WFDASH-WEBAPP/scripts/backfill-item-intel.ts

/**
 * Backfill `ai_item_insights` for every quote_item and order_item
 * that doesn't already have the new-shape blob.
 *
 *   1. Direct CLI (this script):
 *        node --env-file=.env.local $(which npx) tsx scripts/backfill-item-intel.ts
 *
 *      The `--env-file` flag is required because the PB admin URL
 *      is captured at module-load time and the env loader can't
 *      run before the imports (ESM hoists imports above body code).
 *      Node 20+ supports `--env-file` natively.
 *
 *   2. HTTP endpoint (so the operator's n8n workflow can trigger it):
 *        POST /api/v1/items/backfill-intel
 *        Body: { "collection": "quote_items", "limit": 1000 }
 *
 * The script calls the HTTP endpoint by default so the same code path
 * is exercised in both modes. Override with `--direct` to call the
 * service in-process (skips the HTTP layer, slightly faster).
 *
 * Safety:
 *   - `--dry-run` previews the count without writing.
 *   - `--limit N` caps the slice per collection (default 1000).
 *   - Rows that already have the new-shape intel are skipped.
 */

// Load .env.local inline BEFORE any other imports (ESM hoists
// imports, so the load call has to run before the next import is
// evaluated). We use a self-invoking sync block at the top.
(() => {
  try {
    for (const f of [".env", ".env.local"]) {
      try {
        const { readFileSync } = require("node:fs");
        const { resolve } = require("node:path");
        const raw = readFileSync(resolve(process.cwd(), f), "utf8");
        for (const line of raw.split("\n")) {
          const t = line.trim();
          if (!t || t.startsWith("#")) continue;
          const m = t.match(/^([A-Z0-9_]+)\s*=\s*(.+)$/i);
          if (m && !process.env[m[1]]) {
            process.env[m[1]] = m[2].replace(/^["']|["']$/g, "").trim();
          }
        }
      } catch {}
    }
  } catch {}
})();

import { populateItemIntel } from "../src/lib/server/item-intel-service";
import { getAdminPocketBase } from "../src/lib/server/pocketbase-admin";

const PB_URL = process.env.NEXT_PUBLIC_POCKETBASE_URL || `http://127.0.0.1:809${""}`;

function isNewShape(value: unknown): boolean {
  if (!value || typeof value !== "object") return false;
  const obj = value as Record<string, unknown>;
  return ["stock_list", "supplier_history", "price_list", "recent_quote", "recent_order"].every(
    (k) => Array.isArray(obj[k]),
  );
}

async function backfill(
  collection: "quote_items" | "order_items",
  limit: number,
  dryRun: boolean,
  batchSize: number,
  thisMonthOnly: boolean,
): Promise<{ scanned: number; updated: number; skipped: number; errors: number; duration_ms: number }> {
  const t0 = Date.now();
  const pb = await getAdminPocketBase();

  // PB filter: only items whose parent (quote / order) is in-flight
  // and (optionally) created in the current month. Historical data
  // is excluded — the operator only needs intel on work that the
  // team is actively quoting / fulfilling right now.
  //
  // Active statuses per collection:
  //   quote_items → parent quote.status IN (new, claimed,
  //                 approval_pending, approved)
  //   order_items → parent order.item_status / .status IN
  //                 (pending, approval_needed, processed,
  //                  po_generated, collection_pool, assigned,
  //                  collected, picking, ready_delivery, delivered,
  //                  ready_invoice, invoiced)
  //
  // The parent filter is applied via the PB expand syntax: PB can't
  // filter on expanded fields directly, so we pull the parent IDs
  // first, then filter the child collection by `id IN (...)`.
  const ACTIVE_QUOTE_STATUSES = ["new", "claimed", "approval_pending", "approved"];
  const ACTIVE_ORDER_STATUSES = [
    "pending", "approval_needed", "processed", "po_generated",
    "collection_pool", "assigned", "collected", "picking",
    "ready_delivery", "delivered", "ready_invoice", "invoiced",
  ];

  // quote_items: filter via the parent quote (active status + this
  // month). order_items: the parent `orders.status` field is
  // often empty — the active flag lives on the child itself
  // (`item_status`). So for orders we filter the child directly
  // and skip the parent lookup.
  let childFilter: string | null = null;
  let parentIdFilter: string[] | null = null;
  const isQuote = collection === "quote_items";

  if (isQuote) {
    const dateFilter = thisMonthOnly ? buildCurrentMonthFilter("created") : "";
    const statusFilter = ACTIVE_QUOTE_STATUSES.map((s) => `status="${s}"`).join(" || ");
    const filter = `(${statusFilter})${dateFilter ? " && " + dateFilter : ""}`;
    process.stdout.write(`  parent filter: ${filter}\n`);
    const parents = await pb.collection("quotes").getFullList({ filter, fields: "id" });
    parentIdFilter = parents.map((p) => p.id);
    if (parentIdFilter.length === 0) {
      return { scanned: 0, updated: 0, skipped: 0, errors: 0, duration_ms: Date.now() - t0 };
    }
  } else {
    // order_items: direct filter on item_status + created.
    const dateFilter = thisMonthOnly ? buildCurrentMonthFilter("created") : "";
    const statusFilter = ACTIVE_ORDER_STATUSES.map((s) => `item_status="${s}"`).join(" || ");
    childFilter = `(${statusFilter})${dateFilter ? " && " + dateFilter : ""}`;
    process.stdout.write(`  child filter: ${childFilter}\n`);
  }

