[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8},"actionData","errors","root",{"_110":111,"_112":113},"routes/embed.x.$handle.$name.$",{"_9":10,"_11":12,"_13":14,"_15":16,"_17":18,"_19":20,"_21":-5},"handle","efrenlin","name","x-trader-consensus","path","lib/discoveryV2.ts","project",{"_22":48,"_11":12,"_35":49,"_50":-5,"_51":44,"_52":53,"_54":53,"_55":56,"_57":-5,"_58":59,"_60":61,"_62":34,"_63":-5,"_64":-5,"_65":66,"_67":68,"_69":70,"_71":72,"_73":74,"_75":-5,"_76":77,"_78":77,"_79":80,"_81":80,"_82":80,"_83":77,"_84":77,"_85":86,"_87":88,"_89":90,"_91":44},"file",{"_22":23,"_11":24,"_25":26,"_27":28,"_29":30,"_31":32,"_33":34,"_35":36,"_37":38,"_39":40,"_41":42,"_43":44,"_45":-5,"_13":46},"permalink","https://www.val.town/x/efrenlin/x-trader-consensus/code/lib/discoveryV2.ts","extension","id","6ce6c942-04fb-45fd-8f0a-b968401c7cba","discoveryV2.ts","fileId","8e3de3bc-ac26-11f1-ab14-1607ee4eb77e","parentId","8dccd4ba-aa00-11f1-906c-1607ee4eb77e","branchId","012dced4-9569-4aa1-b4df-043c162f1aa5","branchVersion",112,"latest",true,"createdAt",["D",1788958592929],"version",11,"fileType","script","content","// lib/discoveryV2.ts — ADDITIVE \"Market Discovery & Consensus V2\" engine.\n//\n// Builds on the validated V1 data (existing analyzed posts, consensus,\n// Discovery candidates from lib/discovery.ts). It NEVER re-analyzes posts,\n// never runs the collector, never calls Anthropic, never touches the 3-hour\n// cron, and never modifies V1 tables.\n//\n// PERFORMANCE ARCHITECTURE (precomputed + cached):\n// - computeDiscovery() runs a bounded full computation ONLY when the stored\n// V2 results are stale (empty or older than FRESH_MS). Otherwise the\n// read-only endpoints read the precomputed V2 tables directly — no\n// whole-dataset rescan per request.\n//\n// Honesty rules enforced throughout:\n// - Only dimensions with real evidence are scored; everything else stays\n// \"insufficient data\" (never fabricated, never zeroed as if negative).\n// - Semantic mappings require High confidence to influence scoring.\n// - Not Mentioned is never Neutral.\n// - Discovery tiers are research priorities, NOT buy recommendations.\n\nimport { sqlite } from \"https://esm.town/v/std/sqlite/main.ts\";\nimport { TRACKED_ACCOUNTS } from \"../config.ts\";\nimport { computeTickerConsensus } from \"./consensus.ts\";\nimport { getExcludedTickers } from \"./tickerIdentity.ts\";\nimport { extractCandidates } from \"./discovery.ts\";\nimport { scoreMarketDimensions, ensureMarketDataSchema } from \"./marketData.ts\";\nimport type { MarketData } from \"./marketData.ts\";\n\nexport const WATCHLIST = [\n \"QQQ\", \"TSLA\", \"PYPL\", \"META\", \"CAKE\", \"AMZN\", \"NVDA\", \"SE\",\n \"GOOGL\", \"AVGO\", \"XYZ\", \"SOXX\", \"AMD\", \"TSM\", \"ABSI\",\n];\n\n// Reliable company names only; never guessed. Unknown -> null.\nconst KNOWN_COMPANIES: Record = {\n NVDA: \"NVIDIA Corporation\", MU: \"Micron Technology\", INTC: \"Intel Corporation\",\n AMD: \"Advanced Micro Devices\", BE: \"Bloom Energy\", HOOD: \"Robinhood Markets\",\n AVGO: \"Broadcom Inc.\", AAPL: \"Apple Inc.\", AMZN: \"Amazon.com\", META: \"Meta Platforms\",\n GOOGL: \"Alphabet Inc.\", MSFT: \"Microsoft Corporation\", TSLA: \"Tesla Inc.\",\n QCOM: \"Qualcomm Inc.\", ORCL: \"Oracle Corporation\", CRDO: \"Credo Technology\",\n LITE: \"Lumentum Holdings\", COHR: \"Coherent Corp.\", MRVL: \"Marvell Technology\",\n AAOI: \"Applied Optoelectronics\", DELL: \"Dell Technologies\", SMCI: \"Super Micro\",\n SE: \"Sea Limited\", PYPL: \"PayPal Holdings\", TSM: \"Taiwan Semiconductor\",\n SOXX: \"iShares Semiconductor ETF\", QQQ: \"Invesco QQQ Trust\", ABSI: \"Absci Corporation\",\n};\n\n// How long a computed V2 snapshot is considered fresh before recompute.\nconst FRESH_MS = 15 * 60 * 1000;\n\n// Dimension weights sum to 100 (the 0-100 score scale).