fix(pipeline): deduplicate opportunities in auto-sync

ROADMAP.md

@@ -689,7 +689,7 @@
 
 - [x] DB: `operations_journal` table (created in Phase 14.1)
 - [x] API: `GET/POST/PUT/DELETE /api/journal`
-- [ ] **⏳ Lógica: Auto-generar entrada diaria desde activities (deferred)** (est. ~3-4h)
+- [x] **Lógica: Auto-generar entrada diaria desde activities** ✅ (commit cbf8a62, 2026-06-13)
 - [x] UI: Página `/journal` con timeline narrativo
 - [x] UI: Entry editor para añadir highlights manuales
 - **Archivos:** `src/app/api/journal/route.ts`, `src/app/api/journal/[id]/route.ts`, `src/app/(dashboard)/journal/page.tsx`, `src/lib/kanban-db.ts`, `src/components/journal/`

scripts/cleanup-duplicate-opportunities.sh

@@ -0,0 +1,118 @@
+#!/usr/bin/env bash
+#
+# Cleanup script: Remove duplicate opportunities from kanban.db
+#
+# Dedup strategy:
+# - Group by LOWER(TRIM(company))
+# - For each group, keep the OLDEST entry (lowest created_at)
+# - Delete all newer duplicates
+#
+# Usage:
+#   ./scripts/cleanup-duplicate-opportunities.sh [--dry-run] [--db /path/to/kanban.db]
+#
+# Options:
+#   --dry-run    Show what would be deleted without deleting
+#   --db PATH    Path to kanban.db (default: data/kanban.db)
+
+set -euo pipefail
+
+# Parse args
+DRY_RUN=false
+DB_PATH="data/kanban.db"
+
+while [[ $# -gt 0 ]]; do
+    case "$1" in
+        --dry-run) DRY_RUN=true; shift ;;
+        --db) DB_PATH="$2"; shift 2 ;;
+        *) echo "Unknown option: $1"; exit 1 ;;
+    esac
+done
+
+if [[ ! -f "$DB_PATH" ]]; then
+    echo "❌ Database not found: $DB_PATH"
+    exit 1
+fi
+
+echo "📦 Database: $DB_PATH"
+if $DRY_RUN; then
+    echo "🔍 DRY RUN — no changes will be made"
+else
+    echo "🗑️  LIVE MODE — duplicates WILL be deleted"
+fi
+echo ""
+
+# Get stats
+TOTAL=$(sqlite3 "$DB_PATH" "SELECT COUNT(*) FROM opportunities;")
+UNIQUE=$(sqlite3 "$DB_PATH" "SELECT COUNT(DISTINCT LOWER(TRIM(company))) FROM opportunities;")
+DUPES=$(sqlite3 "$DB_PATH" "SELECT COUNT(*) - COUNT(DISTINCT LOWER(TRIM(company))) FROM opportunities;")
+
+echo "📊 Total opportunities: $TOTAL"
+echo "📈 Unique companies: $UNIQUE"
+echo "🔁 Duplicates to remove: $DUPES"
+echo ""
+
+if [[ "$DUPES" -eq 0 ]]; then
+    echo "✅ No duplicates found. Database is clean."
+    exit 0
+fi
+
+# Show top offenders
+echo "Top duplicate offenders:"
+sqlite3 -header -column "$DB_PATH" "
+    SELECT company, COUNT(*) as entries, COUNT(*) - 1 as to_remove
+    FROM opportunities
+    GROUP BY LOWER(TRIM(company))
+    HAVING COUNT(*) > 1
+    ORDER BY entries DESC
+    LIMIT 15;
+"
+echo ""
+
+if $DRY_RUN; then
+    echo "🔍 DRY RUN — showing entries that would be deleted (first 30):"
+    sqlite3 -header -column "$DB_PATH" "
+        SELECT o.id, o.company, SUBSTR(o.title, 1, 40) as title, o.stage, o.value, o.created_at
+        FROM opportunities o
+        WHERE o.id NOT IN (
+            SELECT MIN(o2.id)
+            FROM opportunities o2
+            GROUP BY LOWER(TRIM(o2.company))
+        )
+        ORDER BY o.company, o.created_at
+        LIMIT 30;
+    "
+    echo ""
+    echo "✅ DRY RUN complete. No changes made."
+    echo "Run without --dry-run to actually delete duplicates."
+else
+    # Delete duplicates — keep the oldest (MIN(id)) per company
+    DELETED=$(sqlite3 "$DB_PATH" "
+        DELETE FROM opportunities
+        WHERE id NOT IN (
+            SELECT MIN(o.id)
+            FROM opportunities o
+            GROUP BY LOWER(TRIM(o.company))
+        );
+        SELECT changes();
+    ")
+    echo "✅ Deleted $DELETED duplicate opportunities"
+
+    # Verify
+    REMAINING=$(sqlite3 "$DB_PATH" "SELECT COUNT(*) FROM opportunities;")
+    echo "📊 Remaining opportunities: $REMAINING"
+
+    # Check for remaining duplicates
+    STILL_DUPES=$(sqlite3 "$DB_PATH" "
+        SELECT COUNT(*) FROM (
+            SELECT LOWER(TRIM(company)) as key
+            FROM opportunities
+            GROUP BY key
+            HAVING COUNT(*) > 1
+        );
+    ")
+    if [[ "$STILL_DUPES" -eq 0 ]]; then
+        echo "✅ No more duplicates by company."
+    else
+        echo "⚠️  Still have $STILL_DUPES company groups with duplicates"
+    fi
+fi

