WFDASH-WEBAPP/scripts/migrations/2026-07-06-purchase-orders-documents-field.ts

/**
 * Migration — consolidate PO documents onto `purchase_orders`.
 *
 * Adds two fields to the `purchase_orders` collection:
 *   - `documents`        (file, multi)        — the latest PO PDF + history.
 *   - `documents_history` (json)              — read-only audit log of
 *                                                every regen (version,
 *                                                generatedAt, generatedBy,
 *                                                reason).
 *
 * Why consolidate:
 *   Today the PO PDF lives in a separate `purchase_order_pdfs` collection
 *   (one row per regen, status="current" flag for the latest). Every
 *   Send PO flow does THREE PB round-trips to fetch the attachment:
 *     1. fetch PO
 *     2. fetch latest row from `purchase_order_pdfs` for that PO
 *     3. fetch the file bytes via the proxy
 *   Consolidating onto `purchase_orders.documents` (file, multi) shrinks
 *   this to TWO round-trips — and lets the webhook payload include the
 *   filename directly without a separate query. `documents_history`
 *   keeps the per-regen audit trail without a separate collection.
 *
 * Backfill:
 *   For every PO that already has `purchase_order_pdfs` rows, copy:
 *     - `documents`        ← the file from the row with status="current"
 *                            (or the highest-version row if no "current")
 *     - `documents_history` ← array of `{version, file, generatedAt,
 *                              generatedBy, reason}` for ALL rows
 *                            (ordered by version ASC), so future regens
 *                            preserve the audit trail.
 *
 * Idempotent — skips the field-add when fields already exist, and the
 * backfill is a no-op when documents is already populated.
 *
 * Run with:
 *   npx tsx scripts/migrations/2026-07-06-purchase-orders-documents-field.ts
 *
 * Archive step (separate, run after this is verified):
 *   After the dashboard reads from `documents` everywhere, mark the
 *   `purchase_order_pdfs` collection `is_active=false` in PB admin so
 *   new writes are blocked. The collection can stay around for audit
 *   queries (the file proxy still whitelists it).
 */
/* eslint-disable */
import PocketBase from "pocketbase";
import { readFileSync } from "node:fs";
import { resolve } from "node:path";

function loadEnvFile(name: string) {
  try {
    const raw = readFileSync(resolve(process.cwd(), name), "utf8");
    for (const line of raw.split("\n")) {
      const trimmed = line.trim();
      if (!trimmed || trimmed.startsWith("#")) continue;
      const m = trimmed.match(/^([A-Z0-9_]+)\s*=\s*(.+)$/i);
      if (m && !process.env[m[1]]) {
        process.env[m[1]] = m[2].replace(/^["']|["']$/g, "").trim();
      }
    }
  } catch {
    /* fine */
  }
}
loadEnvFile(".env");
loadEnvFile(".env.local");

const PO_COLLECTION = "purchase_orders";
const PDF_COLLECTION = "purchase_order_pdfs";

interface PbField {
  name: string;
  type: string;
  required?: boolean;
  options?: Record<string, unknown>;
  maxSelect?: number | null;
}

interface PdfRow {
  id: string;
  purchase_order: string;
  file: string;
  version: number;
  status: string;
  generated_at: string;
  generated_by: string;
  notes: string;
}

interface PoRow {
  id: string;
  documents?: string[];
  documents_history?: Array<{
    version: number;
    file: string;
    generatedAt: string;
    generatedBy: string;
    reason: string;
    status: string;
  }>;
}

async function main() {
  const url = process.env.NEXT_PUBLIC_POCKETBASE_URL;
  const email = process.env.PB_ADMIN_EMAIL;
  const password = process.env.PB_ADMIN_PASSWORD;
  if (!url || !email || !password) {
    throw new Error(
      "Set NEXT_PUBLIC_POCKETBASE_URL, PB_ADMIN_EMAIL, PB_ADMIN_PASSWORD.",
    );
  }

  const pb = new PocketBase(url);
  await pb.collection("_superusers").authWithPassword(email, password);
  console.log(`[po-documents] Authenticated against ${url}`);

  // ─── 1. Add fields to purchase_orders ─────────────────────────────
  const poCol = await pb.collections.getOne(PO_COLLECTION);
  const poFields = (poCol.fields as PbField[]).slice();

  const additions: PbField[] = [];
  if (!poFields.some((f) => f.name === "documents")) {
    additions.push({
      name: "documents",
      type: "file",
      required: false,
      options: {
        maxSelect: 20,
        maxSize: 10485760, // 10 MB per file — most PO PDFs are <500KB
        mimeTypes: ["application/pdf", "image/png", "image/jpeg"],
        // PB v0.22 needs these for file fields even though it's
        // conceptually `[]` (no thumbs) for headless server usage.
        thumbs: [],
      },
    });
    console.log("[po-documents] documents field: ADD");
  } else {
    console.log("[po-documents] documents field: skip (already exists)");
  }
  if (!poFields.some((f) => f.name === "documents_history")) {
    additions.push({
      name: "documents_history",
      type: "json",
      required: false,
      options: { maxSize: 50000 },
    });
    console.log("[po-documents] documents_history field: ADD");
  } else {
    console.log("[po-documents] documents_history field: skip (already exists)");
  }

  if (additions.length > 0) {
    poFields.push(...additions);
    await pb.collections.update(poCol.id, { fields: poFields });
    console.log(
      `[po-documents] purchase_orders — added ${additions.length} field(s)`,
    );
  }