\nconst DIM = {\n traderConsensus: 15,\n independentConsensus: 15,\n attentionAcceleration: 15,\n newsCatalyst: 10,\n capitalFlow: 10, // scored from market_data when present, else insufficient\n fundamentalQuality: 10, // scored from market_data when present, else insufficient\n valuation: 5, // scored from market_data when present, else insufficient\n priceMomentum: 10, // scored from market_data when present, else insufficient\n themeFit: 10,\n} as const;\nconst TOTAL_MAX = Object.values(DIM).reduce((a, b) => a + b, 0); // 100\n\nconst safeParseArray = (json: string | null | undefined): string[] => {\n if (!json) return [];\n try { const v = JSON.parse(json); return Array.isArray(v) ? v.filter((x) => typeof x === \"string\") : []; }\n catch { return []; }\n};\nconst safeParseSemantic = (json: string | null | undefined): any[] => {\n if (!json) return [];\n try { const v = JSON.parse(json); return Array.isArray(v) ? v : []; }\n catch { return []; }\n};\nfunction tickersForPost(p: any): string[] {\n const s = new Set();\n safeParseArray(p.mentioned_tickers).forEach((t) => s.add(t.toUpperCase()));\n safeParseArray(p.stance_tickers).forEach((t) => s.add(t.toUpperCase()));\n safeParseSemantic(p.semantic_tickers).forEach((x) => x && x.ticker && s.add(String(x.ticker).toUpperCase()));\n return [...s];\n}\nfunction postMentions(p: any, ticker: string): boolean {\n const t = ticker.toUpperCase();\n return safeParseArray(p.mentioned_tickers).some((x) => x.toUpperCase() === t) ||\n safeParseArray(p.stance_tickers).some((x) => x.toUpperCase() === t) ||\n safeParseSemantic(p.semantic_tickers).some((x) => x && x.ticker && String(x.ticker).toUpperCase() === t);\n}\n\nexport async function ensureV2Schema(): Promise {\n await sqlite.batch([\n { sql: `CREATE TABLE IF NOT EXISTS discovery_candidates (\n ticker TEXT PRIMARY KEY, company_name TEXT,\n score_normalized INTEGER, score_available REAL, available_max INTEGER, coverage REAL,\n confidence TEXT, tier TEXT, why_discovered TEXT,\n direction TEXT, disagreement_level TEXT, independent_consensus INTEGER,\n bullish INTEGER, bearish INTEGER, unclear INTEGER,\n attention_trend TEXT, catalysts TEXT, theme TEXT, evidence TEXT,\n sources TEXT, first_detected TEXT, last_updated TEXT, is_watchlist INTEGER DEFAULT 0\n )` },\n { sql: `CREATE TABLE IF NOT EXISTS discovery_signals (\n id INTEGER PRIMARY KEY AUTOINCREMENT, ticker TEXT NOT NULL, dimension TEXT NOT NULL,\n source TEXT NOT NULL, evidence TEXT, confidence TEXT, signal_time TEXT NOT NULL, payload TEXT\n )` },\n { sql: `CREATE INDEX IF NOT EXISTS idx_dsig_ticker ON discovery_signals(ticker, signal_time)` },\n { sql: `CREATE TABLE IF NOT EXISTS discovery_runs (\n id INTEGER PRIMARY KEY AUTOINCREMENT, computed_at TEXT NOT NULL,\n candidates INTEGER NOT NULL, themes INTEGER NOT NULL, avg_coverage REAL\n )` },\n { sql: `CREATE TABLE IF NOT EXISTS emerging_accounts (\n handle TEXT PRIMARY KEY, category TEXT NOT NULL, times_referenced INTEGER DEFAULT 0,\n referencing_accounts TEXT NOT NULL DEFAULT '[]', tickers TEXT NOT NULL DEFAULT '[]',\n evidence TEXT NOT NULL DEFAULT '[]', promo_flagged INTEGER DEFAULT 0,\n confidence TEXT, repost_ratio REAL, last_evaluated TEXT,\n promotion_status TEXT NOT NULL DEFAULT 'Not Promoted', notes TEXT\n )` },\n { sql: `CREATE TABLE IF NOT EXISTS themes (\n id INTEGER PRIMARY KEY AUTOINCREMENT, name TEXT NOT NULL, label TEXT,\n trend TEXT, related_tickers TEXT NOT NULL DEFAULT '[]', related_accounts TEXT NOT NULL DEFAULT '[]',\n evidence TEXT NOT NULL DEFAULT '[]', attention INTEGER DEFAULT 0,\n first_detected TEXT, last_updated TEXT\n )` },\n { sql: `CREATE TABLE IF NOT EXISTS watchlist (\n ticker TEXT PRIMARY KEY, name TEXT, source TEXT NOT NULL DEFAULT 'user', added_at TEXT NOT NULL\n )` },\n ]);\n // Seed the user's existing watchlist (idempotent) in one batched statement.\n const now = new Date().toISOString();\n const seed = WATCHLIST.map((t) => ({\n sql: `INSERT OR IGNORE INTO watchlist (ticker, name, source, added_at) VALUES (?, ?, 'user', ?)`,\n args: [t, KNOWN_COMPANIES[t] ?? null, now],\n }));\n await sqlite.batch(seed);\n await ensureMarketDataSchema();\n}\n\nasync function loadAnalyzedPosts(): Promise {\n const res = await sqlite.execute(\n `SELECT post_id, username, post_type, posted_at, text, mentioned_tickers, stance_tickers,\n explicit_stance, semantic_tickers, evidence, thesis, catalyst, original_url\n FROM posts WHERE analyzed_at IS NOT NULL`,\n );\n return res.rows as unknown as any[];\n}\n\nfunction attention(posts: any[], t: string) {\n const rel = posts.filter((p) => postMentions(p, t));\n if (rel.length < 2) return { recent: rel.length, prior: 0, pct: null, rising: false };\n rel.sort((a, b) => (a.posted_at < b.posted_at ? -1 : 1));\n const first = new Date(rel[0].posted_at).getTime();\n const last = new Date(rel[rel.length - 1].posted_at).getTime();\n const span = Math.max(last - first, 1);\n const mid = first + span * 0.25;\n const recent = rel.filter((p) => new Date(p.posted_at).getTime() >= mid).length;\n const prior = rel.length - recent;\n const pct = prior > 0 ? Math.round(((recent - prior) / prior) * 100) : (recent > 0 ? 