[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8},"actionData","errors","root",{"_108":109,"_110":111},"routes/embed.x.$handle.$name.$",{"_9":10,"_11":12,"_13":14,"_15":16,"_17":18,"_19":20,"_21":-5},"handle","mdevslabs","name","mai","path","vibe-feed.ts","project",{"_22":45,"_11":12,"_33":46,"_47":-5,"_48":42,"_49":50,"_51":50,"_52":53,"_54":55,"_56":57,"_58":59,"_60":32,"_61":-5,"_62":-5,"_63":64,"_65":66,"_67":68,"_69":70,"_71":72,"_73":-5,"_74":75,"_76":75,"_77":78,"_79":78,"_80":78,"_81":75,"_82":75,"_83":84,"_85":86,"_87":88,"_89":42},"file",{"_22":23,"_11":14,"_24":25,"_26":-5,"_27":28,"_29":30,"_31":32,"_33":34,"_35":36,"_37":38,"_39":40,"_41":42,"_43":-5,"_13":44},"permalink","https://www.val.town/x/mdevslabs/mai/code/vibe-feed.ts","extension","id","a095f7d0-d51f-4b25-99e4-2838b086e19a","fileId","06a630e8-a9c5-11f1-9947-1607ee4eb77e","parentId","branchId","3345cb6c-eeb2-4b61-ab03-8e0abc2dcd73","branchVersion",307,"latest",true,"createdAt",["D",1789229613228],"version",7,"fileType","script","content","/**\n * ============================================================================\n * VIBE SOCIAL PLATFORM — FEEDS & DISCOVERY (vibe-feed.ts)\n * Timeline feeds (Pour Vous, Abonnements, Tendances), real-time trends & search\n * ============================================================================\n */\n\nimport type { Hono } from \"npm:hono@4\";\nimport { extractToken, getDb, verifyToken } from \"./config.ts\";\nimport {\n isBlockEitherWay,\n type RegisterMultiFn,\n resolveHiddenUserIds,\n} from \"./vibe-common.ts\";\nimport { HybridRecommender } from \"./vibe-recommender.ts\";\nimport { ensureCircleTable } from \"./vibe-circle.ts\";\nimport {\n attachBookRefs,\n attachPollsAndCollabs,\n attachQuotedPosts,\n ensurePostColumns,\n fetchPostMedia,\n publishDuePosts,\n} from \"./vibe-posts-core.ts\";\n\n// ── Signaux d'affinement d'algorithme (« Cela m'intéresse / pas ») ──\nexport const FEEDBACK_STOP_WORDS = new Set([\n \"avec\",\n \"dans\",\n \"cette\",\n \"pour\",\n \"plus\",\n \"moins\",\n \"tout\",\n \"tous\",\n \"être\",\n \"fait\",\n \"comme\",\n \"mais\",\n \"vous\",\n \"nous\",\n \"elle\",\n \"ils\",\n \"quoi\",\n \"ainsi\",\n \"alors\",\n \"très\",\n \"this\",\n \"that\",\n \"with\",\n \"from\",\n \"your\",\n \"have\",\n \"will\",\n \"about\",\n \"just\",\n \"they\",\n \"http\",\n \"https\",\n \"www\",\n]);\n\nexport const extractFeedbackTokens = (content: string): string[] => {\n const lower = (content || \"\").toLowerCase();\n const tokens: string[] = [];\n for (const m of lower.matchAll(/#([\\p{L}\\p{N}_]{2,30})/gu)) {\n tokens.push(`#${m[1]}`);\n }\n for (const w of lower.split(/[^\\p{L}\\p{N}#']+/u)) {\n if (w.length >= 4 && !FEEDBACK_STOP_WORDS.has(w)) tokens.push(w);\n }\n return Array.from(new Set(tokens));\n};\n\nexport function registerVibeFeedRoutes(\n app: Hono,\n registerMulti: RegisterMultiFn,\n) {\n // Protection contre le double enregistrement (idempotence)\n if ((app as any).__vibe_feed_registered) return;\n (app as any).__vibe_feed_registered = true;\n\n // Initialisation paresseuse des colonnes critiques en tâche de fond\n ensurePostColumns().catch(() => {});\n\n // 1. TIMELINE FEED\n const handleFeed = async (c: any) => {\n try {\n await ensurePostColumns().catch(() => {});\n const type = c.req.query(\"type\") || \"for_you\";\n const token = extractToken(c.req.raw);\n let currentUserId: number | null = null;\n if (token) {\n try {\n const payload = await verifyToken(token);\n currentUserId = Number(payload.sub || (payload as any).id);\n } catch {}\n }\n\n const sql = getDb();\n\n // Comptes masqués (mute) et bloqués : exclus de tous les modes de feed\n // (le blocage est bidirectionnel — cf. resolveHiddenUserIds).\n const { mutedIds, blockedIds } = await resolveHiddenUserIds(\n currentUserId,\n );\n const hiddenAuthorIds = new Set([...mutedIds, ...blockedIds]);\n const isHiddenAuthor = (authorId: any) =>\n hiddenAuthorIds.has(Number(authorId));\n\n // Publication paresseuse des vibes planifiées arrivées à échéance\n await publishDuePosts();\n\n // Pagination par curseur : soit keyset chronologique \"ts|id\", soit\n // rang dans une liste scorée \"rank:N\" (trending / pour vous).\n const PAGE_SIZE = 20;\n const rawCursor = (c.req.query(\"cursor\") || \"\").trim();\n const parseKeyset = (cur: string): { ts: string; id: string } | null => {\n const [ts, id] = cur.split(\"|\");\n if (!ts || !id || Number.isNaN(Date.parse(ts))) return null;\n return { ts, id };\n };\n const parseRank = (cur: string): number | null => {\n const m = cur.match(/^rank:(\\d+)$/);\n return m ? Number(m[1]) : null;\n };\n\n // Médias hydratés via le helper top-level fetchPostMedia (tagged-template only).