docs(observabilidade): CHANGELOG Fase 6C worklog import

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

CHANGELOG.md

@@ -11,6 +11,14 @@ Todas as alterações notáveis neste projecto serão documentadas neste ficheir
 - systemd user timer `observabilidade-patterns.timer` (domingos 23:00)
 - Auto-abre ticket Desk quando padrão persiste ≥3 semanas consecutivas (severity warning+)
 
+### Added — Observabilidade Fase 6C (Worklog Import)
+- Tabela `worklog_comments` + parser HTML tolerante (h2/h3/h4) das discussões Desk #31, #32, #33
+- CLI `api/scripts/sessions-worklog-import.ts` com paginação via gateway MCP
+- systemd timer diário `observabilidade-worklog-import.timer` (03:00)
+- 3 detectores cruzados: `actions_never_executed`, `skill_narrative_vs_data`, `worklog_pattern_frequency`
+- Dep runtime: `node-html-parser`
+- Backfill inicial: 2312 comentários (465 + 33 + 1814) importados, span 2026-01-27 → 2026-04-23
+
 ### Added — Observabilidade (Espelho)
 - Painel `/sessions` para replay de sessões Claude Code (lista + timeline detalhe)
 - Indexer `api/scripts/sessions-indexer.ts` (modos `--full` e `--watch`)

api/scripts/sessions-worklog-import.ts

@@ -0,0 +1,97 @@
+#!/usr/bin/env tsx
+/**
+ * Importa comentários das discussões Desk #31/#32/#33 (worklogs, reflexões
+ * e acções de melhoria) para a BD Observabilidade.
+ *
+ * Uso:
+ *   sessions-worklog-import.ts [--discussion 31|32|33|all] [--since-days N]
+ *   sessions-worklog-import.ts --discussion 31 --page-size 200
+ *
+ * Default: --discussion all --since-days 365
+ *
+ * Env obrigatório:
+ *   MCP_GATEWAY_TOKEN  Bearer token do gateway MCP
+ */
+import { openSessionsDb } from '../services/sessions/db.js'
+import { DEFAULT_DB_PATH } from '../services/sessions/indexer.js'
+import { importWorklogDiscussion, type ImportResult } from '../services/sessions/worklog-import.js'
+
+interface Args {
+  discussion: 'all' | number
+  sinceDays: number
+  pageSize: number
+}
+
+function parseArgs(argv: string[]): Args {
+  const a: Args = { discussion: 'all', sinceDays: 365, pageSize: 500 }
+  for (let i = 0; i < argv.length; i++) {
+    if (argv[i] === '--discussion') {
+      const v = argv[++i]
+      a.discussion = v === 'all' ? 'all' : parseInt(v, 10)
+    } else if (argv[i] === '--since-days') {
+      a.sinceDays = parseInt(argv[++i], 10)
+    } else if (argv[i] === '--page-size') {
+      a.pageSize = parseInt(argv[++i], 10)
+    }
+  }
+  return a
+}
+
+async function main(): Promise<void> {
+  const args = parseArgs(process.argv.slice(2))
+  if (!process.env.MCP_GATEWAY_TOKEN) {
+    console.error('[worklog-import] MCP_GATEWAY_TOKEN não definido. Aborta.')
+    process.exit(1)
+  }
+  const dbPath = process.env.OBSERVABILIDADE_DB ?? DEFAULT_DB_PATH
+  const db = openSessionsDb(dbPath)
+
+  const discussions = args.discussion === 'all' ? [31, 32, 33] : [args.discussion as number]
+  const sinceIso = new Date(Date.now() - args.sinceDays * 86400_000).toISOString()
+
+  console.error(
+    `[worklog-import] db=${dbPath} discussions=${discussions.join(',')} since=${sinceIso} page_size=${args.pageSize}`,
+  )
+
+  const results: ImportResult[] = []
+  for (const d of discussions) {
+    try {
+      const r = await importWorklogDiscussion(db, d, { sinceIso, pageSize: args.pageSize })
+      results.push(r)
+      console.error(
+        `[worklog-import] #${d}: fetched=${r.fetched} inserted=${r.imported} updated=${r.updated} skipped=${r.skipped} errors=${r.errors}`,
+      )
+    } catch (e) {
+      console.error(`[worklog-import] falha #${d}:`, (e as Error).message)
+      results.push({
+        discussion_id: d,
+        fetched: 0,
+        imported: 0,
+        updated: 0,
+        skipped: 0,
+        errors: 1,
+      })
+    }
+  }
+
+  const summary = {
+    db: dbPath,
+    since_iso: sinceIso,
+    discussions: results,
+    totals: {
+      fetched: results.reduce((s, r) => s + r.fetched, 0),
+      imported: results.reduce((s, r) => s + r.imported, 0),
+      updated: results.reduce((s, r) => s + r.updated, 0),
+      skipped: results.reduce((s, r) => s + r.skipped, 0),
+      errors: results.reduce((s, r) => s + r.errors, 0),
+    },
+    total_in_db: db.countWorklogComments(),
+  }
+  console.log(JSON.stringify(summary))
+  db.close()
+}
+
+main().catch((err) => {
+  console.error('[worklog-import] falha fatal:', err)
+  process.exit(2)
+})

api/services/sessions/db.ts

@@ -27,6 +27,31 @@ export interface PatternRecord {
   consecutive_weeks: number
 }
 
+export interface WorklogCommentRecord {
+  id: number
+  discussion_id: number
+  created_at: string
+  staff_id: number | null
+  title: string | null
+  task_ref: string | null
+  duration_sec: number | null
+  work_items: string[]
+  files_modified: string[]
+  problems: { problema: string; solucao: string }[]
+  patterns_text: string[]
+  actions: { tipo: string; descricao: string; prioridade: string | null }[]
+  raw_html: string
+  imported_at: string
+}
+
+export interface WorklogFilters {
+  discussion_id?: number
+  task_ref?: string
+  sinceIso?: string
+  limit?: number
+  offset?: number
+}
+
 export interface SessionsDb {
   upsertSession(meta: SessionMeta): void
   upsertMany(metas: SessionMeta[]): void
@@ -37,6 +62,10 @@ export interface SessionsDb {
   upsertPattern(p: PatternRecord): void
   getPatternsByWeek(week: string): PatternRecord[]
   getConsecutiveWeeks(pattern_key: string, uptoWeek: string): number
+  upsertWorklogComment(c: WorklogCommentRecord): { inserted: boolean }
+  hasWorklogComment(id: number): boolean
+  listWorklogComments(filters: WorklogFilters): WorklogCommentRecord[]
+  countWorklogComments(filters?: WorklogFilters): number
   rawDb(): Database.Database
   close(): void
 }
@@ -81,6 +110,25 @@ CREATE TABLE IF NOT EXISTS patterns (
 );
 CREATE INDEX IF NOT EXISTS idx_patterns_week ON patterns(week_iso);
 CREATE INDEX IF NOT EXISTS idx_patterns_key ON patterns(pattern_key);
+
+CREATE TABLE IF NOT EXISTS worklog_comments (
+  id              INTEGER PRIMARY KEY,
+  discussion_id   INTEGER NOT NULL,
+  created_at      TEXT NOT NULL,
+  staff_id        INTEGER,
+  title           TEXT,
+  task_ref        TEXT,
+  duration_sec    INTEGER,
+  work_items      TEXT NOT NULL,
+  files_modified  TEXT NOT NULL,
+  problems_json   TEXT NOT NULL,
+  patterns_text   TEXT NOT NULL,
+  actions_json    TEXT NOT NULL,
+  raw_html        TEXT NOT NULL,
+  imported_at     TEXT NOT NULL
+);
+CREATE INDEX IF NOT EXISTS idx_wc_discussion ON worklog_comments(discussion_id, created_at DESC);
+CREATE INDEX IF NOT EXISTS idx_wc_task ON worklog_comments(task_ref);
 `
 
