Frontend Updates on Brand Discovery

src/api/client.js

@@ -387,6 +387,17 @@ export const api = {
       body: JSON.stringify(payload),
     }),
 
+  // Web & retail listings: a background search (minutes, not seconds) that
+  // the panel starts, polls, and then reads through the preview's `use_web`.
+  startWebDiscovery: (brand, { refresh = false } = {}) =>
+    request('/api/admin/brand-discovery/web-jobs', {
+      method: 'POST',
+      body: JSON.stringify({ brand, refresh }),
+    }),
+
+  getWebDiscoveryJob: (jobId) =>
+    request(`/api/admin/brand-discovery/web-jobs/${encodeURIComponent(jobId)}`),
+
   ingestBrandDiscovery: (
     brand,
     products,

src/pages/BrandDiscoveryPanel.jsx

@@ -1,7 +1,7 @@
 import React, { useMemo, useState } from 'react';
 import {
   Search, Sparkles, Loader2, AlertTriangle, PlayCircle, Database,
-  ShieldCheck, PackageSearch, Barcode,
+  ShieldCheck, PackageSearch, Barcode, Store,
 } from 'lucide-react';
 import { api } from '../api/client';
 
@@ -25,6 +25,14 @@ import { api } from '../api/client';
  * The review step stays regardless. A person reading the list is still the last
  * check before anything is written, so the list is shown.
  *
+ * WEB & RETAIL LISTINGS (optional, off by default). For brands Open Food Facts
+ * does not carry - non-food above all - the server can search retailer listings
+ * (Blinkit, Zepto, Instamart, BigBasket, JioMart, Amazon, Flipkart). It reads
+ * only the search engine's result titles and links, never a shop's page and
+ * never a language model, and keeps a product only when a real listing names
+ * the brand and states a pack size. That search takes minutes, so it runs as a
+ * job: Discover starts it, this panel polls it, then asks for the preview.
+ *
  * There is no progress UI here on purpose. Ingest returns a BatchOut, and
  * AdminPage hands it to BatchCatalogPanel through the `adoptBatch` prop it
  * already has - which draws the eleven stages, and owns Resume and Cancel.
@@ -32,16 +40,38 @@ import { api } from '../api/client';
 
 const EVIDENCE_TONE = {
   openfacts: 'bg-emerald-500/10 text-emerald-700 border-emerald-500/20',
+  retail: 'bg-violet-500/10 text-violet-700 border-violet-500/20',
   catalog: 'bg-sky-500/10 text-sky-700 border-sky-500/20',
   registry: 'bg-amber-500/10 text-amber-700 border-amber-500/20',
 };
 
 const EVIDENCE_LABEL = {
   openfacts: 'Open Food Facts',
+  retail: 'retail listing',
   catalog: 'already in catalog',
   registry: 'known sub-brand',
 };
 
+const WEB_POLL_MS = 2500;
+const WEB_FINISHED = new Set(['done', 'partial', 'failed', 'interrupted']);
+const sleep = (ms) => new Promise((resolve) => setTimeout(resolve, ms));
+
+// The badge for one row. A brand-store row has no `evidence` tier (the store IS
+// the source), and used to fall through to "model only" - which it is not.
+function evidenceBadge(product) {
+  if (product.evidence === 'retail' && product.retailer_count) {
+    const n = product.retailer_count;
+    return { tone: EVIDENCE_TONE.retail, label: `retail - ${n} shop${n === 1 ? '' : 's'}` };
+  }
+  if (product.evidence) {
+    return { tone: EVIDENCE_TONE[product.evidence], label: EVIDENCE_LABEL[product.evidence] };
+  }
+  if ((product.sources || []).includes('store')) {
+    return { tone: EVIDENCE_TONE.openfacts, label: "brand's own shop" };
+  }
+  return null;
+}
+
 function Toggle({ checked, onChange, label, hint }) {
   return (
     <label className="flex items-start gap-2 text-xs text-slate-600 cursor-pointer">
@@ -82,6 +112,9 @@ export function BrandDiscoveryPanel({ onBatchStarted }) {
   const [useOpenfacts, setUseOpenfacts] = useState(true);
   const [requireEvidence, setRequireEvidence] = useState(true);
   const [fetchImages, setFetchImages] = useState(true);
+  // Off by default: it costs minutes of paced web searches.
+  const [useWeb, setUseWeb] = useState(false);
+  const [webJob, setWebJob] = useState(null);
   // Set by the 409 the ingest route answers with when the brand sits outside
   // ACTIVE_BRANDS. Staging anyway is a deliberate second press, never a default.
   const [acknowledgeInactive, setAcknowledgeInactive] = useState(false);
@@ -100,10 +133,24 @@ export function BrandDiscoveryPanel({ onBatchStarted }) {
     setError('');
     setResult(null);
     setAcknowledgeInactive(false);
+    setWebJob(null);
     try {
+      let webJobId = null;
+      if (useWeb) {
+        let job = await api.startWebDiscovery(brand.trim());
+        setWebJob(job);
+        while (!WEB_FINISHED.has(job.status)) {
+          await sleep(WEB_POLL_MS);
+          job = await api.getWebDiscoveryJob(job.job_id);
+          setWebJob(job);
+        }
+        webJobId = job.job_id;
+      }
       const body = await api.previewBrandDiscovery({
         brand: brand.trim(),
         use_openfacts: useOpenfacts,
+        use_web: useWeb,
+        web_job_id: webJobId,
         // SENT, NOT OMITTED. DiscoveryPreviewRequest.use_llm defaults to True
         // on the server, so leaving the key out would quietly re-enable the
         // source this panel no longer exposes.
@@ -147,6 +194,10 @@ export function BrandDiscoveryPanel({ onBatchStarted }) {
           fssai_license: p.fssai_license,
           barcode: p.barcode,
           image_url: p.image_url,
+          // Web-found rows only; recorded as provenance after the batch.
+          ...(p.listings?.length
+            ? { listings: p.listings, price_range: p.price_range, retailer_count: p.retailer_count }
+            : {}),
         })),
         { fetch_images: fetchImages, acknowledge_inactive: acknowledgeInactive },
       );
@@ -210,7 +261,11 @@ export function BrandDiscoveryPanel({ onBatchStarted }) {
             ) : (
               <PackageSearch className="h-4 w-4" />
             )}
-            {discovering ? 'Discovering...' : 'Discover products'}
+            {discovering
+              ? webJob && !WEB_FINISHED.has(webJob.status)
+                ? `Searching retailers ${webJob.queries_done}/${webJob.queries_total || '?'}...`
+                : 'Discovering...'
+              : 'Discover products'}
           </button>
         </div>
 