  // ─── 2. Backfill documents + documents_history from purchase_order_pdfs ─
  // PB file fields don't accept plain filename strings — they need an
  // actual file upload via FormData. So for each PO with PDF rows we
  // need to:
  //   (a) download each PDF's bytes from PB (admin client URL)
  //   (b) re-upload to purchase_orders.documents via FormData
  //   (c) populate documents_history with the metadata (no file, just JSON)
  const pdfList = await pb.collection(PDF_COLLECTION).getFullList<PdfRow>({
    sort: "purchase_order,version",
  });
  console.log(`[po-documents] Found ${pdfList.length} purchase_order_pdfs rows`);

  // Group by purchase_order
  const byPo = new Map<string, PdfRow[]>();
  for (const row of pdfList) {
    const arr = byPo.get(row.purchase_order) ?? [];
    arr.push(row);
    byPo.set(row.purchase_order, arr);
  }
  console.log(`[po-documents] Grouped into ${byPo.size} PO(s)`);

  let migrated = 0;
  let skipped = 0;
  let failed = 0;

  for (const [poId, rows] of byPo) {
    try {
      // Skip if PO already has documents (re-run safety).
      const current = await pb
        .collection(PO_COLLECTION)
        .getOne<PoRow>(poId, { fields: "id,documents,documents_history" });

      if (
        Array.isArray(current.documents) &&
        current.documents.length > 0
      ) {
        skipped++;
        continue;
      }

      // Latest = status="current" first, else highest version.
      const sorted = rows.slice().sort((a, b) => {
        if (a.status === "current" && b.status !== "current") return -1;
        if (b.status === "current" && a.status !== "current") return 1;
        return b.version - a.version;
      });
      const latest = sorted[0];

      // Build the history array (asc by version — newest at the END).
      const history = rows
        .slice()
        .sort((a, b) => a.version - b.version)
        .map((r) => ({
          version: r.version,
          file: r.file,
          generatedAt: r.generated_at || "",
          generatedBy: r.generated_by || "",
          reason: r.notes || "",
          status: r.status,
        }));

      // Download the latest PDF bytes via admin PB file URL.
      const downloadUrl = pb.files.getURL(
        { collectionId: PDF_COLLECTION, id: latest.id },
        latest.file,
      );
      const fileRes = await fetch(downloadUrl, {
        headers: { Authorization: pb.authStore.token ?? "" },
      });
      if (!fileRes.ok) {
        throw new Error(
          `PDF download failed: HTTP ${fileRes.status} on ${downloadUrl}`,
        );
      }
      const fileBlob = await fileRes.blob();
      // PB expects a File object, not a Blob — wrap with the original name.
      const file = new File([fileBlob], latest.file, {
        type: fileRes.headers.get("content-type") || "application/pdf",
      });

      // Re-upload to purchase_orders.documents (file, multi — array of File).
      // We only migrate the LATEST file (history captures the rest as
      // metadata; if the operator ever wants to download an older
      // version, they still can via the legacy `purchase_order_pdfs`
      // collection until we archive it).
      const formData = new FormData();
      formData.append("documents", file);
      const updated = await pb.collection(PO_COLLECTION).update<PurchaseOrder>(
        poId,
        formData,
      );

      // PB appends a random suffix to file uploads to prevent
      // collisions (e.g. `foo.pdf` → `foo_uc56p4ttjl.pdf`). Read back
      // the actual stored filename from the response and update the
      // history entry to match — otherwise the download URL won't
      // resolve. Same applies to every entry's `file` field.
      const actualStoredFiles: string[] = Array.isArray(updated.documents)
        ? updated.documents
        : updated.documents
          ? [updated.documents]
          : [];
      const lastStoredFile = actualStoredFiles[actualStoredFiles.length - 1];

      const syncedHistory = history.map((h, idx) =>
        // The LAST history entry corresponds to the LAST stored file.
        // For pre-migration data we can only sync that one reliably —
        // older versions stay archived in the legacy collection.
        idx === history.length - 1 && lastStoredFile
          ? { ...h, file: lastStoredFile }
          : h,
      );

      // Then write documents_history separately (it's JSON, not a file).
      await pb.collection(PO_COLLECTION).update(poId, {
        documents_history: syncedHistory,
      });

      migrated++;
      console.log(
        `[po-documents] PO ${poId}: ${rows.length} PDF row(s) → documents=[${latest.file}] (${fileBlob.size}B), history[${history.length}]`,
      );
    } catch (err) {
      failed++;
      console.warn(`[po-documents] PO ${poId}: failed to migrate:`, err);
    }
  }

  console.log(
    `[po-documents] Backfill complete — migrated=${migrated}, skipped=${skipped}, failed=${failed}`,
  );
  console.log("[po-documents] Migration complete.");
}

main().catch((err) => {
  console.error("[po-documents] Migration failed:", err);
  process.exit(1);
});