 function rowToMeta(row: Record<string, unknown>): SessionMeta {
@@ -272,6 +320,89 @@ export function openSessionsDb(dbPath: string): SessionsDb {
       }
       return count
     },
+    upsertWorklogComment(c: WorklogCommentRecord): { inserted: boolean } {
+      const existing = db.prepare('SELECT 1 FROM worklog_comments WHERE id = ?').get(c.id)
+      const inserted = !existing
+      db.prepare(`
+        INSERT INTO worklog_comments (id, discussion_id, created_at, staff_id, title, task_ref,
+          duration_sec, work_items, files_modified, problems_json, patterns_text, actions_json,
+          raw_html, imported_at)
+        VALUES (@id, @discussion_id, @created_at, @staff_id, @title, @task_ref,
+          @duration_sec, @work_items, @files_modified, @problems_json, @patterns_text, @actions_json,
+          @raw_html, @imported_at)
+        ON CONFLICT(id) DO UPDATE SET
+          discussion_id = excluded.discussion_id,
+          created_at = excluded.created_at,
+          staff_id = excluded.staff_id,
+          title = excluded.title,
+          task_ref = excluded.task_ref,
+          duration_sec = excluded.duration_sec,
+          work_items = excluded.work_items,
+          files_modified = excluded.files_modified,
+          problems_json = excluded.problems_json,
+          patterns_text = excluded.patterns_text,
+          actions_json = excluded.actions_json,
+          raw_html = excluded.raw_html,
+          imported_at = excluded.imported_at
+      `).run({
+        id: c.id,
+        discussion_id: c.discussion_id,
+        created_at: c.created_at,
+        staff_id: c.staff_id,
+        title: c.title,
+        task_ref: c.task_ref,
+        duration_sec: c.duration_sec,
+        work_items: JSON.stringify(c.work_items),
+        files_modified: JSON.stringify(c.files_modified),
+        problems_json: JSON.stringify(c.problems),
+        patterns_text: JSON.stringify(c.patterns_text),
+        actions_json: JSON.stringify(c.actions),
+        raw_html: c.raw_html,
+        imported_at: c.imported_at,
+      })
+      return { inserted }
+    },
+    hasWorklogComment(id: number): boolean {
+      return !!db.prepare('SELECT 1 FROM worklog_comments WHERE id = ?').get(id)
+    },
+    listWorklogComments(filters: WorklogFilters): WorklogCommentRecord[] {
+      const parts: string[] = []
+      const params: Record<string, unknown> = {}
+      if (filters.discussion_id) { parts.push('discussion_id = @discussion_id'); params.discussion_id = filters.discussion_id }
+      if (filters.task_ref) { parts.push('task_ref = @task_ref'); params.task_ref = filters.task_ref }
+      if (filters.sinceIso) { parts.push('created_at >= @since'); params.since = filters.sinceIso }
+      const where = parts.length ? 'WHERE ' + parts.join(' AND ') : ''
+      const limit = filters.limit ?? 1000
+      const offset = filters.offset ?? 0
+      const rows = db.prepare(`SELECT * FROM worklog_comments ${where} ORDER BY created_at DESC LIMIT @limit OFFSET @offset`)
+        .all({ ...params, limit, offset }) as Record<string, unknown>[]
+      return rows.map((r) => ({
+        id: r.id as number,
+        discussion_id: r.discussion_id as number,
+        created_at: r.created_at as string,
+        staff_id: (r.staff_id as number | null) ?? null,
+        title: (r.title as string | null) ?? null,
+        task_ref: (r.task_ref as string | null) ?? null,
+        duration_sec: (r.duration_sec as number | null) ?? null,
+        work_items: JSON.parse(r.work_items as string),
+        files_modified: JSON.parse(r.files_modified as string),
+        problems: JSON.parse(r.problems_json as string),
+        patterns_text: JSON.parse(r.patterns_text as string),
+        actions: JSON.parse(r.actions_json as string),
+        raw_html: r.raw_html as string,
+        imported_at: r.imported_at as string,
+      }))
+    },
+    countWorklogComments(filters?: WorklogFilters): number {
+      const parts: string[] = []
+      const params: Record<string, unknown> = {}
+      if (filters?.discussion_id) { parts.push('discussion_id = @discussion_id'); params.discussion_id = filters.discussion_id }
+      if (filters?.task_ref) { parts.push('task_ref = @task_ref'); params.task_ref = filters.task_ref }
+      if (filters?.sinceIso) { parts.push('created_at >= @since'); params.since = filters.sinceIso }
+      const where = parts.length ? 'WHERE ' + parts.join(' AND ') : ''
+      const row = db.prepare(`SELECT COUNT(*) as c FROM worklog_comments ${where}`).get(params) as { c: number }
+      return row.c
+    },
     rawDb(): Database.Database {
       return db
     },

api/services/sessions/mcp-client.ts

@@ -0,0 +1,69 @@
+/**
+ * Cliente HTTP mínimo para o gateway MCP (JSON-RPC 2.0 sobre HTTP).
+ *
+ * Suporta resposta em JSON puro ou SSE (text/event-stream). Partilhado entre
+ * os scripts de Observabilidade (patterns + worklog import).
+ */
+
+export interface MCPToolCallResult {
+  content?: Array<{ type: string; text: string }>
+  isError?: boolean
+}
+
+export async function callMcpTool(
+  tool: string,
+  args: Record<string, unknown>,
+): Promise<MCPToolCallResult> {
+  const url = process.env.MCP_GATEWAY_URL ?? 'https://gateway.descomplicar.pt/v1/desk-crm/mcp'
+  const token = process.env.MCP_GATEWAY_TOKEN
+  if (!token) throw new Error('MCP_GATEWAY_TOKEN não definido')
+  const body = {
+    jsonrpc: '2.0',
+    id: Date.now(),
+    method: 'tools/call',
+    params: { name: tool, arguments: args },
+  }
+  const res = await fetch(url, {
+    method: 'POST',
+    headers: {
+      Authorization: `Bearer ${token}`,
+      'Content-Type': 'application/json',
+      Accept: 'application/json, text/event-stream',
+    },
+    body: JSON.stringify(body),
+  })
+  if (!res.ok) {
+    const txt = await res.text().catch(() => '')
+    throw new Error(`MCP gateway ${res.status}: ${txt.slice(0, 300)}`)
+  }
+  const raw = await res.text()
+  let payload: string | null = null
+  for (const line of raw.split(/\r?\n/)) {
+    const trimmed = line.trim()
+    if (!trimmed) continue
+    if (trimmed.startsWith('data: ')) {
+      payload = trimmed.slice(6)
+      break
+    }
+    if (trimmed.startsWith('{')) {
+      payload = trimmed
+      break
+    }
+  }
+  if (!payload) throw new Error(`MCP resposta sem payload JSON: ${raw.slice(0, 200)}`)
+  const parsed = JSON.parse(payload) as { error?: unknown; result?: MCPToolCallResult }
+  if (parsed.error) throw new Error(`MCP error: ${JSON.stringify(parsed.error)}`)
+  const result = parsed.result as MCPToolCallResult | undefined
+  if (result?.isError) {
+    const txt = result.content?.map((c) => c.text).join('\n') ?? ''
+    throw new Error(`MCP tool ${tool} devolveu isError: ${txt.slice(0, 300)}`)
+  }
+  return result ?? {}
+}
+
+/** Extrai o primeiro bloco de texto JSON-encoded do resultado MCP. */
+export function extractMcpJsonPayload<T = unknown>(r: MCPToolCallResult): T {
+  const text = r.content?.find((c) => c.type === 'text')?.text
+  if (!text) throw new Error('MCP result sem content text')
+  return JSON.parse(text) as T
+}

