fix: Instagram data collection pipeline — handle normalization + DB persistence

- enrich-channels: Instagram fallback — auto-try _ps, .ps, _clinic suffixes when <100 followers
- enrich-channels: YouTube URL normalization via normalizeYouTubeChannel (handles /c/, /user/, @handle)
- enrich-channels: Google Maps multi-query search for better hit rate
- generate-report: AI-found social handles prioritized over Firecrawl scrape
- generate-report: Added socialMedia field to AI prompt for accurate handle discovery
- normalizeHandles: Added normalizeYouTubeChannel for /c/, /user/, /channel/, @handle URLs

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

supabase/functions/_shared/normalizeHandles.ts

@@ -46,3 +46,62 @@ export function normalizeInstagramHandle(
 
   return handle || null;
 }
+
+/**
+ * Normalize a YouTube channel identifier from various URL formats.
+ *
+ * Returns an object with the best identifier type for API lookup:
+ *  - "https://www.youtube.com/@banobagips"        → { type: 'handle', value: 'banobagips' }
+ *  - "https://youtube.com/c/banobagips"           → { type: 'username', value: 'banobagips' }
+ *  - "https://youtube.com/user/banobagi"          → { type: 'username', value: 'banobagi' }
+ *  - "https://youtube.com/channel/UCxxxx"         → { type: 'channelId', value: 'UCxxxx' }
+ *  - "@banobagips"                                → { type: 'handle', value: 'banobagips' }
+ *  - "UCxxxx"                                     → { type: 'channelId', value: 'UCxxxx' }
+ *  - "banobagips"                                 → { type: 'username', value: 'banobagips' }
+ */
+export function normalizeYouTubeChannel(
+  raw: string | null | undefined,
+): { type: 'handle' | 'username' | 'channelId'; value: string } | null {
+  if (!raw || typeof raw !== "string") return null;
+
+  let input = raw.trim();
+  if (!input) return null;
+
+  // Parse YouTube URLs
+  if (input.includes("youtube.com") || input.includes("youtu.be")) {
+    try {
+      const urlStr = input.startsWith("http") ? input : `https://${input}`;
+      const url = new URL(urlStr);
+      const segments = url.pathname.split("/").filter(Boolean);
+
+      if (segments[0] === "channel" && segments[1]?.startsWith("UC")) {
+        return { type: "channelId", value: segments[1] };
+      }
+      if (segments[0] === "c" && segments[1]) {
+        return { type: "username", value: segments[1] };
+      }
+      if (segments[0] === "user" && segments[1]) {
+        return { type: "username", value: segments[1] };
+      }
+      if (segments[0]?.startsWith("@")) {
+        return { type: "handle", value: segments[0].slice(1) };
+      }
+      // Fallback: first path segment
+      if (segments[0]) {
+        return { type: "username", value: segments[0] };
+      }
+    } catch {
+      // URL parsing failed
+    }
+  }
+
+  // Non-URL formats
+  if (input.startsWith("@")) {
+    return { type: "handle", value: input.slice(1) };
+  }
+  if (input.startsWith("UC") && input.length > 20) {
+    return { type: "channelId", value: input };
+  }
+
+  return { type: "username", value: input };
+}

supabase/functions/enrich-channels/index.ts

@@ -1,6 +1,6 @@
 import "@supabase/functions-js/edge-runtime.d.ts";
 import { createClient } from "https://esm.sh/@supabase/supabase-js@2";
-import { normalizeInstagramHandle } from "../_shared/normalizeHandles.ts";
+import { normalizeInstagramHandle, normalizeYouTubeChannel } from "../_shared/normalizeHandles.ts";
 
 const corsHeaders = {
   "Access-Control-Allow-Origin": "*",
@@ -58,18 +58,33 @@ Deno.serve(async (req) => {
     // Run all enrichment tasks in parallel
     const tasks = [];
 
