof TenantCSVRow, number>();
  records[0].forEach((column, index) => {
    const normalized = normalizeHeader(column);
    const key = headerAliases[normalized];
    if (key) header.set(key, index);
  });

  return records.slice(1).flatMap((record, index) => {
    if (record.every((value) => value.trim() === "")) return [];
    const value = (key: keyof TenantCSVRow) => {
      const columnIndex = header.get(key);
      if (columnIndex === undefined) return "";
      return (record[columnIndex] ?? "").trim();
    };

    return {
      rowNumber: index + 2,
      tenantId: value("tenantId"),
      name: value("name"),
      type: value("type"),
      parentTenantId: value("parentTenantId"),
      slug: value("slug"),
      memo: value("memo"),
      emailDomain: value("emailDomain"),
    };
  });
}

export function buildTenantImportPreview(
  rows: TenantCSVRow[],
  tenants: TenantSummary[],
): TenantImportPreviewRow[] {
  return rows
    .map((row) => {
      const candidates = row.tenantId ? [] : findTenantCandidates(row, tenants);
      return {
        row,
        candidates,
        defaultTenantId:
          candidates[0] && candidates[0].score >= 0.95
            ? candidates[0].tenantId
            : "",
      };
    })
    .sort((a, b) => {
      const aScore = a.candidates[0]?.score ?? 0;
      const bScore = b.candidates[0]?.score ?? 0;
      if (bScore !== aScore) return bScore - aScore;
      return a.row.rowNumber - b.row.rowNumber;
    });
}

export function serializeTenantImportCSV(
  previewRows: TenantImportPreviewRow[],
  selectedTenantIds: Record<number, string>,
) {
  const lines = [importHeaders];
  for (const preview of [...previewRows].sort(
    (a, b) => a.row.rowNumber - b.row.rowNumber,
  )) {
    const selectedTenantId = selectedTenantIds[preview.row.rowNumber] ?? "";
    lines.push([
      preview.row.tenantId || selectedTenantId,
      preview.row.name,
      preview.row.type,
      preview.row.parentTenantId,
      preview.row.slug,
      preview.row.memo,
      preview.row.emailDomain,
    ]);
  }
  return lines.map(formatCSVRecord).join("\n") + "\n";
}

function findTenantCandidates(
  row: TenantCSVRow,
  tenants: TenantSummary[],
): TenantImportCandidate[] {
  return tenants
    .map((tenant) => {
      const nameScore = similarity(row.name, tenant.name);
      const slugScore =
        normalizeToken(row.slug) &&
        normalizeToken(row.slug) === normalizeToken(tenant.slug)
          ? 0.98
          : 0;
      const exactName =
        normalizeToken(row.name) === normalizeToken(tenant.name);
      const score = exactName ? 1 : Math.max(slugScore, nameScore);
      const reason: TenantImportCandidate["reason"] = exactName
        ? "exact_name"
        : slugScore >= 0.98
          ? "exact_slug"
          : "similar_name";
      return {
        tenantId: tenant.id,
        name: tenant.name,
        slug: tenant.slug,
        score,
        reason,
      };
    })
    .filter((candidate) => candidate.score >= 0.45)
    .sort((a, b) => b.score - a.score)
    .slice(0, 3);
}

function parseCSV(text: string): string[][] {
  const rows: string[][] = [];
  let current = "";
  let row: string[] = [];
  let quoted = false;

  for (let i = 0; i < text.length; i += 1) {
    const char = text[i];
    const next = text[i + 1];

    if (char === '"' && quoted && next === '"') {
      current += '"';
      i += 1;
      continue;
    }
    if (char === '"') {
      quoted = !quoted;
      continue;
    }
    if (char === "," && !quoted) {
      row.push(current);
      current = "";
      continue;
    }
    if ((char === "\n" || char === "\r") && !quoted) {
      if (char === "\r" && next === "\n") i += 1;
      row.push(current);
      rows.push(row);
      row = [];
      current = "";
      continue;
    }
    current += char;
  }

  if (current !== "" || row.length > 0) {
    row.push(current);
    rows.push(row);
  }

  return rows;
}

function formatCSVRecord(record: string[]) {
  return record
    .map((value) => {
      if (!/[",\r\n]/.test(value)) return value;
      return `"${value.replaceAll('"', '""')}"`;
    })
    .join(",");
}

function normalizeHeader(value: string) {
  return value.trim().toLowerCase().replaceAll(" ", "_");
}

function normalizeToken(value: string) {
  return value
    .trim()
    .toLowerCase()
    .replace(/[\s_-]+/g, "")
    .replace(/[^\p{L}\p{N}]/gu, "");
}

function similarity(left: string, right: string) {
  const a = normalizeToken(left);
  const b = normalizeToken(right);
  if (!a || !b) return 0;
  if (a === b) return 1;
  if (a.includes(b) || b.includes(a)) {
    return Math.min(a.length, b.length) / Math.max(a.length, b.length);
  }

  const distance = levenshtein(a, b);
  return 1 - distance / Math.max(a.length, b.length);
}

function levenshtein(left: string, right: string) {
  const previous = Array.from({ length: right.length + 1 }, (_, i) => i);
  const current = Array.from({ length: right.length + 1 }, () => 0);

  for (let i = 1; i <= left.length; i += 1) {
    current[0] = i;
    for (let j = 1; j <= right.length; j += 1) {
      const cost = left[i - 1] === right[j - 1] ? 0 : 1;
      current[j] = Math.min(
        current[j - 1] + 1,
        previous[j] + 1,
        previous[j - 1] + cost,
      );
    }
    previous.splice(0, previous.length, ...current);
  }

  return previous[right.length];
}