\n\n if (type === \"trending\") {\n const tag = (c.req.query(\"tag\") || \"\").trim();\n const rankOffset = rawCursor ? parseRank(rawCursor) ?? 0 : 0;\n let posts;\n if (tag) {\n posts = await sql`\n SELECT p.*, pr.display_name, pr.avatar_url, u.username, u.tier,\n (COALESCE(u.is_verified, FALSE) OR LOWER(COALESCE(u.tier, '')) IN ('plus', 'pro', 'max')) as is_verified,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM post_interactions WHERE post_id = p.id AND user_id = ${currentUserId} AND interaction_type = 'like') > 0`\n : sql`FALSE`\n } as has_liked,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM post_interactions WHERE post_id = p.id AND user_id = ${currentUserId} AND interaction_type = 'repost') > 0`\n : sql`FALSE`\n } as has_reposted,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM bookmarks WHERE post_id = p.id AND user_id = ${currentUserId}) > 0`\n : sql`FALSE`\n } as has_bookmarked,\n ${\n currentUserId\n ? sql`(SELECT pi.interaction_type FROM post_interactions pi WHERE pi.post_id = p.id AND pi.user_id = ${currentUserId} AND pi.interaction_type IN ('interest_more', 'interest_less') LIMIT 1)`\n : sql`NULL`\n } as my_feedback\n FROM posts p\n JOIN users u ON u.id = p.author_id\n LEFT JOIN profiles pr ON pr.user_id = u.id\n WHERE p.visibility = 'public' AND COALESCE(p.status, 'published') = 'published' AND p.content ILIKE ('%' || ${tag} || '%')\n ORDER BY (p.likes_count * 3 + p.reposts_count * 2 + p.replies_count * 2) DESC, p.published_at DESC\n LIMIT ${PAGE_SIZE} OFFSET ${rankOffset}\n `;\n } else {\n posts = await sql`\n SELECT p.*, pr.display_name, pr.avatar_url, u.username, u.tier,\n (COALESCE(u.is_verified, FALSE) OR LOWER(COALESCE(u.tier, '')) IN ('plus', 'pro', 'max')) as is_verified,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM post_interactions WHERE post_id = p.id AND user_id = ${currentUserId} AND interaction_type = 'like') > 0`\n : sql`FALSE`\n } as has_liked,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM post_interactions WHERE post_id = p.id AND user_id = ${currentUserId} AND interaction_type = 'repost') > 0`\n : sql`FALSE`\n } as has_reposted,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM bookmarks WHERE post_id = p.id AND user_id = ${currentUserId}) > 0`\n : sql`FALSE`\n } as has_bookmarked,\n ${\n currentUserId\n ? sql`(SELECT pi.interaction_type FROM post_interactions pi WHERE pi.post_id = p.id AND pi.user_id = ${currentUserId} AND pi.interaction_type IN ('interest_more', 'interest_less') LIMIT 1)`\n : sql`NULL`\n } as my_feedback\n FROM posts p\n JOIN users u ON u.id = p.author_id\n LEFT JOIN profiles pr ON pr.user_id = u.id\n WHERE p.visibility = 'public' AND COALESCE(p.status, 'published') = 'published'\n ORDER BY (p.likes_count * 3 + p.reposts_count * 2 + p.replies_count * 2) DESC, p.published_at DESC\n LIMIT ${PAGE_SIZE} OFFSET ${rankOffset}\n `;\n }\n\n await fetchPostMedia(posts);\n await attachQuotedPosts(posts);\n await attachBookRefs(posts, currentUserId).catch(() => {});\n await attachPollsAndCollabs(posts, currentUserId).catch(() => {});\n\n // Filtrage mute/block (après pagination OFFSET — les trous éventuels\n // sont acceptables pour un classement de tendances).\n const visibleTrending = posts.filter((p: any) =>\n !isHiddenAuthor(p.author_id)\n );\n\n return c.json({\n mode: \"trending\",\n title: tag ? `Tendances : ${tag}` : \"Tendances Populaires\",\n count: visibleTrending.length,\n nextCursor: posts.length === PAGE_SIZE\n ? `rank:${rankOffset + PAGE_SIZE}`\n : null,\n posts: visibleTrending,\n });\n }\n\n if (type === \"stream\" || type === \"following\") {\n // Keyset pagination : (published_at, id) < (ts, id) — stable et indexable\n const keyset = rawCursor ? parseKeyset(rawCursor) : null;\n const cursorFilter = (uid: number | null) =>\n keyset\n ? sql`AND (p.published_at, p.id) < (${keyset.ts}::timestamptz, ${keyset.id}::uuid)`\n : sql``;\n let posts;\n if (currentUserId) {\n await ensureCircleTable().catch(() => {});\n posts = await sql`\n SELECT p.