-    // 1. Instagram Profile
+    // 1. Instagram Profile — with fallback for wrong handle
     const cleanIgHandle = normalizeInstagramHandle(instagramHandle);
     if (cleanIgHandle) {
       tasks.push(
         (async () => {
+          // Try the given handle first, then common clinic variants
+          const handleCandidates = [
+            cleanIgHandle,
+            `${cleanIgHandle}_ps`,           // banobagi → banobagi_ps
+            `${cleanIgHandle}.ps`,           // banobagi → banobagi.ps
+            `${cleanIgHandle}_clinic`,       // banobagi → banobagi_clinic
+            `${cleanIgHandle}_official`,     // banobagi → banobagi_official
+          ];
+
+          for (const handle of handleCandidates) {
             const items = await runApifyActor(
               "apify~instagram-profile-scraper",
-            { usernames: [cleanIgHandle], resultsLimit: 12 },
+              { usernames: [handle], resultsLimit: 12 },
               APIFY_TOKEN
             );
             const profile = (items as Record<string, unknown>[])[0];
+
             if (profile && !profile.error) {
+              const followers = (profile.followersCount as number) || 0;
+
+              // Accept if: has meaningful followers OR is a business account with posts
+              if (followers >= 100 || ((profile.isBusinessAccount as boolean) && (profile.postsCount as number) > 10)) {
                 enrichment.instagram = {
                   username: profile.username,
                   followers: profile.followersCount,
@@ -88,6 +103,9 @@ Deno.serve(async (req) => {
                       timestamp: p.timestamp,
                     })),
                 };
+                break; // Found a valid account
+              }
+            }
           }
         })()
       );
@@ -97,17 +115,27 @@ Deno.serve(async (req) => {
     if (clinicName || address) {
       tasks.push(
         (async () => {
-          const searchQuery = `${clinicName} ${address || "강남"}`;
+          // Try multiple search queries for better hit rate
-          const items = await runApifyActor(
+          const queries = [
+            `${clinicName} 성형외과`,
+            clinicName,
+            `${clinicName} ${address || "강남"}`,
+          ];
+
+          let items: unknown[] = [];
+          for (const query of queries) {
+            items = await runApifyActor(
               "compass~crawler-google-places",
               {
-              searchStringsArray: [searchQuery],
+                searchStringsArray: [query],
-              maxCrawledPlacesPerSearch: 1,
+                maxCrawledPlacesPerSearch: 3,
                 language: "ko",
                 maxReviews: 10,
               },
               APIFY_TOKEN
             );
+            if ((items as Record<string, unknown>[]).length > 0) break;
+          }
           const place = (items as Record<string, unknown>[])[0];
           if (place) {
             enrichment.googleMaps = {
@@ -140,17 +168,24 @@ Deno.serve(async (req) => {
           (async () => {
             const YT_BASE = "https://www.googleapis.com/youtube/v3";
 
-            // Resolve handle/username to channel ID
+            // Normalize YouTube URL/handle to structured identifier
-            let channelId = youtubeChannelId;
+            const ytNormalized = normalizeYouTubeChannel(youtubeChannelId);
-            if (channelId.startsWith("@") || !channelId.startsWith("UC")) {
+            if (!ytNormalized) return;
-              // Use forHandle for @handles, forUsername for legacy usernames
+
-              const param = channelId.startsWith("@") ? "forHandle" : "forUsername";
+            let channelId = "";
-              const handle = channelId.startsWith("@") ? channelId.slice(1) : channelId;
+
+            if (ytNormalized.type === "channelId") {
+              channelId = ytNormalized.value;
+            } else {
+              // Try forHandle first, then forUsername as fallback
+              for (const param of ["forHandle", "forUsername"]) {
                 const lookupRes = await fetch(
-                `${YT_BASE}/channels?part=id&${param}=${handle}&key=${YOUTUBE_API_KEY}`
+                  `${YT_BASE}/channels?part=id&${param}=${ytNormalized.value}&key=${YOUTUBE_API_KEY}`
                 );
                 const lookupData = await lookupRes.json();
                 channelId = lookupData.items?.[0]?.id || "";
+                if (channelId) break;
+              }
             }
 
             if (!channelId) return;

supabase/functions/generate-report/index.ts

@@ -92,7 +92,13 @@ ${JSON.stringify(analyzeResult.data?.analysis || {}, null, 2)}
     "address": "주소",
     "phone": "전화번호",
     "services": ["시술1", "시술2"],
-    "doctors": [{"name": "의사명", "specialty": "전문분야"}]
+    "doctors": [{"name": "의사명", "specialty": "전문분야"}],
+    "socialMedia": {
+      "instagram": "정확한 Instagram 핸들 (@ 없이, 예: banobagi_ps)",
+      "youtube": "YouTube 채널 핸들 또는 URL",
+      "facebook": "Facebook 페이지명",
+      "naverBlog": "네이버 블로그 ID"
+    }
   },
   "executiveSummary": "경영진 요약 (3-5문장)",
   "overallScore": 0-100,
@@ -165,13 +171,14 @@ ${JSON.stringify(analyzeResult.data?.analysis || {}, null, 2)}
       report = { raw: reportText, parseError: true };
     }
 
-    // Normalize social handles from scrape data
+    // Merge social handles: AI-found (more accurate) > Firecrawl-extracted (fallback)
-    const socialMedia = clinic.socialMedia || {};
+    const scrapeSocial = clinic.socialMedia || {};
+    const aiSocial = report?.clinicInfo?.socialMedia || {};
     const normalizedHandles = {
-      instagram: normalizeInstagramHandle(socialMedia.instagram),
+      instagram: normalizeInstagramHandle(aiSocial.instagram) || normalizeInstagramHandle(scrapeSocial.instagram),
-      youtube: socialMedia.youtube || null,
+      youtube: aiSocial.youtube || scrapeSocial.youtube || null,
-      facebook: socialMedia.facebook || null,
+      facebook: aiSocial.facebook || scrapeSocial.facebook || null,
-      blog: socialMedia.blog || null,
+      blog: aiSocial.naverBlog || scrapeSocial.blog || null,
     };
 
     // Embed normalized handles in report for DB persistence