100 : 0);\n return { recent, prior, pct, rising: recent > prior };\n}\n\nfunction stanceChanges(posts: any[], t: string): number {\n const byAcct = new Map();\n for (const p of posts) {\n if (!postMentions(p, t)) continue;\n const st = p.explicit_stance;\n if (st !== \"Bullish\" && st !== \"Bearish\") continue;\n const a = byAcct.get(p.username) ?? {};\n if (!a.first) a.first = st;\n a.last = st;\n byAcct.set(p.username, a);\n }\n let flips = 0;\n for (const a of byAcct.values()) if (a.first && a.last && a.first !== a.last) flips++;\n return flips;\n}\n\nfunction buildThemes(posts: any[]): any[] {\n const postTickers = posts.map((p) => ({ p, ticks: tickersForPost(p) })).filter((x) => x.ticks.length >= 1);\n const co = new Map>();\n for (const { ticks } of postTickers) {\n for (const a of ticks) for (const b of ticks) {\n if (a === b) continue;\n const m = co.get(a) ?? new Map();\n m.set(b, (m.get(b) ?? 0) + 1);\n co.set(a, m);\n }\n }\n const parent = new Map();\n const root = (x: string) => { while (parent.get(x) !== x) { parent.set(x, parent.get(parent.get(x)!)!); x = parent.get(x)!; } return x; };\n const allT = new Set();\n postTickers.forEach((x) => x.ticks.forEach((t) => allT.add(t)));\n allT.forEach((t) => parent.set(t, t));\n for (const [a, m] of co) for (const [b, n] of m) if (n >= 2) { const ra = root(a), rb = root(b); if (ra !== rb) parent.set(rb, ra); }\n const clusters = new Map>();\n for (const t of allT) { const r = root(t); if (!clusters.has(r)) clusters.set(r, new Set()); clusters.get(r)!.add(t); }\n\n const out: any[] = [];\n for (const members of clusters.values()) {\n if (members.size < 2) continue;\n const ticks = [...members].sort();\n const related: any[] = [];\n postTickers.forEach(({ p, ticks: pt }) => {\n const hit = pt.filter((x) => members.has(x));\n if (hit.length >= 2) related.push({ account: p.username, tickers: hit, evidence: p.evidence ?? p.catalyst ?? p.thesis ?? p.text?.slice(0, 160), postedAt: p.posted_at, url: p.original_url });\n });\n const accounts = new Set(related.map((r) => r.account));\n const names = ticks.map((t) => `${t} (${KNOWN_COMPANIES[t] ?? \"—\"})`).slice(0, 4);\n out.push({\n name: ticks.join(\" · \"), label: names.join(\" · \"), related_tickers: ticks, related_accounts: [...accounts],\n evidence: related.slice(0, 5), attention: related.length,\n first_detected: related.length ? related.map((r) => r.postedAt).sort()[0] : null,\n last_updated: related.length ? related.map((r) => r.postedAt).sort().slice(-1)[0] : null,\n trend: `attention ${related.length} posts across ${accounts.size} accounts`,\n });\n }\n return out.sort((a, b) => b.attention - a.attention);\n}\n\n// Bounded full computation: builds candidates + signals + themes, persists\n// them to the V2 tables using batched writes so it completes quickly.\nexport async function computeDiscovery(\n opts: { onlyTickers?: string[] } = {},\n): Promise<{ generatedAt: string; candidates: number; themes: number }> {\n await ensureV2Schema();\n const posts = await loadAnalyzedPosts();\n const excluded = await getExcludedTickers();\n const rows = posts.map((p) => ({\n post_id: p.post_id, username: p.username, post_type: p.post_type, posted_at: p.posted_at,\n text: p.text, mentioned_tickers: p.mentioned_tickers ?? \"[]\", stance_tickers: p.stance_tickers ?? \"[]\",\n explicit_stance: p.explicit_stance, semantic_tickers: p.semantic_tickers ?? \"[]\",\n }));\n // Scoped mode: opts.onlyTickers restricts which tickers get a\n // discovery_candidates/discovery_signals row written this call (e.g. a\n // controlled single-ticker market-data test). Read-only derivations\n // (consensus math, attention, themes) still use the full analyzed-post\n // set — only WHICH rows get WRITTEN is scoped, so every other ticker's\n // row is left completely untouched. Never re-analyzes posts, never calls\n // Anthropic, never touches ticker_consensus.