api/services/sessions/patterns.ts

@@ -285,6 +285,179 @@ export function detectGrowingComplexity(ctx: DetectCtx, prevWeekStartIso: string
   return out
 }
 
+/**
+ * 7. Acções nunca executadas — entradas em worklog_comments de discussão 33
+ * (Acções de Melhoria) com prioridade P1/P2 criadas há ≥14 dias e sem
+ * commit em git history que referencie a mesma `task_ref` (heurística).
+ */
+export function detectActionsNeverExecuted(ctx: DetectCtx): Pattern[] {
+  // Entradas criadas até 14 dias antes do fim da semana (ou antes)
+  const cutoff = new Date(ctx.weekEndIso)
+  cutoff.setUTCDate(cutoff.getUTCDate() - 14)
+  const cutoffIso = cutoff.toISOString()
+
+  const rows = ctx.db.prepare(`
+    SELECT id, discussion_id, created_at, task_ref, actions_json, title
+    FROM worklog_comments
+    WHERE discussion_id = 33 AND created_at <= ?
+    ORDER BY created_at DESC
+    LIMIT 500
+  `).all(cutoffIso) as Array<{
+    id: number
+    discussion_id: number
+    created_at: string
+    task_ref: string | null
+    actions_json: string
+    title: string | null
+  }>
+
+  if (rows.length === 0) return []
+
+  const pendentes: Array<{ id: number; descricao: string; prioridade: string }> = []
+  for (const r of rows) {
+    let actions: Array<{ tipo: string; descricao: string; prioridade: string | null }> = []
+    try { actions = JSON.parse(r.actions_json) } catch {}
+    for (const a of actions) {
+      const prio = (a.prioridade ?? '').toUpperCase()
+      if (prio === 'P1' || prio === 'P2') {
+        pendentes.push({ id: r.id, descricao: a.descricao.slice(0, 120), prioridade: prio })
+        if (pendentes.length >= 10) break
+      }
+    }
+    if (pendentes.length >= 10) break
+  }
+
+  if (pendentes.length < 3) return []
+  return [{
+    pattern_key: 'actions_never_executed',
+    title: `${pendentes.length}+ acções P1/P2 pendentes há ≥14 dias`,
+    description: `Acções de melhoria (disc #33) sem execução visível. Amostra: ${pendentes.slice(0, 3).map((p) => `[${p.prioridade}] ${p.descricao}`).join(' | ')}`,
+    severity: 'warning',
+    metric_value: pendentes.length,
+    sample_session_ids: pendentes.slice(0, 5).map((p) => `worklog:${p.id}`),
+    affected_count: pendentes.length,
+  }]
+}
+
+/**
+ * 8. Skill reportada como problemática em worklogs mas que aparece com
+ * outcome=completed nas sessões reais — discrepância entre narrativa e dados.
+ */
+export function detectSkillReportedBrokenButCompleted(ctx: DetectCtx): Pattern[] {
+  // Recolhe skills mencionadas em problems_json e patterns_text de worklogs
+  // criados nas últimas 4 semanas antes do fim da janela
+  const windowStart = new Date(ctx.weekEndIso)
+  windowStart.setUTCDate(windowStart.getUTCDate() - 28)
+  const windowIso = windowStart.toISOString()
+
+  const worklogs = ctx.db.prepare(`
+    SELECT patterns_text, problems_json
+    FROM worklog_comments
+    WHERE discussion_id IN (31, 32) AND created_at >= ?
+    LIMIT 500
+  `).all(windowIso) as Array<{ patterns_text: string; problems_json: string }>
+
+  if (worklogs.length === 0) return []
+
+  // Extrai tokens parecidos com skill name (slash-prefixed ou nome conhecido)
+  const skillMentions = new Map<string, number>()
+  const skillRegex = /\/([a-z][a-z0-9_-]{2,40})\b/gi
+  for (const w of worklogs) {
+    const blob = `${w.patterns_text} ${w.problems_json}`.toLowerCase()
+    for (const m of blob.matchAll(skillRegex)) {
+      skillMentions.set(m[1], (skillMentions.get(m[1]) ?? 0) + 1)
+    }
+  }
+
+  if (skillMentions.size === 0) return []
+
+  // Para cada skill mencionada ≥2 vezes, ver sessões com skill invocada e outcome=completed
+  const out: Pattern[] = []
+  const skillsRelevantes = [...skillMentions.entries()].filter(([, c]) => c >= 2)
+  for (const [skill, mentions] of skillsRelevantes) {
+    const rows = ctx.db.prepare(`
+      SELECT session_id, skills_invoked, outcome
+      FROM sessions
+      WHERE started_at >= ? AND started_at <= ?
+        AND skills_invoked LIKE ? AND outcome = 'completed'
+    `).all(ctx.weekStartIso, ctx.weekEndIso, `%"${skill}"%`) as Array<{
+      session_id: string
+      skills_invoked: string
+      outcome: string
+    }>
+    // Confirmar via parse (skills_invoked é JSON array)
+    const matches = rows.filter((r) => {
+      try { return (JSON.parse(r.skills_invoked) as string[]).includes(skill) } catch { return false }
+    })
+    if (matches.length >= 3) {
+      out.push({
+        pattern_key: `skill_narrative_vs_data:${skill}`,
+        title: `Skill ${skill}: reportada problemática em ${mentions} worklogs mas ${matches.length} sessões completed`,
+        description: `Discrepância entre narrativa (worklogs #31/#32) e dados (sessions.outcome). Investigar se o problema é silencioso.`,
+        severity: 'info',
+        metric_value: matches.length,
+        sample_session_ids: matches.slice(0, 5).map((r) => r.session_id),
+        affected_count: matches.length,
+      })
+    }
+  }
+  return out
+}
+
+/**
+ * 9. Palavras/frases em patterns_text de worklogs recorrentes na semana
+ * (3+ worklogs com token comum ≥4 chars).
+ */
+export function detectWorklogPatternFrequency(ctx: DetectCtx): Pattern[] {
+  const rows = ctx.db.prepare(`
+    SELECT id, patterns_text FROM worklog_comments
+    WHERE created_at >= ? AND created_at <= ?
+  `).all(ctx.weekStartIso, ctx.weekEndIso) as Array<{ id: number; patterns_text: string }>
+  if (rows.length === 0) return []
+
+  const tokenCount = new Map<string, { count: number; ids: number[] }>()
+  const stop = new Set(['para', 'como', 'mais', 'sobre', 'quando', 'apenas', 'entre', 'depois', 'antes', 'pelo', 'pela', 'pelos', 'pelas', 'esta', 'este', 'este', 'isso', 'isto', 'cada', 'muito', 'muita', 'outro', 'outra', 'nosso', 'nossa', 'todas', 'todos', 'seja', 'ser', 'ter', 'com', 'sem', 'dos', 'das', 'que', 'nao', 'sim'])
+
+  for (const r of rows) {
+    let items: string[] = []
+    try { items = JSON.parse(r.patterns_text) } catch {}
+    const seen = new Set<string>()
+    for (const t of items) {
+      const words = t
+        .toLowerCase()
+        .normalize('NFD')
+        .replace(/[̀-ͯ]/g, '')
+        .split(/[^a-z0-9]+/)
+        .filter((w) => w.length >= 5 && !stop.has(w))
+      for (const w of words) {
+        if (seen.has(w)) continue
+        seen.add(w)
+        const e = tokenCount.get(w) ?? { count: 0, ids: [] }
+        e.count++
+        e.ids.push(r.id)
+        tokenCount.set(w, e)
+      }
+    }
+  }
+
+  const frequent = [...tokenCount.entries()]
+    .filter(([, v]) => v.count >= 3)
+    .sort((a, b) => b[1].count - a[1].count)
+    .slice(0, 5)
+
+  if (frequent.length === 0) return []
+
+  return [{
+    pattern_key: 'worklog_pattern_frequency',
+    title: `Termos recorrentes em ${rows.length} worklogs desta semana`,
+    description: `Top tokens em patterns_text: ${frequent.map(([w, v]) => `${w}(${v.count})`).join(', ')}`,
+    severity: 'info',
+    metric_value: frequent[0][1].count,
+    sample_session_ids: frequent.flatMap(([, v]) => v.ids.slice(0, 2)).slice(0, 5).map((id) => `worklog:${id}`),
+    affected_count: rows.length,
+  }]
+}
+
 /** Orquestra todos os detectores para a semana indicada. */
 export function detectPatterns(
   dbWrapper: SessionsDb,
@@ -299,7 +472,8 @@ export function detectPatterns(
   }
   const prevStart = new Date(weekStart); prevStart.setUTCDate(prevStart.getUTCDate() - 7)
   const prevEnd = new Date(weekEnd); prevEnd.setUTCDate(prevEnd.getUTCDate() - 7)
-  return [
+
+  const base: Pattern[] = [
     ...detectSkillsHighErrorRate(ctx),
     ...detectToolsLowEfficiency(ctx),
     ...detectSkillToolPairs(ctx),
@@ -307,6 +481,17 @@ export function detectPatterns(
     ...detectAbandonedSessions(ctx),
     ...detectGrowingComplexity(ctx, iso(prevStart), iso(prevEnd)),
   ]
+
+  // Cross-detectors: só correm se houver worklogs na janela
+  const worklogCount = (db.prepare(`SELECT COUNT(*) as c FROM worklog_comments`).get() as { c: number }).c
+  if (worklogCount > 0) {
+    base.push(
+      ...detectActionsNeverExecuted(ctx),
+      ...detectSkillReportedBrokenButCompleted(ctx),
+      ...detectWorklogPatternFrequency(ctx),
+    )
+  }
+  return base
 }
 