*, pr.display_name, pr.avatar_url, u.username, u.tier,\n (COALESCE(u.is_verified, FALSE) OR LOWER(COALESCE(u.tier, '')) IN ('plus', 'pro', 'max')) as is_verified,\n (SELECT COUNT(*) FROM post_interactions WHERE post_id = p.id AND user_id = ${currentUserId} AND interaction_type = 'like') > 0 as has_liked,\n (SELECT COUNT(*) FROM post_interactions WHERE post_id = p.id AND user_id = ${currentUserId} AND interaction_type = 'repost') > 0 as has_reposted,\n (SELECT COUNT(*) FROM bookmarks WHERE post_id = p.id AND user_id = ${currentUserId}) > 0 as has_bookmarked,\n (SELECT pi.interaction_type FROM post_interactions pi WHERE pi.post_id = p.id AND pi.user_id = ${currentUserId} AND pi.interaction_type IN ('interest_more', 'interest_less') LIMIT 1) as my_feedback\n FROM posts p\n JOIN users u ON u.id = p.author_id\n LEFT JOIN profiles pr ON pr.user_id = u.id\n WHERE (\n p.author_id = ${currentUserId}\n OR p.author_id IN (SELECT following_id FROM follows WHERE follower_id = ${currentUserId})\n ) AND COALESCE(p.status, 'published') = 'published'\n AND (\n p.visibility = 'public'\n OR p.author_id = ${currentUserId}\n OR (p.visibility = 'followers' AND EXISTS (SELECT 1 FROM follows f2 WHERE f2.follower_id = ${currentUserId} AND f2.following_id = p.author_id))\n OR (p.visibility = 'circle' AND EXISTS (SELECT 1 FROM circle_members cm2 WHERE cm2.user_id = p.author_id AND cm2.member_user_id = ${currentUserId}))\n ) ${cursorFilter(currentUserId)}\n ORDER BY p.published_at DESC, p.id DESC\n LIMIT ${PAGE_SIZE}\n `;\n } else {\n // Visiteur non connecté : aucun abonnement\n posts = [];\n }\n\n await fetchPostMedia(posts);\n await attachQuotedPosts(posts);\n await attachBookRefs(posts, currentUserId).catch(() => {});\n await attachPollsAndCollabs(posts, currentUserId).catch(() => {});\n\n // Filtrage mute/block dans le flux Abonnements\n const visibleStream = posts.filter((p: any) =>\n !isHiddenAuthor(p.author_id)\n );\n\n const last = visibleStream[visibleStream.length - 1];\n return c.json({\n mode: \"stream\",\n title: \"Abonnements\",\n count: visibleStream.length,\n nextCursor: posts.length === PAGE_SIZE && last\n ? `${new Date(last.published_at).toISOString()}|${last.id}`\n : null,\n posts: visibleStream,\n });\n }\n\n // \"Pour Vous\" — Algorithme de recommandation sophistiqué\n let followedAuthorIds = new Set();\n let affinityByAuthor = new Map();\n let blockedKeywords: string[] = [];\n let shouldHideReposts = false;\n\n if (currentUserId) {\n try {\n const [followsRows, settingsRows, affinityRows] = await Promise.all([\n sql`SELECT following_id FROM follows WHERE follower_id = ${currentUserId}`,\n sql`SELECT blocked_keywords, hide_reposts FROM user_settings WHERE user_id = ${currentUserId} LIMIT 1`,\n // Affinité réelle : historique d'interactions de l'utilisateur par auteur\n sql`\n SELECT p.author_id, COUNT(*)::int AS n\n FROM post_interactions pi\n JOIN posts p ON p.id = pi.post_id\n WHERE pi.user_id = ${currentUserId}\n AND pi.interaction_type IN ('like', 'repost')\n GROUP BY p.author_id\n ORDER BY n DESC\n LIMIT 50\n `,\n ]);\n followedAuthorIds = new Set(\n followsRows.map((f: any) => Number(f.following_id)),\n );\n for (const row of affinityRows) {\n affinityByAuthor.set(\n Number(row.author_id),\n Math.min(1, Number(row.n) / 5),\n );\n }\n if (settingsRows[0]) {\n blockedKeywords = (settingsRows[0].blocked_keywords || []).map((\n k: string,\n ) => k.toLowerCase().trim());\n shouldHideReposts = Boolean(settingsRows[0].hide_reposts);\n }\n } catch {}\n }\n\n const rawCandidates = await sql`\n SELECT p.*, pr.display_name, pr.avatar_url, u.username, u.tier,\n (COALESCE(u.is_verified, FALSE) OR LOWER(COALESCE(u.tier, '')) IN ('plus', 'pro', 'max')) as is_verified,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM post_interactions WHERE post_id = p.id AND user_id = ${currentUserId} AND interaction_type = 'like') > 0`\n : sql`FALSE`\n } as has_liked,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM post_interactions WHERE post_id = p.id AND user_id = ${currentUserId} AND interaction_type = 'repost') > 0`\n : sql`FALSE`\n } as has_reposted,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM bookmarks WHERE post_id = p.id AND user_id = ${currentUserId}) > 0`\n : sql`FALSE`\n } as has_bookmarked,\n ${\n currentUserId\n ? sql`(SELECT pi.interaction_type FROM post_interactions pi WHERE pi.post_id = p.id AND pi.user_id = ${currentUserId} AND pi.interaction_type IN ('interest_more', 'interest_less') LIMIT 1)`\n : sql`NULL`\n } as my_feedback\n FROM posts p\n JOIN users u ON u.id = p.author_id\n LEFT JOIN profiles pr ON pr.user_id = u.id\n WHERE p.visibility = 'public' AND COALESCE(p.status, 'published') = 'published'\n ORDER BY p.published_at DESC\n LIMIT 300\n `;\n\n await fetchPostMedia(rawCandidates);\n\n // Filtrage selon les paramètres utilisateur\n const filteredCandidates = rawCandidates.filter((post: any) => {\n if (shouldHideReposts && post.is_repost) return false;\n if (currentUserId && Number(post.author_id) === currentUserId) {\n return true;\n }\n if (isHiddenAuthor(post.author_id)) return false;\n if (blockedKeywords.length > 0) {\n const contentLc = (post.content || \"\").toLowerCase();\n const hasBlocked = blockedKeywords.some((kw: string) =>\n kw && contentLc.includes(kw)\n );\n if (hasBlocked) return false;\n }\n return true;\n });\n\n // Signaux d'affinement : retours « Cela m'intéresse / pas » + centres d'intérêt\n let moreAuthorWeights = new Map();\n let lessAuthorWeights = new Map();\n let moreTokenWeights = new Map();\n let lessTokenWeights = new Map();\n let interestTags: string[] = [];\n let dwellByPost = new Map();\n let topicAffinity = new Map();\n\n if (currentUserId) {\n try {\n const [feedbackRows, interestRows, dwellRows, affinityRows] =\n await Promise.all([\n sql`\n SELECT pi.interaction_type, p.author_id, p.content\n FROM post_interactions pi\n JOIN posts p ON p.id = pi.post_id\n WHERE pi.user_id = ${currentUserId}\n AND pi.interaction_type IN ('interest_more', 'interest_less')\n ORDER BY pi.created_at DESC\n LIMIT 200\n `,\n sql`SELECT interests FROM profiles WHERE user_id = ${currentUserId} LIMIT 1`,\n sql`\n SELECT post_id, SUM(duration_ms)::int AS total_ms\n FROM post_views WHERE user_id = ${currentUserId} AND created_at > NOW() - INTERVAL '30 days'\n GROUP BY post_id\n `.catch(() => []),\n sql`SELECT topic, score FROM user_topic_affinity WHERE user_id = ${currentUserId} ORDER BY score DESC LIMIT 20`\n .catch(() => []),\n ]);\n for (const row of feedbackRows) {\n const authorId = Number(row.author_id);\n const isMore = row.interaction_type === \"interest_more\";\n const authorMap = isMore ? moreAuthorWeights : lessAuthorWeights;\n const tokenMap = isMore ? moreTokenWeights : lessTokenWeights;\n authorMap.set(\n authorId,\n Math.min(3, (authorMap.get(authorId) || 0) + 1),\n );\n for (const token of extractFeedbackTokens(row.content)) {\n tokenMap.set(token, Math.min(3, (tokenMap.get(token) || 0) + 1));\n }\n }\n const rawInterests = interestRows[0]?.interests;\n if (Array.isArray(rawInterests)) {\n interestTags = rawInterests.map((t: any) =>\n String(t).toLowerCase().trim()\n ).filter(Boolean);\n } else if (typeof rawInterests === \"string\") {\n interestTags = rawInterests.split(\",\").map((t) =>\n t.toLowerCase().trim()\n ).filter(Boolean);\n }\n for (const d of (dwellRows as any[]) || []) {\n dwellByPost.set(\n String((d as any).post_id),\n Number((d as any).total_ms || 0),\n );\n }\n for (const a of (affinityRows as any[]) || []) {\n topicAffinity.set(\n String((a as any).topic).toLowerCase(),\n Number((a as any).score || 0),\n );\n }\n } catch (feedbackErr) {\n console.warn(\"[Vibe API] Erreur signaux feedback:\", feedbackErr);\n }\n }\n\n const clamp = (v: number, lo: number, hi: number) =>\n Math.max(lo, Math.min(hi, v));\n\n const computeInterestSignal = (\n post: any,\n ): { signal: number; matched: string[] } => {\n if (!currentUserId) return { signal: 0, matched: [] };\n const authorId = Number(post.author_id);\n const authorDelta = (moreAuthorWeights.get(authorId) || 0) -\n (lessAuthorWeights.get(authorId) || 0);\n const tokens = extractFeedbackTokens(post.content);\n let topicDelta = 0;\n let interestMatches = 0;\n let affinityBoost = 0;\n const matched: string[] = [];\n for (const token of tokens) {\n const wMore = moreTokenWeights.get(token) || 0;\n const wLess = lessTokenWeights.get(token) || 0;\n if (wMore > 0 || wLess > 0) topicDelta += wMore - wLess;\n const bare = token.replace(/^#/, \"\");\n if (interestTags.includes(bare) || interestTags.includes(token)) {\n interestMatches += 1;\n if (matched.length < 3) matched.push(bare);\n }\n const aff = topicAffinity.get(bare) ?? topicAffinity.get(token) ?? 0;\n if (aff > 0) {\n affinityBoost += aff;\n if (matched.length < 3 && !matched.includes(bare)) {\n matched.push(bare);\n }\n }\n }\n // Dwell : temps passé sur ce post précis → boost implicite (vrai profil)\n const dwellMs = dwellByPost.get(String(post.id)) || 0;\n const dwellBoost = Math.min(1, dwellMs / 30000);\n const signal = clamp(\n 0.35 * (authorDelta / 3) +\n 0.3 * clamp(topicDelta / 3, -3, 3) / 3 +\n 0.15 * (Math.min(2, interestMatches) / 2) +\n 0.1 * Math.min(1, affinityBoost) +\n 0.1 * dwellBoost,\n -1,\n 1,\n );\n return { signal, matched };\n };\n\n const scoredPosts = filteredCandidates.map((post: any) => {\n const interest = computeInterestSignal(post);\n // Affinité topic → similarité sémantique (remplace le 0.75 en dur)\n const topicScore = interest.matched.length > 0 ? 