\n const onlyTickers = opts.onlyTickers ? new Set(opts.onlyTickers.map((t) => t.toUpperCase())) : null;\n const consensus = computeTickerConsensus(rows, excluded)\n .filter((c) => !excluded.has(c.ticker))\n .filter((c) => !onlyTickers || onlyTickers.has(c.ticker));\n const themes = buildThemes(posts);\n const watchRow = new Set(((await sqlite.execute(`SELECT ticker FROM watchlist`)).rows as any[]).map((r) => r.ticker.toUpperCase()));\n const now = new Date().toISOString();\n const threeDaysAgo = new Date(Date.now() - 3 * 86400000).toISOString();\n // Persisted market/fundamental data (separate table). Tick without a row\n // keeps the four market dimensions unavailable (NaN) — scores unchanged.\n const marketRows = (await sqlite.execute(`SELECT * FROM market_data`)).rows as any[];\n const marketByTicker = new Map(marketRows.map((r) => [r.ticker.toUpperCase(), r as unknown as MarketData]));\n\n const candStmts: any[] = [];\n const sigStmts: any[] = [];\n\n for (const c of consensus) {\n const t = c.ticker;\n const rel = posts.filter((p) => postMentions(p, t));\n const attn = attention(posts, t);\n const flips = stanceChanges(posts, t);\n const firstDetected = rel.length ? rel.map((p) => p.posted_at).sort()[0] : null;\n const lastUpdated = rel.length ? rel.map((p) => p.posted_at).sort().slice(-1)[0] : now;\n const isNew = firstDetected ? firstDetected >= threeDaysAgo : false;\n const catalysts = [...new Set(rel.map((p) => p.catalyst).filter(Boolean))];\n const theses = [...new Set(rel.map((p) => p.thesis).filter(Boolean))];\n const theme = themes.find((th) => th.related_tickers.includes(t));\n\n const scores: Record = {};\n const strength = c.bullish_accounts + c.bearish_accounts > 0 ? Math.max(c.bullish_ratio ?? 0, c.bearish_ratio ?? 0) : 0;\n scores.traderConsensus = Math.min(15, Math.round(strength * 7.5 + (c.bullish_accounts + c.bearish_accounts >= 2 ? 7.5 : 0)));\n scores.independentConsensus = Math.min(15, (c.bullish_accounts + c.bearish_accounts) * 3);\n if (attn.pct === null) scores.attentionAcceleration = NaN; else scores.attentionAcceleration = attn.rising ? Math.min(15, 6 + attn.pct / 10) : 2;\n scores.newsCatalyst = Math.min(10, catalysts.length * 5);\n scores.themeFit = theme ? Math.min(10, Math.ceil((theme.related_accounts.length + theme.related_tickers.length) / 2)) : 0;\n // Market/fundamental dims score ONLY when a persisted market_data row exists\n // for this ticker; otherwise NaN => dimension unavailable (never zero).\n const mds = scoreMarketDimensions(marketByTicker.get(t) ?? null);\n scores.fundamentalQuality = mds.fundamentalQuality;\n scores.valuation = mds.valuation;\n scores.priceMomentum = mds.priceMomentum;\n scores.capitalFlow = mds.capitalFlow;\n // No hardcoded \"unavailable\" list: NaN dims are excluded automatically, so\n // existing scores are unchanged until market data is actually provided.\n const unavailable: string[] = [];\n\n const availableMax = Object.entries(DIM).filter(([k]) => !unavailable.includes(k) && !Number.isNaN(scores[k] as number)).reduce((a, [, w]) => a + w, 0);\n const availableScore = Object.entries(scores).filter(([k, v]) => !unavailable.includes(k) && !Number.isNaN(v)).reduce((a, [, v]) => a + (v as number), 0);\n const normalized = Math.round(availableMax > 0 ? (availableScore / availableMax) * 100 : 0);\n const coverage = availableMax / TOTAL_MAX;\n const confidence = Math.min(1, coverage * (0.5 + 0.1 * Math.min(c.unique_accounts_mentioning, 5)));\n\n let tier: string;\n if (normalized >= 70 && c.unique_accounts_mentioning >= 3 && coverage >= 0.6) tier = \"A\";\n else if (normalized >= 40 && c.unique_accounts_mentioning >= 2) tier = \"B\";\n else tier = \"C\";\n\n const why: string[] = [];\n if (isNew) why.push(\"newly emerging (first seen within 3 days)\");\n if (attn.rising) why.push(`attention rising (${attn.pct}%)`);\n if (flips > 0) why.push(`${flips} account stance change(s)`);\n if (c.unique_accounts_mentioning >= 3) why.push(`${c.unique_accounts_mentioning} independent accounts`);\n if (c.consensus_direction === \"Bullish\" || c.consensus_direction === \"Bearish\") why.push(`${c.consensus_direction} trader consensus`);\n if (c.disagreement_level === \"High\") why.push(\"high disagreement\");\n if (catalysts.length) why.push(\"news/catalyst present\");\n if (theme) why.push(`theme: ${theme.name}`);\n\n const rec = {\n ticker: t, company_name: KNOWN_COMPANIES[t] ?? null,\n score_normalized: normalized, score_available: Math.round(availableScore * 10) / 10,\n available_max: availableMax, coverage: Math.round(coverage * 100) / 100,\n confidence: confidence >= 0.66 ? \"High\" : confidence >= 0.4 ? \"Medium\" : \"Low\", tier,\n why_discovered: why.length ? why.join(\"; \") : \"mentioned by tracked accounts\",\n direction: c.consensus_direction, disagreement_level: c.disagreement_level,\n independent_consensus: c.unique_accounts_mentioning, bullish: c.bullish_accounts,\n bearish: c.bearish_accounts, unclear: c.unclear_accounts,\n attention_trend: attn.pct === null ? null : `${attn.rising ? \"rising\" : \"falling\"} ${Math.abs(attn.pct)}%`,\n catalysts: JSON.stringify(catalysts.slice(0, 4)),\n theme: theme ? theme.name : null,\n evidence: JSON.stringify([...new Set([...catalysts, ...theses])].slice(0, 3)),\n sources: JSON.stringify([\n { source: \"trader_consensus\", evidence: `${c.unique_accounts_mentioning} unique accounts`, confidence: \"high\" },\n { source: \"attention_acceleration\", evidence: attn.pct === null ? \"insufficient\" : `${attn.pct}% recent vs prior`, confidence: attn.pct === null ? \"low\" : \"medium\" },\n ]),\n first_detected: firstDetected, last_updated: lastUpdated,\n is_watchlist: watchRow.has(t) ? 1 : 0,\n };\n\n candStmts.push({\n sql: `INSERT INTO discovery_candidates (\n ticker, company_name, score_normalized, score_available, available_max, coverage, confidence, tier,\n why_discovered, direction, disagreement_level, independent_consensus, bullish, bearish, unclear,\n attention_trend, catalysts, theme, evidence, sources, first_detected, last_updated, is_watchlist)\n VALUES (?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?)\n ON CONFLICT(ticker) DO UPDATE SET\n score_normalized=excluded.score_normalized, score_available=excluded.score_available,\n available_max=excluded.available_max, coverage=excluded.coverage, confidence=excluded.confidence,\n tier=excluded.tier, why_discovered=excluded.why_discovered, direction=excluded.direction,\n disagreement_level=excluded.disagreement_level, independent_consensus=excluded.independent_consensus,\n bullish=excluded.bullish, bearish=excluded.bearish, unclear=excluded.unclear,\n attention_trend=excluded.attention_trend, catalysts=excluded.catalysts, theme=excluded.theme,\n evidence=excluded.evidence, sources=excluded.sources, first_detected=excluded.first_detected,\n last_updated=excluded.last_updated, is_watchlist=excluded.is_watchlist`,\n args: [rec.ticker, rec.company_name, rec.score_normalized, rec.score_available,\n rec.available_max, rec.coverage, rec.confidence, rec.tier, rec.why_discovered,\n rec.direction, rec.disagreement_level, rec.independent_consensus, rec.bullish,\n rec.bearish, rec.unclear, rec.attention_trend, rec.catalysts, rec.theme,\n rec.evidence, rec.sources, rec.first_detected, rec.last_updated, rec.is_watchlist],\n });\n\n for (const dim of Object.keys(DIM)) {\n const scr = (scores as any)[dim];\n if (Number.isNaN(scr)) continue;\n sigStmts.push({\n sql: `INSERT INTO discovery_signals (ticker, dimension, source, evidence, confidence, signal_time, payload) VALUES (?,?,?,?,?,?,?)`,\n args: [t, dim, \"discovery_v2_compute\", `score ${scr}/${DIM[dim as keyof typeof DIM]}`, coverage >= 0.6 ? \"high\" : \"medium\", now, JSON.stringify({ dimension_max: DIM[dim as keyof typeof DIM] })],\n });\n }\n }\n\n // Persist candidates + signals in batched writes (fast, no per-row awaits).\n for (let i = 0; i < candStmts.length; i += 20) await sqlite.batch(candStmts.slice(i, i + 20));\n for (let i = 0; i < sigStmts.length; i += 20) await sqlite.batch(sigStmts.slice(i, i + 20));\n\n // Scoped mode (onlyTickers set, e.g. a single-ticker controlled test):\n // never touch the shared themes table or append a production\n // discovery_runs row — blast radius stays limited to the requested\n // ticker(s)' own rows in discovery_candidates/discovery_signals\n // (upserted above via ON CONFLICT(ticker), so no other ticker's row is\n // read, written, or otherwise affected).