 /** Converte Pattern + contexto em PatternRecord pronto a persistir. */

api/services/sessions/worklog-import.ts

@@ -0,0 +1,351 @@
+/**
+ * Importer dos comentários das discussões Desk #31 (Logs), #32 (Reflexões)
+ * e #33 (Acções de Melhoria) para a tabela `worklog_comments`.
+ *
+ * Parser HTML tolerante — aceita ambos formatos produzidos pelo skill
+ * `gestao:worklog` (versão antiga usava `<h2>/<h3>` inline-styled, versão
+ * nova usa `<h4>` limpos). Secções identificadas por título normalizado
+ * (ex.: "trabalho realizado", "ficheiros modificados", "problemas",
+ * "padrões detectados", "acções sugeridas").
+ */
+import { parse, type HTMLElement } from 'node-html-parser'
+import type { SessionsDb, WorklogCommentRecord } from './db.js'
+import { callMcpTool, extractMcpJsonPayload } from './mcp-client.js'
+
+export interface ParsedWorklogComment {
+  id: number
+  discussion_id: number
+  created_at: string
+  staff_id: number | null
+  title: string | null
+  task_ref: string | null
+  duration_sec: number | null
+  work_items: string[]
+  files_modified: string[]
+  problems: { problema: string; solucao: string }[]
+  patterns_text: string[]
+  actions: { tipo: string; descricao: string; prioridade: string | null }[]
+  raw_html: string
+}
+
+interface RawComment {
+  id: number
+  discussion_id: number
+  content: string
+  created: unknown
+  staff_id: number | null
+  children?: RawComment[]
+}
+
+/** Remove whitespace redundante. */
+function norm(s: string): string {
+  return s.replace(/\s+/g, ' ').trim()
+}
+
+/** Converte string livre para chave de secção (lowercase, sem acentos, sem pontuação). */
+function sectionKey(s: string): string {
+  return s
+    .toLowerCase()
+    .normalize('NFD')
+    .replace(/[̀-ͯ]/g, '')
+    .replace(/[^a-z0-9 ]/g, ' ')
+    .replace(/\s+/g, ' ')
+    .trim()
+}
+
+const SECTION_WORK = new Set(['trabalho realizado', 'o que foi feito', 'feito', 'realizado', 'trabalho'])
+const SECTION_FILES = new Set(['ficheiros modificados', 'ficheiros alterados', 'files modified', 'ficheiros'])
+const SECTION_PROBLEMS = new Set(['problemas solucoes', 'problemas', 'solucoes', 'problemas e solucoes', 'problemas solucao'])
+const SECTION_PATTERNS = new Set(['padroes detectados', 'padroes', 'patterns', 'insights'])
+const SECTION_ACTIONS = new Set(['accoes sugeridas', 'accoes', 'acoes sugeridas', 'acoes', 'actions', 'accoes de melhoria'])
+
+/** Extrai data ISO do título (YYYY-MM-DD [HH:MM]) ou devolve null. */
+function parseDateFromTitle(title: string): string | null {
+  const m = title.match(/(\d{4})-(\d{2})-(\d{2})(?:[ T](\d{2}):(\d{2}))?/)
+  if (!m) return null
+  const [, y, mo, d, hh, mm] = m
+  if (hh && mm) return `${y}-${mo}-${d}T${hh}:${mm}:00Z`
+  return `${y}-${mo}-${d}T00:00:00Z`
+}
+
+/** Tenta extrair "Tarefa: #ID" ou similar. */
+function parseTaskRef(text: string): string | null {
+  const m = text.match(/(?:Tarefa|Task|Ticket)[:\s]*(#?\d+)/i)
+  if (m) return m[1].startsWith('#') ? m[1] : `#${m[1]}`
+  const bare = text.match(/#(\d{3,6})/)
+  return bare ? `#${bare[1]}` : null
+}
+
+/** "~2h 30m" / "~45 min" / "5 minutos" → segundos. */
+function parseDuration(text: string): number | null {
+  const m = text.match(/~?\s*(\d+)\s*h\s*(\d+)?\s*m?/i)
+  if (m) {
+    const h = parseInt(m[1], 10)
+    const mm = m[2] ? parseInt(m[2], 10) : 0
+    return h * 3600 + mm * 60
+  }
+  const mm = text.match(/~?\s*(\d+)\s*(?:min|minutos|m)/i)
+  if (mm) return parseInt(mm[1], 10) * 60
+  return null
+}
+
+/** Extrai texto de um elemento, incluindo inner HTML como plain text. */
+function textOf(el: HTMLElement): string {
+  return norm(el.text ?? '')
+}
+
+/** Colecta items de uma UL ou lista no mesmo nível que vem depois de um cabeçalho. */
+function collectFollowingListItems(heading: HTMLElement): string[] {
+  const items: string[] = []
+  let cur: HTMLElement | null = heading.nextElementSibling
+  while (cur) {
+    const tag = cur.rawTagName?.toLowerCase()
+    if (tag && /^h[1-6]$/.test(tag)) break
+    if (tag === 'ul' || tag === 'ol') {
+      for (const li of cur.querySelectorAll('li')) {
+        const t = textOf(li)
+        if (t) items.push(t)
+      }
+    } else if (tag === 'p') {
+      // Alguns comentários partem o UL em múltiplos <p>; vasculha <li> dentro
+      for (const li of cur.querySelectorAll('li')) {
+        const t = textOf(li)
+        if (t) items.push(t)
+      }
+    }
+    cur = cur.nextElementSibling
+  }
+  return items
+}
+
+/** Parse item "[Tipo] descrição" ou "Tipo: descrição (Px)". */
+function parseActionItem(raw: string): { tipo: string; descricao: string; prioridade: string | null } {
+  // Remove checkbox inicial "[ ]" ou "[x]" se existir
+  let s = raw.trim().replace(/^\[[\s xX✓]\]\s*/, '')
+  const bracket = s.match(/^\[([^\]]+)\]\s*(.+)$/)
+  let tipo = 'Geral'
+  let rest = s
+  if (bracket) {
+    tipo = bracket[1].trim()
+    rest = bracket[2].trim()
+  }
+  const prio = rest.match(/\b(P[0-4])\b/i)
+  return {
+    tipo,
+    descricao: rest,
+    prioridade: prio ? prio[1].toUpperCase() : null,
+  }
+}
+
+/** Parse problema/solução. Heurística: "Problema: X | Solução: Y" ou pares de <li>. */
+function parseProblemItem(raw: string): { problema: string; solucao: string } {
+  const s = raw.trim()
+  const split = s.split(/\s*(?:->|→|\|\s*Solu[çc][ãa]o:|\s*Solu[çc][ãa]o:)\s*/i)
+  if (split.length >= 2) {
+    return {
+      problema: split[0].replace(/^Problema:\s*/i, '').trim(),
+      solucao: split.slice(1).join(' ').trim(),
+    }
+  }
+  return { problema: s, solucao: '' }
+}
+
+/** Extrai lista "bruta" de todas as <li> dentro do HTML (fallback). */
+function extractAllLiItems(root: HTMLElement): string[] {
+  return root
+    .querySelectorAll('li')
+    .map((li) => textOf(li))
+    .filter(Boolean)
+}
+
+export function parseWorklogHtml(
+  html: string,
+  meta: { id: number; discussion_id: number; created_at: string; staff_id?: number | null },
+): ParsedWorklogComment {
+  const root = parse(html || '')
+  const headings = root.querySelectorAll('h1, h2, h3, h4, h5, h6')
+
+  // Título: primeiro heading não vazio