0.85 : 0.6;\n const signal = HybridRecommender.scorePost({\n postId: post.id,\n authorId: Number(post.author_id),\n publishedAt: new Date(post.published_at),\n likes: Number(post.likes_count || 0),\n reposts: Number(post.reposts_count || 0),\n replies: Number(post.replies_count || 0),\n views: Number(post.views_count || 0),\n hasMedia: Array.isArray(post.media_assets) &&\n post.media_assets.length > 0,\n isVerifiedAuthor: Boolean(post.is_verified),\n isFollowedAuthor: followedAuthorIds.has(Number(post.author_id)),\n affinity: affinityByAuthor.get(Number(post.author_id)) || 0,\n semanticSimilarity: topicScore,\n candidateSentiment: Number(post.sentiment_score || 0.5),\n toxicityScore: Number(post.toxicity_score || 0),\n interestSignal: interest.signal,\n matchedInterestTags: interest.matched,\n userDwellMs: dwellByPost.get(String(post.id)) || 0,\n });\n\n return {\n ...post,\n recommendationScore: signal.totalScore,\n explanation: signal.explanationText,\n scoreBreakdown: signal.breakdown,\n };\n });\n\n scoredPosts.sort((a: any, b: any) =>\n b.recommendationScore - a.recommendationScore\n );\n\n // Diversification : max 3 posts consécutifs du même auteur\n const diversified: any[] = [];\n let streakAuthor: number | null = null;\n let streakLen = 0;\n const deferred: any[] = [];\n for (const post of scoredPosts) {\n const authorId = Number(post.author_id);\n if (authorId === streakAuthor && streakLen >= 3) {\n deferred.push(post);\n continue;\n }\n if (authorId === streakAuthor) streakLen += 1;\n else {\n streakAuthor = authorId;\n streakLen = 1;\n }\n diversified.push(post);\n }\n const ranked = diversified.concat(deferred);\n\n const rankOffset = rawCursor ? parseRank(rawCursor) ?? 0 : 0;\n const page = ranked.slice(rankOffset, rankOffset + PAGE_SIZE);\n await attachQuotedPosts(page);\n await attachBookRefs(page, currentUserId).catch(() => {});\n await attachPollsAndCollabs(page, currentUserId).catch(() => {});\n\n return c.json({\n mode: \"for_you\",\n title: \"Pour Vous\",\n count: page.length,\n nextCursor:\n page.length === PAGE_SIZE && rankOffset + PAGE_SIZE < ranked.length\n ? `rank:${rankOffset + PAGE_SIZE}`\n : null,\n posts: page,\n });\n } catch (err: any) {\n console.error(\"[Vibe API] Error fetching feed:\", err);\n return c.json({ error: \"Erreur lors de la récupération du flux.\" }, 500);\n }\n };\n\n registerMulti(\n \"get\",\n [\"/api/vibe/feed\", \"/vibe/feed\", \"/v1/feed\", \"/feed\"],\n handleFeed,\n );\n\n // 2. REAL TRENDS & HASHTAGS\n const handleGetTrends = async (c: any) => {\n try {\n const sql = getDb();\n\n // Chercher les posts récents (48h d'abord, puis 30j si pas assez)\n let recentPosts = await sql`\n SELECT content, likes_count, reposts_count, replies_count, views_count, published_at\n FROM posts\n WHERE published_at > NOW() - INTERVAL '48 hours'\n AND visibility = 'public' AND status = 'published'\n ORDER BY (likes_count * 2 + reposts_count * 3 + replies_count) DESC\n LIMIT 300\n `;\n\n if (recentPosts.length < 5) {\n recentPosts = await sql`\n SELECT content, likes_count, reposts_count, replies_count, views_count, published_at\n FROM posts\n WHERE published_at > NOW() - INTERVAL '30 days'\n AND visibility = 'public' AND status = 'published'\n ORDER BY (likes_count * 2 + reposts_count * 3 + replies_count) DESC\n LIMIT 300\n `;\n }\n\n const tagMap: Record<\n string,\n { count: number; engagement: number; recencyBoost: number }\n > = {};\n const now = Date.now();\n\n for (const p of recentPosts) {\n const text = p.content || \"\";\n const matches = text.match(/#[\\p{L}\\p{N}_]+/gu) || [];\n\n const eng = Number(p.likes_count || 0) * 2 +\n Number(p.reposts_count || 0) * 3 +\n Number(p.replies_count || 0) * 2 +\n Number(p.views_count || 0) * 0.1 +\n 1;\n\n const ageHours = (now - new Date(p.published_at).getTime()) /\n (1000 * 3600);\n const recency = ageHours < 6 ? 3 : ageHours < 24 ? 1.5 : 1;\n\n for (const rawTag of matches) {\n const normalized = rawTag.toLowerCase().trim();\n if (normalized.length <= 1 || normalized.length > 35) continue;\n\n if (!tagMap[rawTag]) {\n tagMap[rawTag] = { count: 0, engagement: 0, recencyBoost: 0 };\n }\n tagMap[rawTag].count += 1;\n tagMap[rawTag].engagement += eng;\n tagMap[rawTag].recencyBoost += recency;\n }\n }\n\n const sortedTrends = Object.entries(tagMap)\n .map(([tag, data]) => {\n const score = data.engagement * data.recencyBoost + data.count * 5;\n const formatted = data.count > 1000\n ? `${(data.count / 1000).toFixed(1)}k`\n : `${data.count}`;\n const lc = tag.toLowerCase();\n const category =\n lc.includes(\"mai\") || lc.includes(\"ia\") || lc.includes(\"ai\") ||\n