\n if (!onlyTickers) {\n await sqlite.execute(`DELETE FROM themes`);\n const thStmts = themes.map((th) => ({\n sql: `INSERT INTO themes (name, label, trend, related_tickers, related_accounts, evidence, attention, first_detected, last_updated)\n VALUES (?,?,?,?,?,?,?,?,?)`,\n args: [th.name, th.label, th.trend, JSON.stringify(th.related_tickers), JSON.stringify(th.related_accounts),\n JSON.stringify(th.evidence), th.attention, th.first_detected, th.last_updated],\n }));\n for (let i = 0; i < thStmts.length; i += 20) await sqlite.batch(thStmts.slice(i, i + 20));\n\n await sqlite.execute({\n sql: `INSERT INTO discovery_runs (computed_at, candidates, themes, avg_coverage) VALUES (?,?,?,?)`,\n args: [now, consensus.length, themes.length, Math.round((65 / TOTAL_MAX) * 100) / 100],\n });\n }\n\n return { generatedAt: now, candidates: consensus.length, themes: themes.length };\n}\n\nasync function isFresh(): Promise {\n const cand = await sqlite.execute(`SELECT COUNT(*) c FROM discovery_candidates`);\n const last = await sqlite.execute(`SELECT MAX(computed_at) m FROM discovery_runs`);\n const m = (last.rows[0] as any)?.m ?? null;\n return Number((cand.rows[0] as any).c) > 0 && !!m && (Date.now() - new Date(m).getTime()) < FRESH_MS;\n}\n\nasync function ensureDiscoveryComputed(): Promise {\n await ensureV2Schema();\n if (await isFresh()) return;\n await computeDiscovery(); // only runs when stale/empty\n}\n\nasync function ensureEmergingAccounts(): Promise {\n await ensureV2Schema();\n const last = await sqlite.execute(`SELECT MAX(last_evaluated) m FROM emerging_accounts`);\n const m = (last.rows[0] as any)?.m ?? null;\n if (m && (Date.now() - new Date(m).getTime()) < FRESH_MS) return;\n await buildEmergingAccounts();\n}\n\nexport async function buildEmergingAccounts(): Promise {\n await ensureV2Schema();\n const { candidates } = await extractCandidates();\n const now = new Date().toISOString();\n const stmts: any[] = [];\n for (const c of candidates) {\n if (TRACKED_ACCOUNTS.some((h) => h.toLowerCase() === c.handle.toLowerCase())) continue;\n const category = c.has_ticker_evidence\n ? (c.times_referenced >= 3 ? \"Independent Market Voice\" : \"Sector Specialist\")\n : \"Emerging Account\";\n const confidence = c.api_resolved_evidence_count > 0 ? (c.times_referenced >= 3 ? \"Medium\" : \"Low\") : \"Low\";\n const promo = c.promotional_flag ? 1 : 0;\n stmts.push({\n sql: `INSERT INTO emerging_accounts (\n handle, category, times_referenced, referencing_accounts, tickers, evidence, promo_flagged,\n confidence, repost_ratio, last_evaluated, promotion_status, notes)\n VALUES (?,?,?,?,?,?,?,?,?,?,'Not Promoted',?)\n ON CONFLICT(handle) DO UPDATE SET category=excluded.category, times_referenced=excluded.times_referenced,\n referencing_accounts=excluded.referencing_accounts, tickers=excluded.tickers, evidence=excluded.evidence,\n promo_flagged=excluded.promo_flagged, confidence=excluded.confidence, repost_ratio=excluded.repost_ratio,\n last_evaluated=excluded.last_evaluated, notes=excluded.notes`,\n args: [c.handle, category, c.times_referenced, JSON.stringify(c.referencing_core_accounts),\n JSON.stringify(c.tickers_discussed), JSON.stringify(c.evidence.slice(0, 3)), promo,\n confidence, null, now,\n promo ? \"Flagged promotional — requires explicit review before any consideration.\" : \"Not promoted; explicit approval required to move toward tracked status.\"],\n });\n }\n for (let i = 0; i < stmts.length; i += 20) await sqlite.batch(stmts.slice(i, i + 20));\n const rows = (await sqlite.execute(`SELECT * FROM emerging_accounts ORDER BY times_referenced DESC LIMIT 60`)).rows as any[];\n return rows.map((r) => ({ ...r, referencing_accounts: JSON.parse(r.referencing_accounts || \"[]\"), tickers: JSON.parse(r.tickers || \"[]\"), evidence: JSON.parse(r.evidence || \"[]\") }));\n}\n\nexport async function watchlistView(): Promise {\n await ensureV2Schema();\n const res = await sqlite.execute(`SELECT ticker, name, source, added_at FROM watchlist`);\n const cand = await sqlite.execute(`SELECT ticker, score_normalized, tier, direction, independent_consensus FROM discovery_candidates`);\n const candMap = new Map((cand.rows as any[]).map((r) => [r.ticker.toUpperCase(), r]));\n return (res.rows as any[]).map((w) => {\n const c = candMap.get(w.ticker.toUpperCase());\n return { ticker: w.ticker, name: w.name ?? null, source: w.source, added_at: w.added_at,\n status: c ? \"Existing Watchlist\" : \"Watchlist (no consensus data yet)\",\n score_normalized: c?.score_normalized ?? null, tier: c?.tier ?? null,\n direction: c?.direction ?? null, independent_consensus: c?.independent_consensus ?? 