+  let title: string | null = null
+  for (const h of headings) {
+    const t = textOf(h)
+    if (t) { title = t; break }
+  }
+
+  // Data: preferir `meta.created_at` se válido; senão extrair do título ou do texto
+  let createdAt = meta.created_at
+  if (!createdAt || createdAt === '1970-01-01T00:00:00.000Z' || createdAt.startsWith('1970')) {
+    const fromTitle = title ? parseDateFromTitle(title) : null
+    if (fromTitle) createdAt = fromTitle
+    else {
+      const fromText = parseDateFromTitle(textOf(root).slice(0, 500))
+      createdAt = fromText ?? new Date().toISOString()
+    }
+  }
+
+  const fullText = textOf(root)
+  const taskRef = parseTaskRef(fullText)
+  const durationSec = parseDuration(fullText)
+
+  // Indexa secções por chave normalizada
+  const sections = new Map<string, HTMLElement>()
+  for (const h of headings) {
+    const key = sectionKey(textOf(h))
+    if (!sections.has(key)) sections.set(key, h)
+  }
+
+  function findSection(target: Set<string>): HTMLElement | null {
+    for (const [k, el] of sections) {
+      if (target.has(k)) return el
+    }
+    // match parcial (ex.: "trabalho realizado manutenção" — começa com)
+    for (const [k, el] of sections) {
+      for (const t of target) {
+        if (k.startsWith(t) || t.startsWith(k)) return el
+      }
+    }
+    return null
+  }
+
+  const workHeading = findSection(SECTION_WORK)
+  const filesHeading = findSection(SECTION_FILES)
+  const problemsHeading = findSection(SECTION_PROBLEMS)
+  const patternsHeading = findSection(SECTION_PATTERNS)
+  const actionsHeading = findSection(SECTION_ACTIONS)
+
+  const workItems = workHeading ? collectFollowingListItems(workHeading) : []
+  const filesModified = filesHeading ? collectFollowingListItems(filesHeading) : []
+  const problemsRaw = problemsHeading ? collectFollowingListItems(problemsHeading) : []
+  const patternsText = patternsHeading ? collectFollowingListItems(patternsHeading) : []
+  const actionsRaw = actionsHeading ? collectFollowingListItems(actionsHeading) : []
+
+  // Fallback: se nenhuma secção encontrada mas existem <li>, e a discussão é #33,
+  // tratar tudo como acções (formato diferente das outras discussões)
+  let actions = actionsRaw.map(parseActionItem)
+  if (meta.discussion_id === 33 && actions.length === 0) {
+    actions = extractAllLiItems(root).map(parseActionItem)
+  }
+
+  const problems = problemsRaw.map(parseProblemItem)
+
+  return {
+    id: meta.id,
+    discussion_id: meta.discussion_id,
+    created_at: createdAt,
+    staff_id: meta.staff_id ?? null,
+    title,
+    task_ref: taskRef,
+    duration_sec: durationSec,
+    work_items: workItems,
+    files_modified: filesModified,
+    problems,
+    patterns_text: patternsText,
+    actions,
+    raw_html: html,
+  }
+}
+
+/** Converte o campo `created` devolvido pelo Desk MCP (pode ser objecto vazio). */
+function normalizeMcpDate(v: unknown): string {
+  if (!v) return ''
+  if (typeof v === 'string') return v
+  if (typeof v === 'object') {
+    const obj = v as Record<string, unknown>
+    if (typeof obj.date === 'string') return obj.date
+    if (typeof obj.datetime === 'string') return obj.datetime
+  }
+  return ''
+}
+
+/** Achata a árvore de comentários (comentários com children recursivos). */
+function flattenComments(comments: RawComment[]): RawComment[] {
+  const out: RawComment[] = []
+  for (const c of comments) {
+    out.push(c)
+    if (c.children && c.children.length) {
+      out.push(...flattenComments(c.children))
+    }
+  }
+  return out
+}
+
+export interface ImportResult {
+  discussion_id: number
+  fetched: number
+  imported: number
+  updated: number
+  skipped: number
+  errors: number
+}
+
+/**
+ * Importa todos os comentários de uma discussão Desk. Paginação por `limit`/`offset`.
+ * Idempotente por `id` — comentários já existentes sofrem update (raw_html pode mudar).
+ */
+export async function importWorklogDiscussion(
+  db: SessionsDb,
+  discussionId: number,
+  opts: { sinceIso?: string; pageSize?: number; maxPages?: number } = {},
+): Promise<ImportResult> {
+  // O MCP desk-crm parece clampar resultados em 200/página independentemente do limit.
+  // Pedimos 200 e iteramos offset até a resposta vir vazia.
+  const pageSize = opts.pageSize ?? 200
+  const maxPages = opts.maxPages ?? 20
+  const result: ImportResult = {
+    discussion_id: discussionId,
+    fetched: 0,
+    imported: 0,
+    updated: 0,
+    skipped: 0,
+    errors: 0,
+  }
+
+  let offset = 0
+  for (let page = 0; page < maxPages; page++) {
+    const raw = await callMcpTool('get_discussion_comments', {
+      discussion_id: discussionId,
+      limit: pageSize,
+      offset,
+    })
+    const payload = extractMcpJsonPayload<{
+      success?: boolean
+      comments?: RawComment[]
+    }>(raw)
+    const pageComments = flattenComments(payload.comments ?? [])
+    if (pageComments.length === 0) break
+    result.fetched += pageComments.length
+
+    const importedAt = new Date().toISOString()
+    for (const c of pageComments) {
+      try {
+        const createdStr = normalizeMcpDate(c.created)
+        const parsed = parseWorklogHtml(c.content ?? '', {
+          id: c.id,
+          discussion_id: c.discussion_id ?? discussionId,
+          created_at: createdStr || '',
+          staff_id: c.staff_id,
+        })
+        if (opts.sinceIso && parsed.created_at < opts.sinceIso) {
+          result.skipped++
+          continue
+        }
+        const record: WorklogCommentRecord = {
+          ...parsed,
+          imported_at: importedAt,
+        }
+        const { inserted } = db.upsertWorklogComment(record)
+        if (inserted) result.imported++
+        else result.updated++
+      } catch (e) {
+        console.error(`[worklog-import] erro a parsear comentário #${c.id}:`, (e as Error).message)
+        result.errors++
+      }
+    }
+
+    // Avança offset; quando próxima página vier vazia, o while quebra na próxima iter.
+    offset += pageComments.length
+    // Safety: se MCP devolveu 0, para
+    if (pageComments.length === 0) break
+  }
+  return result
+}