lc.includes(\"llm\") || lc.includes(\"gpt\")\n ? \"Intelligence Artificielle\"\n : lc.includes(\"tech\") || lc.includes(\"dev\") ||\n lc.includes(\"code\") || lc.includes(\"web\")\n ? \"Technologie\"\n : lc.includes(\"design\") || lc.includes(\"art\") ||\n lc.includes(\"photo\") || lc.includes(\"ux\")\n ? \"Design & Création\"\n : lc.includes(\"vibe\") || lc.includes(\"social\") ||\n lc.includes(\"community\")\n ? \"Communauté\"\n : \"Tendances\";\n return {\n tag,\n category,\n posts: `${formatted} publications`,\n post_count: data.count,\n score,\n };\n })\n .sort((a, b) => b.score - a.score)\n .slice(0, 10);\n\n return c.json({ success: true, trends: sortedTrends });\n } catch (err: any) {\n console.error(\"[Get Trends Error]:\", err);\n return c.json({ success: true, trends: [] });\n }\n };\n\n registerMulti(\"get\", [\n \"/api/vibe/trends\",\n \"/vibe/trends\",\n \"/v1/trends\",\n \"/trends\",\n ], handleGetTrends);\n\n // 2bis. TOP VIBE — meilleures publications des N derniers jours (Explorer)\n // Score d'engagement : likes×1 + reposts×2.5 + réponses×2 + vues×0.1\n // (miroir SQL de rawEngagement — lib/recommendation/signals.ts).\n const handleTopPosts = async (c: any) => {\n try {\n await ensurePostColumns().catch(() => {});\n const token = extractToken(c.req.raw);\n let currentUserId: number | null = null;\n if (token) {\n try {\n const payload = await verifyToken(token);\n currentUserId = Number(payload.sub || (payload as any).id);\n } catch {}\n }\n\n const limit = Math.min(\n 10,\n Math.max(1, Number(c.req.query(\"limit\") || 5)),\n );\n const days = Math.min(90, Math.max(1, Number(c.req.query(\"days\") || 30)));\n\n const sql = getDb();\n await publishDuePosts();\n\n const posts = await sql`\n SELECT p.*, pr.display_name, pr.avatar_url, u.username, u.tier,\n (COALESCE(u.is_verified, FALSE) OR LOWER(COALESCE(u.tier, '')) IN ('plus', 'pro', 'max')) as is_verified,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM post_interactions WHERE post_id = p.id AND user_id = ${currentUserId} AND interaction_type = 'like') > 0`\n : sql`FALSE`\n } as has_liked,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM post_interactions WHERE post_id = p.id AND user_id = ${currentUserId} AND interaction_type = 'repost') > 0`\n : sql`FALSE`\n } as has_reposted,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM bookmarks WHERE post_id = p.id AND user_id = ${currentUserId}) > 0`\n : sql`FALSE`\n } as has_bookmarked,\n ${\n currentUserId\n ? sql`(SELECT pi.interaction_type FROM post_interactions pi WHERE pi.post_id = p.id AND pi.user_id = ${currentUserId} AND pi.interaction_type IN ('interest_more', 'interest_less') LIMIT 1)`\n : sql`NULL`\n } as my_feedback\n FROM posts p\n JOIN users u ON u.id = p.author_id\n LEFT JOIN profiles pr ON pr.user_id = u.id\n WHERE p.visibility = 'public' AND COALESCE(p.status, 'published') = 'published'\n AND p.published_at > NOW() - (${days} * INTERVAL '1 day')\n ORDER BY (COALESCE(p.likes_count, 0) * 1.0\n + COALESCE(p.reposts_count, 0) * 2.5\n + COALESCE(p.replies_count, 0) * 2.0\n + COALESCE(p.views_count, 0) * 0.1) DESC,\n p.published_at DESC\n LIMIT ${limit}\n `;\n\n await fetchPostMedia(posts);\n await attachQuotedPosts(posts);\n await attachBookRefs(posts, currentUserId).catch(() => {});\n await attachPollsAndCollabs(posts, currentUserId).catch(() => {});\n\n // Filtrage mute/block\n const { mutedIds, blockedIds } = await resolveHiddenUserIds(\n currentUserId,\n );\n const hiddenAuthorIds = new Set([...mutedIds, ...blockedIds]);\n const visible = posts.filter((p: any) =>\n !hiddenAuthorIds.has(Number(p.author_id))\n );\n\n return c.json({ success: true, posts: visible });\n } catch (err: any) {\n console.error(\"[Top Posts Error]:\", err);\n return c.json({ success: true, posts: [] });\n }\n };\n\n registerMulti(\"get\", [\n \"/api/vibe/posts/top\",\n \"/vibe/posts/top\",\n \"/v1/posts/top\",\n ], handleTopPosts);\n\n // 3. SEARCH POSTS\n const handleSearchPosts = async (c: any) => {\n try {\n await ensurePostColumns().catch(() => {});\n const q = (c.req.query(\"q\") || c.req.query(\"query\") || \"\").trim();\n if (!q) {\n return c.json({ posts: [], count: 0 });\n }\n const token = extractToken(c.req.raw);\n let currentUserId: number | null = null;\n if (token) {\n try {\n const payload = await verifyToken(token);\n currentUserId = Number(payload.sub || (payload as any).id);\n } catch {}\n }\n\n const sql = getDb();\n const limit = Math.min(\n 50,\n Math.max(1, Number(c.req.query(\"limit\") || 20)),\n );\n const offset = Math.max(0, Number(c.req.query(\"offset\") || 0));\n\n const cleanQ = q.startsWith(\"#\") ? q.slice(1) : q;\n const posts = await sql`\n SELECT p.