0 };\n });\n}\n\nfunction decode(r: any) {\n return { ...r, catalysts: JSON.parse(r.catalysts || \"[]\"), evidence: JSON.parse(r.evidence || \"[]\"), sources: JSON.parse(r.sources || \"[]\"), is_watchlist: !!r.is_watchlist };\n}\n\nexport async function discoveryApi(opts: { tier?: string; ticker?: string; period?: string } = {}) {\n await ensureDiscoveryComputed();\n await ensureEmergingAccounts();\n const rows = (await sqlite.execute(`SELECT * FROM discovery_candidates ORDER BY score_normalized DESC, independent_consensus DESC LIMIT 300`)).rows as any[];\n const themes = (await sqlite.execute(`SELECT * FROM themes ORDER BY attention DESC LIMIT 100`)).rows as any[];\n const emerging = (await sqlite.execute(`SELECT * FROM emerging_accounts ORDER BY times_referenced DESC LIMIT 60`)).rows as any[];\n const lastRun = (await sqlite.execute(`SELECT computed_at FROM discovery_runs ORDER BY id DESC LIMIT 1`)).rows as any[];\n const wl = await watchlistView();\n\n let list = rows.map(decode);\n if (opts.tier) list = list.filter((c) => c.tier === opts.tier.toUpperCase());\n if (opts.ticker) list = list.filter((c) => c.ticker.toUpperCase() === opts.ticker!.toUpperCase());\n\n const tiers = { A: list.filter((c) => c.tier === \"A\").length, B: list.filter((c) => c.tier === \"B\").length, C: list.filter((c) => c.tier === \"C\").length };\n const newToday = list.filter((c) => c.first_detected && c.first_detected >= new Date(Date.now() - 86400000).toISOString()).length;\n const attentionRising = list.filter((c) => c.attention_trend && c.attention_trend.startsWith(\"rising\")).length;\n const highDisagreement = list.filter((c) => c.disagreement_level === \"High\" || c.disagreement_level === \"Moderate\").length;\n const nonWatchlist = list.filter((c) => !c.is_watchlist).length;\n\n const themesOut = themes.map((t) => ({ ...t, related_tickers: JSON.parse(t.related_tickers || \"[]\"), related_accounts: JSON.parse(t.related_accounts || \"[]\"), evidence: JSON.parse(t.evidence || \"[]\") }));\n\n return {\n generatedAt: lastRun[0]?.computed_at ?? null, period: opts.period ?? \"daily\",\n summary: { totalCandidates: list.length, tiers, newToday, attentionRising, highDisagreement, nonWatchlist },\n candidates: list, themes: themesOut,\n emergingAccounts: emerging, watchlist: wl, watchlistKeys: WATCHLIST,\n };\n}\n\nexport async function claudeContext(): Promise {\n await ensureDiscoveryComputed();\n await ensureEmergingAccounts();\n const d = await sqlite.execute(`SELECT * FROM discovery_candidates ORDER BY score_normalized DESC LIMIT 8`);\n const top = (d.rows as any[]).map((r) => ({ ticker: r.ticker, score: r.score_normalized, tier: r.tier, direction: r.direction, accounts: r.independent_consensus, watch: !!r.is_watchlist }));\n const consensus = await sqlite.execute(`SELECT tc.ticker, tc.consensus_direction, tc.disagreement_level, tc.unique_accounts_mentioning\n FROM ticker_consensus tc INNER JOIN (SELECT ticker, MAX(computed_at) mc FROM ticker_consensus GROUP BY ticker) l\n ON tc.ticker=l.ticker AND tc.computed_at=l.mc ORDER BY tc.unique_accounts_mentioning DESC LIMIT 10`);\n const themes = await sqlite.execute(`SELECT name, attention, related_tickers FROM themes ORDER BY attention DESC LIMIT 6`);\n const acct = await sqlite.execute(`SELECT handle, category, promotion_status, promo_flagged FROM emerging_accounts WHERE promo_flagged=0 ORDER BY times_referenced DESC LIMIT 8`);\n const lastRun = await sqlite.execute(`SELECT computed_at FROM discovery_runs ORDER BY id DESC LIMIT 1`);\n const mdCount = Number((await sqlite.execute(`SELECT COUNT(*) c FROM market_data`)).rows[0]?.c ?? 