  // Step 2: fetch child rows. PB doesn't support `IN` on relation
  // Step 2: fetch child rows. Three modes:
  //   1. quote_items via parentIdFilter — iterate per parent (PB
  //      doesn't support `IN` on relation fields).
  //   2. order_items via childFilter (item_status + date).
  //   3. no filter at all — just take the first `limit` rows.
  const childRows: Array<{ id: string; ai_item_insights: unknown }> = [];
  if (parentIdFilter) {
    for (const id of parentIdFilter) {
      const filter = `quote = "${id}"`;
      const rows = await pb.collection(collection).getList(1, limit, {
        fields: "id,ai_item_insights",
        filter,
      });
      childRows.push(
        ...rows.items.map((r) => ({ id: r.id, ai_item_insights: r.ai_item_insights })),
      );
    }
  } else if (childFilter) {
    const rows = await pb.collection(collection).getList(1, limit, {
      fields: "id,ai_item_insights",
      filter: childFilter,
    });
    childRows.push(
      ...rows.items.map((r) => ({ id: r.id, ai_item_insights: r.ai_item_insights })),
    );
  } else {
    const rows = await pb.collection(collection).getList(1, limit, {
      fields: "id,ai_item_insights",
    });
    childRows.push(
      ...rows.items.map((r) => ({ id: r.id, ai_item_insights: r.ai_item_insights })),
    );
  }
  const rows = { items: childRows };

  let scanned = 0;
  let updated = 0;
  let skipped = 0;
  let errors = 0;

  for (let i = 0; i < rows.items.length; i += batchSize) {
    const batch = rows.items.slice(i, i + batchSize);
    const results = await Promise.all(
      batch.map(async (r) => {
        scanned += 1;
        if (isNewShape(r.ai_item_insights)) {
          return { ok: true, skipped: true };
        }
        if (dryRun) {
          return { ok: true, skipped: false };
        }
        try {
          await populateItemIntel(r.id, collection, { write: true });
          return { ok: true, skipped: false };
        } catch (err) {
          return { ok: false, skipped: false, err };
        }
      }),
    );
    for (const r of results) {
      if (!r.ok) {
        errors += 1;
        continue;
      }
      if (r.skipped) skipped += 1;
      else updated += 1;
    }
    if (updated % 50 === 0 && updated > 0) {
      process.stdout.write(`  ${collection}: ${updated} updated...\n`);
    }
  }

  return { scanned, updated, skipped, errors, duration_ms: Date.now() - t0 };
}

function buildCurrentMonthFilter(field: string): string {
  const now = new Date();
  const y = now.getUTCFullYear();
  const m = now.getUTCMonth(); // 0-indexed
  const start = new Date(Date.UTC(y, m, 1)).toISOString();
  const end = new Date(Date.UTC(y, m + 1, 1)).toISOString();
  return `${field} >= "${start}" && ${field} < "${end}"`;
}

async function main() {
  const args = new Set(process.argv.slice(2));
  const dryRun = args.has("--dry-run");
  const thisMonthOnly = !args.has("--all-time");
  const collection = (args.has("--quote-items") ? "quote_items" : args.has("--order-items") ? "order_items" : "both") as
    | "quote_items"
    | "order_items"
    | "both";
  const limitArg = [...args].find((a) => a.startsWith("--limit="));
  const limit = limitArg ? parseInt(limitArg.split("=")[1] ?? "1000", 10) : 1000;
  const batchArg = [...args].find((a) => a.startsWith("--batch="));
  const batchSize = batchArg ? parseInt(batchArg.split("=")[1] ?? "20", 10) : 20;

  console.log(`PB:        ${PB_URL}`);
  console.log(`Mode:      ${dryRun ? "DRY RUN" : "WRITE"}`);
  console.log(`Collection: ${collection}`);
  console.log(`Limit:      ${limit} per collection`);
  console.log();

  const t0 = Date.now();
  const out: Record<string, unknown> = {};

  if (collection === "quote_items" || collection === "both") {
    out.quote_items = await backfill("quote_items", limit, dryRun, batchSize, thisMonthOnly);
  }
  if (collection === "order_items" || collection === "both") {
    out.order_items = await backfill("order_items", limit, dryRun, batchSize, thisMonthOnly);
  }

  console.log();
  console.log("══════════════════════════════════════════════════");
  console.log("RESULT");
  console.log("══════════════════════════════════════════════════");
  console.log(JSON.stringify(out, null, 2));
  console.log();
  console.log(`Total: ${((Date.now() - t0) / 1000).toFixed(1)}s`);
  if (dryRun) {
    console.log("\n(dry-run — nothing was written. Re-run without --dry-run to apply.)");
  }
}

main().catch((err) => {
  console.error("FATAL:", err);
  process.exit(1);
});