api/tests/sessions-worklog-import.test.ts

@@ -0,0 +1,175 @@
+import { describe, it, expect, beforeEach } from 'vitest'
+import { mkdtempSync } from 'fs'
+import { tmpdir } from 'os'
+import { join } from 'path'
+import { openSessionsDb, type SessionsDb, type WorklogCommentRecord } from '../services/sessions/db.js'
+import { parseWorklogHtml } from '../services/sessions/worklog-import.js'
+import { detectActionsNeverExecuted, weekRange } from '../services/sessions/patterns.js'
+
+const SAMPLE_H4 = `
+<h4>2026-04-15 10:30 - Refactor API sessions</h4>
+<p><strong>Projecto:</strong> DashDescomplicar</p>
+<p><strong>Tarefa:</strong> #2059 - Observabilidade Espelho</p>
+<p><strong>Duração:</strong> ~2h 15m</p>
+<h4>Trabalho Realizado</h4>
+<ul><li>Criar módulo worklog-import</li><li>Integrar detectores cruzados</li></ul>
+<h4>Ficheiros Modificados</h4>
+<ul><li><code>api/services/sessions/db.ts</code></li><li><code>api/scripts/sessions-worklog-import.ts</code></li></ul>
+<h4>Problemas / Soluções</h4>
+<ul><li>Parser HTML frágil → usar node-html-parser</li></ul>
+<h4>Padrões Detectados</h4>
+<ul><li>MCP gateway responde em SSE ou JSON</li></ul>
+<h4>Acções Sugeridas</h4>
+<ul><li>[Refactor] Extrair callMcpTool para módulo partilhado P2</li></ul>
+`
+
+const SAMPLE_H2 = `
+<h2>2026-01-31 - Estratégia Stack</h2>
+<p><strong>Duração:</strong> ~2h</p>
+<h3>Trabalho Realizado</h3>
+<ul><li>Stack Mapeado - 15 sistemas</li></ul>
+<h3>Insights</h3>
+<ul><li>Posicionamento: Marketing alta performance</li></ul>
+`
+
+const SAMPLE_D33 = `
+<ul>
+  <li>[ ] [MCP] Corrigir bug desk-crm-v3 com tabelas de discussões</li>
+</ul>
+<p><strong>Origem:</strong> Sessão 2026-02-02</p>
+<p><strong>Prioridade:</strong> P1</p>
+`
+
+describe('parseWorklogHtml', () => {
+  it('extrai campos de comentário formato <h4>', () => {
+    const parsed = parseWorklogHtml(SAMPLE_H4, { id: 100, discussion_id: 31, created_at: '' })
+    expect(parsed.id).toBe(100)
+    expect(parsed.title).toMatch(/2026-04-15/)
+    expect(parsed.task_ref).toBe('#2059')
+    expect(parsed.duration_sec).toBe(2 * 3600 + 15 * 60)
+    expect(parsed.work_items.length).toBe(2)
+    expect(parsed.files_modified.length).toBe(2)
+    expect(parsed.patterns_text.length).toBe(1)
+    expect(parsed.actions.length).toBe(1)
+    expect(parsed.actions[0].tipo).toBe('Refactor')
+    expect(parsed.actions[0].prioridade).toBe('P2')
+    expect(parsed.created_at.startsWith('2026-04-15')).toBe(true)
+  })
+
+  it('extrai campos de comentário formato <h2>/<h3> (legacy)', () => {
+    const parsed = parseWorklogHtml(SAMPLE_H2, { id: 64, discussion_id: 31, created_at: '' })
+    expect(parsed.title).toMatch(/2026-01-31/)
+    expect(parsed.work_items.length).toBeGreaterThanOrEqual(1)
+    expect(parsed.duration_sec).toBe(2 * 3600)
+    expect(parsed.created_at.startsWith('2026-01-31')).toBe(true)
+  })
+
+  it('extrai acções em formato discussão #33 (lista crua)', () => {
+    const parsed = parseWorklogHtml(SAMPLE_D33, { id: 200, discussion_id: 33, created_at: '2026-02-02T00:00:00Z' })
+    expect(parsed.actions.length).toBe(1)
+    expect(parsed.actions[0].tipo).toBe('MCP')
+  })
+})
+
+describe('upsertWorklogComment idempotência', () => {
+  let db: SessionsDb
+  beforeEach(() => {
+    const dir = mkdtempSync(join(tmpdir(), 'obs-wl-'))
+    db = openSessionsDb(join(dir, 'sessions.db'))
+  })
+
+  it('insert primeiro, update depois', () => {
+    const base: WorklogCommentRecord = {
+      id: 42,
+      discussion_id: 31,
+      created_at: '2026-04-15T10:30:00Z',
+      staff_id: 25,
+      title: 'Test',
+      task_ref: '#100',
+      duration_sec: 600,
+      work_items: ['a'],
+      files_modified: [],
+      problems: [],
+      patterns_text: [],
+      actions: [],
+      raw_html: '<h4>Test</h4>',
+      imported_at: '2026-04-23T00:00:00Z',
+    }
+    const r1 = db.upsertWorklogComment(base)
+    expect(r1.inserted).toBe(true)
+    expect(db.countWorklogComments()).toBe(1)
+    const r2 = db.upsertWorklogComment({ ...base, title: 'Updated' })
+    expect(r2.inserted).toBe(false)
+    expect(db.countWorklogComments()).toBe(1)
+    const list = db.listWorklogComments({ discussion_id: 31 })
+    expect(list[0].title).toBe('Updated')
+  })
+})
+
+describe('detectActionsNeverExecuted', () => {
+  let db: SessionsDb
+  beforeEach(() => {
+    const dir = mkdtempSync(join(tmpdir(), 'obs-act-'))
+    db = openSessionsDb(join(dir, 'sessions.db'))
+  })
+
+  it('sinaliza acções P1/P2 antigas sem execução', () => {
+    const old = new Date('2026-03-01T00:00:00Z').toISOString()
+    for (let i = 0; i < 4; i++) {
+      db.upsertWorklogComment({
+        id: 300 + i,
+        discussion_id: 33,
+        created_at: old,
+        staff_id: 25,
+        title: `Acção ${i}`,
+        task_ref: `#${1000 + i}`,
+        duration_sec: null,
+        work_items: [],
+        files_modified: [],
+        problems: [],
+        patterns_text: [],
+        actions: [{ tipo: 'MCP', descricao: `Corrigir bug X${i}`, prioridade: i % 2 ? 'P1' : 'P2' }],
+        raw_html: '',
+        imported_at: '2026-04-23T00:00:00Z',
+      })
+    }
+    const range = weekRange(new Date('2026-04-22T00:00:00Z'))
+    const patterns = detectActionsNeverExecuted({
+      db: db.rawDb(),
+      weekStartIso: range.start.toISOString(),
+      weekEndIso: range.end.toISOString(),
+    })
+    expect(patterns.length).toBe(1)
+    expect(patterns[0].pattern_key).toBe('actions_never_executed')
+    expect(patterns[0].affected_count).toBeGreaterThanOrEqual(3)
+  })
+
+  it('não sinaliza se acções recentes (<14 dias)', () => {
+    const recent = new Date().toISOString()
+    for (let i = 0; i < 5; i++) {
+      db.upsertWorklogComment({
+        id: 400 + i,
+        discussion_id: 33,
+        created_at: recent,
+        staff_id: 25,
+        title: null,
+        task_ref: null,
+        duration_sec: null,
+        work_items: [],
+        files_modified: [],
+        problems: [],
+        patterns_text: [],
+        actions: [{ tipo: 'MCP', descricao: 'x', prioridade: 'P1' }],
+        raw_html: '',
+        imported_at: recent,
+      })
+    }
+    const range = weekRange(new Date())
+    const patterns = detectActionsNeverExecuted({
+      db: db.rawDb(),
+      weekStartIso: range.start.toISOString(),
+      weekEndIso: range.end.toISOString(),
+    })
+    expect(patterns.length).toBe(0)
+  })
+})