0);\n const mdcfg = !!Deno.env.get(\"MARKETDATA_API_KEY\");\n const miss = mdcfg\n ? (mdCount > 0\n ? `${mdCount} tickers have market-data rows; price/fundamentals/valuation/flow dimensions are scored where data exists.`\n : \"Market-data provider configured, but no data fetched yet — those dimensions remain insufficient data (not zero).\")\n : \"Market-data provider key (MARKETDATA_API_KEY) not configured — price/fundamentals/valuation/flow dimensions remain insufficient data (not zero).\";\n return {\n system_status: { ok: true, last_discovery_computed: (lastRun.rows[0] as any)?.computed_at ?? null },\n last_update: (await sqlite.execute(`SELECT MAX(analyzed_at) m FROM posts WHERE analyzed_at IS NOT NULL`)).rows[0]?.m ?? null,\n top_consensus_stocks: consensus.rows,\n top_discovery_candidates: top,\n emerging_themes: (themes.rows as any[]).map((r) => ({ theme: r.name, attention: r.attention, tickers: JSON.parse(r.related_tickers || \"[]\") })),\n disagreement_major: (await sqlite.execute(`SELECT tc.ticker, tc.disagreement_level, tc.disagreement_score\n FROM ticker_consensus tc INNER JOIN (SELECT ticker, MAX(computed_at) mc FROM ticker_consensus GROUP BY ticker) l\n ON tc.ticker=l.ticker AND tc.computed_at=l.mc WHERE tc.disagreement_level='High' OR tc.disagreement_level='Moderate' LIMIT 6`)).rows,\n new_candidates_today: (await sqlite.execute(`SELECT ticker, score_normalized, tier FROM discovery_candidates WHERE first_detected >= datetime('now','-1 day') LIMIT 10`)).rows,\n emerging_accounts_to_review: acct.rows,\n data_coverage_warnings: miss,\n };\n}\n\nexport async function claudeMarketBrief(): Promise {\n const ctx: any = await claudeContext();\n const d = await sqlite.execute(`SELECT ticker, attention_trend, score_normalized, tier FROM discovery_candidates ORDER BY score_normalized DESC`);\n const cands = (d.rows as any[]);\n return {\n as_of: ctx.last_update,\n discovery: {\n daily: { brief: \"Daily Market Discovery Brief — top new candidates, attention shifts, disagreements.\", candidates: cands.filter((c) => c.tier === \"A\" || c.tier === \"B\").slice(0, 12) },\n weekly: { brief: \"Weekly Market Discovery Review — consensus strengthen/weaken, emerging accounts/themes.\", candidates: cands.slice(0, 15) },\n monthly: { brief: \"Monthly Consensus Review — sustained themes, momentum, watchlist vs new.\", candidates: cands.slice(0, 20) },\n },\n top_candidates: ctx.top_discovery_candidates,\n themes: ctx.emerging_themes,\n acknowledgements: [\"Not investment advice; research prioritization only. Dimensions without data are marked insufficient, never zeroed as a negative signal.\"],\n };\n}\n","deleted",false,"label",[47,24],"lib","5c601043-49d8-4da3-8595-d21725bb7d55",["D",1788705462303],"deletedAt","legacyEnvVars","privacy","public","httpPrivacy","description","X account collector: normalized posts into SQLite","imageUrl","userId","8f36435b-4ee1-42d7-a52e-11508b5c71b0","pinnedFileId","b9b4c5ce-aa00-11f1-b6d1-1607ee4eb77e","pinPreview","pinMigrated","screenshotVersion","likeCount",0,"orgId","01a07719-4bcc-778e-997b-14b67712b6b9","tursoName","5c60104349d84da38595d21725bb7d55","tursoHostname","5c60104349d84da38595d21725bb7d55-val-town.aws-us-east-2.turso.io","tursoMeasuredAt",["D",1789485645684],"pinned","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm","user",{"_9":10,"_109":105},"org",{"_9":10,"_102":103,"_104":105,"_58":59,"_106":107,"_108":44},"branch",{"_37":92,"_11":93,"_22":30,"_94":48,"_95":-5,"_96":-5,"_97":-5,"_98":-5,"_99":-5,"_100":12,"_101":10},"isOwned",146,"main","projectId","forkedBranchId","forkedBranchVersion","forkedProjectPrivacy","forkedProjectId","forkedProjectOrgId","projectName","userHandle","displayName","Efren Lin","avatar","https://lh3.googleusercontent.com/a/ACg8ocITAXmIR316-M-ReEEDUIUESLjIDNnYnCPqoEpOrKLJExkkkWU=s96-c","tier","free","canRestrictVals","profileImageUrl","ctx",{"_85":-5,"_143":144,"_145":146,"_147":148,"_149":44,"_150":44,"_151":44,"_152":34,"_153":-5,"_154":-7,"_155":-7,"_156":66,"_157":44,"_158":44,"_159":-5,"_160":44,"_161":44,"_162":44,"_163":-5},"ENV",{"_114":115,"_116":117,"_118":93,"_119":120,"_121":122,"_123":124,"_125":126,"_127":128,"_129":130,"_131":132,"_133":134,"_135":136,"_137":138,"_139":140,"_141":142},"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"]