*, pr.display_name, pr.avatar_url, u.username, u.tier,\n (COALESCE(u.is_verified, FALSE) OR LOWER(COALESCE(u.tier, '')) IN ('plus', 'pro', 'max')) as is_verified,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM post_interactions WHERE post_id = p.id AND user_id = ${currentUserId} AND interaction_type = 'like') > 0`\n : sql`FALSE`\n } as has_liked,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM post_interactions WHERE post_id = p.id AND user_id = ${currentUserId} AND interaction_type = 'repost') > 0`\n : sql`FALSE`\n } as has_reposted,\n ${\n currentUserId\n ? sql`(SELECT COUNT(*) FROM bookmarks WHERE post_id = p.id AND user_id = ${currentUserId}) > 0`\n : sql`FALSE`\n } as has_bookmarked,\n ${\n currentUserId\n ? sql`(SELECT pi.interaction_type FROM post_interactions pi WHERE pi.post_id = p.id AND pi.user_id = ${currentUserId} AND pi.interaction_type IN ('interest_more', 'interest_less') LIMIT 1)`\n : sql`NULL`\n } as my_feedback\n FROM posts p\n JOIN users u ON u.id = p.author_id\n LEFT JOIN profiles pr ON pr.user_id = u.id\n WHERE p.visibility = 'public' AND COALESCE(p.status, 'published') = 'published'\n AND (\n p.content ILIKE ('%' || ${cleanQ} || '%')\n OR u.username ILIKE ('%' || ${cleanQ} || '%')\n OR pr.display_name ILIKE ('%' || ${cleanQ} || '%')\n )\n ORDER BY p.published_at DESC\n LIMIT ${limit} OFFSET ${offset}\n `;\n\n await fetchPostMedia(posts);\n await attachQuotedPosts(posts);\n await attachBookRefs(posts, currentUserId).catch(() => {});\n await attachPollsAndCollabs(posts, currentUserId).catch(() => {});\n\n // Filtrage mute/block des résultats de recherche\n const { mutedIds: searchMuted, blockedIds: searchBlocked } =\n await resolveHiddenUserIds(currentUserId);\n const searchHidden = new Set([...searchMuted, ...searchBlocked]);\n const visiblePosts = posts.filter((p: any) =>\n !searchHidden.has(Number(p.author_id))\n );\n\n return c.json({\n query: q,\n count: visiblePosts.length,\n posts: visiblePosts,\n });\n } catch (err: any) {\n console.error(\"[Search Posts Error]:\", err);\n return c.json({ error: \"Erreur recherche publications.\" }, 500);\n }\n };\n\n registerMulti(\"get\", [\n \"/api/vibe/search/posts\",\n \"/vibe/search/posts\",\n \"/v1/search/posts\",\n \"/search/posts\",\n ], handleSearchPosts);\n\n // 4. UNIFIED GLOBAL SEARCH (posts + users + books + DMs, limit/section)\n const handleSearchUnified = async (c: any) => {\n try {\n await ensurePostColumns().catch(() => {});\n const q = (c.req.query(\"q\") || c.req.query(\"query\") || \"\").trim();\n if (!q) {\n return c.json({\n posts: [],\n users: [],\n books: [],\n messages: [],\n total: 0,\n });\n }\n const token = extractToken(c.req.raw);\n let currentUserId: number | null = null;\n if (token) {\n try {\n const payload = await verifyToken(token);\n currentUserId = Number(payload.sub || (payload as any).id);\n } catch {}\n }\n const sql = getDb();\n const limit = Math.min(\n 10,\n Math.max(1, Number(c.req.query(\"limit\") || 5)),\n );\n const cleanQ = q.startsWith(\"#\") ? q.slice(1) : q;\n\n const postsPromise = sql`\n SELECT p.*, pr.display_name, pr.avatar_url, u.username, u.tier,\n (COALESCE(u.is_verified, FALSE) OR LOWER(COALESCE(u.tier, '')) IN ('plus', 'pro', 'max')) as is_verified\n FROM posts p\n JOIN users u ON u.id = p.author_id\n LEFT JOIN profiles pr ON pr.user_id = u.id\n WHERE p.visibility = 'public' AND COALESCE(p.status, 'published') = 'published'\n AND (\n p.content ILIKE ('%' || ${cleanQ} || '%')\n OR u.username ILIKE ('%' || ${cleanQ} || '%')\n OR pr.display_name ILIKE ('%' || ${cleanQ} || '%')\n )\n ORDER BY p.published_at DESC\n LIMIT ${limit}\n `.catch(() => []);\n\n const usersPromise = sql`\n SELECT u.id, u.username, u.tier,\n (COALESCE(u.is_verified, FALSE) OR LOWER(COALESCE(u.tier, '')) IN ('plus', 'pro', 'max')) as is_verified,\n pr.display_name, pr.avatar_url, pr.bio, pr.followers_count, pr.posts_count\n FROM users u\n LEFT JOIN profiles pr ON pr.user_id = u.id\n WHERE LOWER(u.username) LIKE ${`%${q.toLowerCase()}%`}\n OR LOWER(COALESCE(pr.display_name, '')) LIKE ${`%${q.toLowerCase()}%`}\n ORDER BY COALESCE(pr.followers_count, 0) DESC\n LIMIT ${limit}\n `.catch(() => []);\n\n const booksPromise = currentUserId\n ? sql`\n SELECT b.id, b.title, b.icon, b.created_at,\n (SELECT COUNT(*) FROM vibe_book_items bi WHERE bi.book_id = b.id) AS items_count\n FROM vibe_books b\n WHERE b.user_id = ${currentUserId} AND b.title