package-lock.json

@@ -21,6 +21,7 @@
         "googleapis": "^171.4.0",
         "lucide-react": "^0.563.0",
         "mysql2": "^3.11.5",
+        "node-html-parser": "^7.1.0",
         "oidc-client-ts": "^3.0.1",
         "pg": "^8.20.0",
         "react": "^19.2.0",
@@ -3414,6 +3415,12 @@
       "integrity": "sha512-Tpp60P6IUJDTuOq/5Z8cdskzJujfwqfOTkrwIwj7IRISpnkJnT6SyJ4PCPnGMoFjC9ddhal5KVIYtAt97ix05A==",
       "license": "MIT"
     },
+    "node_modules/boolbase": {
+      "version": "1.0.0",
+      "resolved": "https://registry.npmjs.org/boolbase/-/boolbase-1.0.0.tgz",
+      "integrity": "sha512-JZOSA7Mo9sNGB8+UjSgzdLtokWAky1zbztM3WRLCbZ70/3cTANmQmOdR7y2g+J0e2WXywy1yS468tY+IruqEww==",
+      "license": "ISC"
+    },
     "node_modules/brace-expansion": {
       "version": "1.1.13",
       "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-1.1.13.tgz",
@@ -3905,6 +3912,22 @@
         "node": ">= 8"
       }
     },
+    "node_modules/css-select": {
+      "version": "5.2.2",
+      "resolved": "https://registry.npmjs.org/css-select/-/css-select-5.2.2.tgz",
+      "integrity": "sha512-TizTzUddG/xYLA3NXodFM0fSbNizXjOKhqiQQwvhlspadZokn1KDy0NZFS0wuEubIYAV5/c1/lAr0TaaFXEXzw==",
+      "license": "BSD-2-Clause",
+      "dependencies": {
+        "boolbase": "^1.0.0",
+        "css-what": "^6.1.0",
+        "domhandler": "^5.0.2",
+        "domutils": "^3.0.1",
+        "nth-check": "^2.0.1"
+      },
+      "funding": {
+        "url": "https://github.com/sponsors/fb55"
+      }
+    },
     "node_modules/css-tree": {
       "version": "3.1.0",
       "resolved": "https://registry.npmjs.org/css-tree/-/css-tree-3.1.0.tgz",
@@ -3919,6 +3942,18 @@
         "node": "^10 || ^12.20.0 || ^14.13.0 || >=15.0.0"
       }
     },
+    "node_modules/css-what": {
+      "version": "6.2.2",
+      "resolved": "https://registry.npmjs.org/css-what/-/css-what-6.2.2.tgz",
+      "integrity": "sha512-u/O3vwbptzhMs3L1fQE82ZSLHQQfto5gyZzwteVIEyeaY5Fc7R4dapF/BvRoSYFeqfBk4m0V1Vafq5Pjv25wvA==",
+      "license": "BSD-2-Clause",
+      "engines": {
+        "node": ">= 6"
+      },
+      "funding": {
+        "url": "https://github.com/sponsors/fb55"
+      }
+    },
     "node_modules/css.escape": {
       "version": "1.5.1",
       "resolved": "https://registry.npmjs.org/css.escape/-/css.escape-1.5.1.tgz",
@@ -4384,6 +4419,73 @@
       "license": "MIT",
       "peer": true
     },
+    "node_modules/dom-serializer": {
+      "version": "2.0.0",
+      "resolved": "https://registry.npmjs.org/dom-serializer/-/dom-serializer-2.0.0.tgz",
+      "integrity": "sha512-wIkAryiqt/nV5EQKqQpo3SToSOV9J0DnbJqwK7Wv/Trc92zIAYZ4FlMu+JPFW1DfGFt81ZTCGgDEabffXeLyJg==",
+      "license": "MIT",
+      "dependencies": {
+        "domelementtype": "^2.3.0",
+        "domhandler": "^5.0.2",
+        "entities": "^4.2.0"
+      },
+      "funding": {
+        "url": "https://github.com/cheeriojs/dom-serializer?sponsor=1"
+      }
+    },
+    "node_modules/dom-serializer/node_modules/entities": {
+      "version": "4.5.0",
+      "resolved": "https://registry.npmjs.org/entities/-/entities-4.5.0.tgz",
+      "integrity": "sha512-V0hjH4dGPh9Ao5p0MoRY6BVqtwCjhz6vI5LT8AJ55H+4g9/4vbHx1I54fS0XuclLhDHArPQCiMjDxjaL8fPxhw==",
+      "license": "BSD-2-Clause",
+      "engines": {
+        "node": ">=0.12"
+      },
+      "funding": {
+        "url": "https://github.com/fb55/entities?sponsor=1"
+      }
+    },
+    "node_modules/domelementtype": {
+      "version": "2.3.0",
+      "resolved": "https://registry.npmjs.org/domelementtype/-/domelementtype-2.3.0.tgz",
+      "integrity": "sha512-OLETBj6w0OsagBwdXnPdN0cnMfF9opN69co+7ZrbfPGrdpPVNBUj02spi6B1N7wChLQiPn4CSH/zJvXw56gmHw==",
+      "funding": [
+        {
+          "type": "github",
+          "url": "https://github.com/sponsors/fb55"
+        }
+      ],
+      "license": "BSD-2-Clause"
+    },
+    "node_modules/domhandler": {
+      "version": "5.0.3",
+      "resolved": "https://registry.npmjs.org/domhandler/-/domhandler-5.0.3.tgz",
+      "integrity": "sha512-cgwlv/1iFQiFnU96XXgROh8xTeetsnJiDsTc7TYCLFd9+/WNkIqPTxiM/8pSd8VIrhXGTf1Ny1q1hquVqDJB5w==",