@@ -233,6 +288,12 @@ export function BrandDiscoveryPanel({ onBatchStarted }) {
             label="Fetch images on ingest"
             hint="Slow; nothing fills them in later"
           />
+          <Toggle
+            checked={useWeb}
+            onChange={setUseWeb}
+            label="Web & retail listings"
+            hint="Blinkit, Zepto, Instamart, BigBasket, JioMart, Amazon, Flipkart - for brands Open Food Facts lacks. Takes a few minutes; results are cached."
+          />
         </div>
       </div>
 
@@ -251,6 +312,9 @@ export function BrandDiscoveryPanel({ onBatchStarted }) {
             <Stat icon={ShieldCheck} value={counts.from_openfacts ?? 0} label="from Open Food Facts" />
             <Stat icon={Barcode} value={counts.with_barcode ?? 0} label="with a real barcode" />
             <Stat icon={Database} value={counts.already_in_catalog ?? 0} label="already in the catalog" />
+            {counts.from_web !== undefined && (
+              <Stat icon={Store} value={counts.from_web} label="from retail listings" />
+            )}
           </div>
 
           {(result.warnings || []).map((warning) => (
@@ -332,6 +396,26 @@ export function BrandDiscoveryPanel({ onBatchStarted }) {
                           {note}
                         </div>
                       ))}
+                      {product.listings?.length > 0 && (
+                        <details className="text-[11px] text-slate-500">
+                          <summary className="cursor-pointer">
+                            {product.listings.length} listing{product.listings.length === 1 ? '' : 's'}
+                            {product.price_range ? ` - ${product.price_range}` : ''}
+                          </summary>
+                          {product.listings.map((listing) => (
+                            <a
+                              key={listing.url}
+                              href={listing.url}
+                              target="_blank"
+                              rel="noopener noreferrer"
+                              className="block truncate text-violet-700 underline"
+                              title={listing.title}
+                            >
+                              {listing.retailer}: {listing.title}
+                            </a>
+                          ))}
+                        </details>
+                      )}
                     </td>
                     <td className="px-3 py-2 text-slate-600">
                       {product.size_variants?.length
@@ -343,11 +427,11 @@ export function BrandDiscoveryPanel({ onBatchStarted }) {
                       {product.barcode || <span className="text-slate-400">-</span>}
                     </td>
                     <td className="px-3 py-2">
-                      {product.evidence ? (
+                      {evidenceBadge(product) ? (
                         <span
-                          className={`rounded-full border px-2 py-0.5 text-[10px] font-bold ${EVIDENCE_TONE[product.evidence]}`}
+                          className={`rounded-full border px-2 py-0.5 text-[10px] font-bold ${evidenceBadge(product).tone}`}
                         >
-                          {EVIDENCE_LABEL[product.evidence]}
+                          {evidenceBadge(product).label}
                         </span>
                       ) : (
                         <span className="rounded-full border border-ink-900/10 bg-ink-100 px-2 py-0.5 text-[10px] font-bold text-slate-500">
@@ -365,7 +449,10 @@ export function BrandDiscoveryPanel({ onBatchStarted }) {
             <p className="text-xs text-slate-500">
               Nothing was found for this brand. Open Food Facts may not carry
               it - non-food brands are thin there - and no verified storefront
-              may be registered for it. Discovery only offers products a real
+              may be registered for it.{' '}
+              {useWeb
+                ? 'No retailer listing named the brand with a pack size either.'
+                : 'Tick "Web & retail listings" to search retailer listings too.'} Discovery only offers products a real
               source can vouch for, so an empty list means nothing was found,
               not that something was hidden.
             </p>