ILIKE ('%' || ${q} || '%')\n ORDER BY b.created_at DESC\n LIMIT ${limit}\n `.catch(() => [])\n : Promise.resolve([]);\n\n const messagesPromise = currentUserId\n ? sql`\n SELECT m.id, m.conversation_id, m.sender_id, m.recipient_id, m.content, m.is_read, m.created_at\n FROM direct_messages m\n WHERE m.content ILIKE ('%' || ${q} || '%')\n AND m.conversation_id IN (\n SELECT dm.id FROM dm_conversations dm\n WHERE dm.participant_one_id = ${currentUserId} OR dm.participant_two_id = ${currentUserId}\n )\n AND (m.status IS NULL OR m.status = 'sent')\n ORDER BY m.created_at DESC\n LIMIT ${limit}\n `.catch(() => [])\n : Promise.resolve([]);\n\n const [posts, users, books, messages] = await Promise.all([\n postsPromise,\n usersPromise,\n booksPromise,\n messagesPromise,\n ]);\n try {\n await fetchPostMedia(posts);\n } catch {}\n try {\n await attachQuotedPosts(posts);\n } catch {}\n try {\n await attachBookRefs(posts, currentUserId);\n } catch {}\n try {\n await attachPollsAndCollabs(posts, currentUserId);\n } catch {}\n\n return c.json({\n posts,\n users,\n books,\n messages,\n total: posts.length + users.length + books.length + messages.length,\n });\n } catch (err: any) {\n console.error(\"[Search Unified Error]:\", err);\n return c.json({ error: \"Erreur recherche unifiée.\" }, 500);\n }\n };\n\n registerMulti(\"get\", [\n \"/api/vibe/search/unified\",\n \"/vibe/search/unified\",\n \"/v1/search/unified\",\n \"/search/unified\",\n ], handleSearchUnified);\n}","deleted",false,"label",[14],"200f8efb-f4ef-4117-ae26-ee737dbffb76",["D",1785339526786],"deletedAt","legacyEnvVars","privacy","public","httpPrivacy","description","Avec l'API d'mAI, récupérez des données sur les projets !","imageUrl","https://imagedelivery.net/iHX6Ovru0O7AjmyT5yZRoA/cdf5d0c3-f220-4dec-712e-87455a037c00/square512?exp=1789538399&sig=cd777e882462fce47bac2003be9a9fe1267d4472bdb46131ff02f8b27f89b2ac","userId","7ed13655-83c2-4934-a981-01578302df33","pinnedFileId","968028a2-8b63-11f1-b718-1607ee4eb77e","pinPreview","pinMigrated","screenshotVersion","likeCount",0,"orgId","019f99eb-d944-7332-99f8-2fbb8b1e7011","tursoName","200f8efbf4ef4117ae26ee737dbffb76","tursoHostname","200f8efbf4ef4117ae26ee737dbffb76-val-town.aws-us-east-2.turso.io","tursoMeasuredAt",["D",1789474235934],"pinned","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm","user",{"_9":10,"_107":103},"org",{"_9":10,"_100":101,"_102":103,"_56":57,"_104":105,"_106":42},"branch",{"_35":90,"_11":91,"_22":28,"_92":45,"_93":-5,"_94":-5,"_95":-5,"_96":-5,"_97":-5,"_98":12,"_99":10},"isOwned",364,"main","projectId","forkedBranchId","forkedBranchVersion","forkedProjectPrivacy","forkedProjectId","forkedProjectOrgId","projectName","userHandle","displayName","mProjects","avatar","https://imagedelivery.net/iHX6Ovru0O7AjmyT5yZRoA/ec72623f-c02d-461a-9b1c-c2f416214400/square512?exp=1789538399&sig=b90c7895d89b2c8a9d976c298024ed06d954dd92e3be0b0905971776951516a4","tier","free","canRestrictVals","profileImageUrl","ctx",{"_83":-5,"_141":142,"_143":144,"_145":146,"_147":42,"_148":42,"_149":42,"_150":32,"_151":-5,"_152":-7,"_153":-7,"_154":64,"_155":42,"_156":42,"_157":-5,"_158":42,"_159":42,"_160":42,"_161":-5},"ENV",{"_112":113,"_114":115,"_116":91,"_117":118,"_119":120,"_121":122,"_123":124,"_125":126,"_127":128,"_129":130,"_131":132,"_133":134,"_135":136,"_137":138,"_139":140},"REACT_APP_SENTRY_DSN","https://586020d86f05489db63a554690acf1e9@o1398359.ingest.sentry.io/6724499","RENDER_GIT_COMMIT","e87eb04e707cf64e7e532c8e8ec395f209e17f15","RENDER_GIT_BRANCH","POSTHOG_API_TOKEN","phc_nmoyuAkStEZuUp69rYlg1DZgrIt19Gx0Y4Hwf8tGgJP","POSTHOG_API_HOST","https://wart.val.town","WEB_API_TEMPLATE","https://{val}.web.val.run","APP_URL","https://www.val.town","MODULE_URL","https://esm.town","VANITY_WEB_API_TEMPLATE","https://{label}.val.run/","DENOLS_URL","https://lsp.val.town","NODE_ENV","production","CLOUDFLARE_TURNSTILE_SITE_KEY","0x4AAAAAACtaIZJH7s9rBMZ6","GOOGLE_CLIENT_ID","875989696899-jtsqkhlbdv5jrlsk8631p7ipjg3c6o62.apps.googleusercontent.com","STRIPE_PUBLISHABLE_KEY","pk_live_51N2zvWEpZTTLipIkVEwXrUG9AT0yeOOEYxszarZT1jpsZV58Ly6wlg3SltZpHN26CUalOBmUSY7adQRQiSkCZz2500qSXNkf8Y","API_URL","https://api.val.town","orgs",[],"preferencesKeybindings","default","preferencesTheme","light","preferencesFixers","preferencesCopilot","preferencesSelectionAi","preferencesCloseBrackets","preferencesChatPanelState","preferencesDprint","preferencesDenofmt","notificationCount","flagDenols","isBusinessOrgMember","businessOrgHandle","isBusinessOrgAdmin","isHigherTownieTier","navCollapsed","currentOrgId"]