+      "license": "BSD-2-Clause",
+      "dependencies": {
+        "domelementtype": "^2.3.0"
+      },
+      "engines": {
+        "node": ">= 4"
+      },
+      "funding": {
+        "url": "https://github.com/fb55/domhandler?sponsor=1"
+      }
+    },
+    "node_modules/domutils": {
+      "version": "3.2.2",
+      "resolved": "https://registry.npmjs.org/domutils/-/domutils-3.2.2.tgz",
+      "integrity": "sha512-6kZKyUajlDuqlHKVX1w7gyslj9MPIXzIFiz/rGu35uC1wMi+kMhQwGhl4lt9unC9Vb9INnY9Z3/ZA3+FhASLaw==",
+      "license": "BSD-2-Clause",
+      "dependencies": {
+        "dom-serializer": "^2.0.0",
+        "domelementtype": "^2.3.0",
+        "domhandler": "^5.0.3"
+      },
+      "funding": {
+        "url": "https://github.com/fb55/domutils?sponsor=1"
+      }
+    },
     "node_modules/dotenv": {
       "version": "16.6.1",
       "resolved": "https://registry.npmjs.org/dotenv/-/dotenv-16.6.1.tgz",
@@ -5837,6 +5939,15 @@
         "node": ">= 0.4"
       }
     },
+    "node_modules/he": {
+      "version": "1.2.0",
+      "resolved": "https://registry.npmjs.org/he/-/he-1.2.0.tgz",
+      "integrity": "sha512-F/1DnUGPopORZi0ni+CvrCgHQ5FyEAHRLSApuYWMmrbSwoN2Mn/7k+Gl38gJnR7yyDZk6WLXwiGod1JOWNDKGw==",
+      "license": "MIT",
+      "bin": {
+        "he": "bin/he"
+      }
+    },
     "node_modules/hermes-estree": {
       "version": "0.25.1",
       "resolved": "https://registry.npmjs.org/hermes-estree/-/hermes-estree-0.25.1.tgz",
@@ -7371,6 +7482,16 @@
         "url": "https://opencollective.com/node-fetch"
       }
     },
+    "node_modules/node-html-parser": {
+      "version": "7.1.0",
+      "resolved": "https://registry.npmjs.org/node-html-parser/-/node-html-parser-7.1.0.tgz",
+      "integrity": "sha512-iJo8b2uYGT40Y8BTyy5ufL6IVbN8rbm/1QK2xffXU/1a/v3AAa0d1YAoqBNYqaS4R/HajkWIpIfdE6KcyFh1AQ==",
+      "license": "MIT",
+      "dependencies": {
+        "css-select": "^5.1.0",
+        "he": "1.2.0"
+      }
+    },
     "node_modules/node-releases": {
       "version": "2.0.27",
       "resolved": "https://registry.npmjs.org/node-releases/-/node-releases-2.0.27.tgz",
@@ -7390,6 +7511,18 @@
         "url": "https://github.com/sponsors/sindresorhus"
       }
     },
+    "node_modules/nth-check": {
+      "version": "2.1.1",
+      "resolved": "https://registry.npmjs.org/nth-check/-/nth-check-2.1.1.tgz",
+      "integrity": "sha512-lqjrjmaOoAnWfMmBPL+XNnynZh2+swxiX3WUE0s4yEHI6m+AwrK2UZOimIRl3X/4QctVqS8AiZjFqyOGrMXb/w==",
+      "license": "BSD-2-Clause",
+      "dependencies": {
+        "boolbase": "^1.0.0"
+      },
+      "funding": {
+        "url": "https://github.com/fb55/nth-check?sponsor=1"
+      }
+    },
     "node_modules/object-assign": {
       "version": "4.1.1",
       "resolved": "https://registry.npmjs.org/object-assign/-/object-assign-4.1.1.tgz",

package.json

@@ -29,6 +29,7 @@
     "googleapis": "^171.4.0",
     "lucide-react": "^0.563.0",
     "mysql2": "^3.11.5",
+    "node-html-parser": "^7.1.0",
     "oidc-client-ts": "^3.0.1",
     "pg": "^8.20.0",
     "react": "^19.2.0",

systemd/observabilidade-worklog-import.service

@@ -0,0 +1,13 @@
+[Unit]
+Description=Observabilidade — import diário de worklogs Desk (#31/#32/#33)
+After=default.target
+
+[Service]
+Type=oneshot
+WorkingDirectory=/media/ealmeida/Dados/Dev/DashDescomplicar
+ExecStart=/home/ealmeida/.nvm/versions/node/v22.22.2/bin/npx tsx api/scripts/sessions-worklog-import.ts --discussion all --since-days 7
+Environment="OBSERVABILIDADE_DB=/home/ealmeida/.claude-work/sessions.db"
+Environment="PATH=/home/ealmeida/.nvm/versions/node/v22.22.2/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin"
+EnvironmentFile=/home/ealmeida/.claude-work/observabilidade-patterns.env
+StandardOutput=append:/home/ealmeida/.claude-work/observabilidade-worklog-import.log
+StandardError=append:/home/ealmeida/.claude-work/observabilidade-worklog-import.log

systemd/observabilidade-worklog-import.timer

@@ -0,0 +1,10 @@
+[Unit]
+Description=Observabilidade — import diário de worklogs Desk
+
+[Timer]
+OnCalendar=daily
+OnCalendar=*-*-* 03:00:00
+Persistent=true
+
+[Install]
+WantedBy=default.target