src/app/(dashboard)/journal/JournalClient.tsx

@@ -1,7 +1,7 @@
 "use client";
 
 import { useState, useEffect } from "react";
-import { Plus } from "lucide-react";
+import { Plus, Sparkles } from "lucide-react";
 import {
   JournalTimeline,
   JournalEntryCard,
@@ -23,6 +23,7 @@ export default function JournalClient({ initialData }: { initialData?: JournalIn
   const [isModalOpen, setIsModalOpen] = useState(false);
   const [startDate, setStartDate] = useState("");
   const [endDate, setEndDate] = useState("");
+  const [generating, setGenerating] = useState(false);
 
   useEffect(() => {
     if (!initialData?.entries) {
@@ -102,6 +103,28 @@ export default function JournalClient({ initialData }: { initialData?: JournalIn
     setIsModalOpen(true);
   };
 
+  const handleAutoGenerate = async () => {
+    setGenerating(true);
+    try {
+      const today = new Date().toISOString().split("T")[0];
+      const res = await authFetch("/api/journal/auto-generate", {
+        method: "POST",
+        headers: { "Content-Type": "application/json" },
+        body: JSON.stringify({ date: today }),
+      });
+      const data = await res.json();
+      if (data.generated && data.entry) {
+        setEntries([data.entry, ...entries]);
+      } else if (data.entry) {
+        setEntries([data.entry, ...entries.filter((e) => e.id !== data.entry.id)]);
+      }
+    } catch {
+      // ignore
+    } finally {
+      setGenerating(false);
+    }
+  };
+
   const handleCloseModal = () => {
     setIsModalOpen(false);
     setEditingEntry(null);
@@ -128,14 +151,25 @@ export default function JournalClient({ initialData }: { initialData?: JournalIn
             Registra y revisa tus actividades diarias
           </p>
         </div>
-        <button
-          onClick={handleNewEntry}
-          className="flex items-center gap-1.5 px-4 py-2 rounded-lg text-sm font-medium transition-colors hover:opacity-80"
-          style={{ backgroundColor: "var(--accent)", color: "var(--text-primary)" }}
-        >
-          <Plus className="w-4 h-4" />
-          Nueva Entrada
-        </button>
+        <div className="flex items-center gap-2">
+          <button
+            onClick={handleAutoGenerate}
+            disabled={generating}
+            className="flex items-center gap-1.5 px-4 py-2 rounded-lg text-sm font-medium transition-colors hover:opacity-80 disabled:opacity-50"
+            style={{ backgroundColor: "var(--card-elevated)", border: "1px solid var(--border)", color: "var(--text-secondary)" }}
+          >
+            <Sparkles className="w-4 h-4" />
+            {generating ? "Generando..." : "Auto-Generar Hoy"}
+          </button>
+          <button
+            onClick={handleNewEntry}
+            className="flex items-center gap-1.5 px-4 py-2 rounded-lg text-sm font-medium transition-colors hover:opacity-80"
+            style={{ backgroundColor: "var(--accent)", color: "var(--text-primary)" }}
+          >
+            <Plus className="w-4 h-4" />
+            Nueva Entrada
+          </button>
+        </div>
       </div>
 
       {/* Filters */}

src/app/api/journal/auto-generate/route.ts

@@ -0,0 +1,136 @@
+import { NextRequest, NextResponse } from "next/server";
+import { getActivities } from "@/lib/activities-db";
+import { createJournalEntry, listJournalEntries } from "@/lib/kanban-db";
+
+export const dynamic = "force-dynamic";
+
+/**
+ * POST /api/journal/auto-generate
+ * Auto-generates a journal entry from the day's activities.
+ * Body: { date: string (YYYY-MM-DD) }
+ * 
+ * Aggregates activities for the given date and creates a narrative
+ * summary with highlights. Skips if entry already exists.
+ */
+export async function POST(request: NextRequest) {
+  try {
+    const body = await request.json();
+    const date = body.date;
+
+    if (!date || !/^\d{4}-\d{2}-\d{2}$/.test(date)) {
+      return NextResponse.json(
+        { error: "date is required (YYYY-MM-DD format)" },
+        { status: 400 }
+      );
+    }
+
+    // Check if entry already exists for this date
+    const existing = listJournalEntries({ startDate: date, endDate: date });
+    if (existing.length > 0) {
+      return NextResponse.json(
+        { message: "Journal entry already exists for this date", entry: existing[0] },
+        { status: 200 }
+      );
+    }
+
+    // Fetch activities for the date
+    const result = getActivities({
+      startDate: date,
+      endDate: date,
+      limit: 500,
+      sort: "oldest",
+    });
+
+    const activities = result.activities;
+    const total = result.total;
+
+    if (total === 0) {
+      return NextResponse.json(
+        { message: "No activities found for this date", entry: null },
+        { status: 200 }
+      );
+    }
+
+    // Build narrative from activities
+    const byType: Record<string, number> = {};
+    const byStatus: Record<string, number> = {};
+    const byAgent: Record<string, number> = {};
+    const errors: string[] = [];
+    const notableActions: string[] = [];
+
+    for (const act of activities) {
+      byType[act.type] = (byType[act.type] || 0) + 1;
+      byStatus[act.status] = (byStatus[act.status] || 0) + 1;
+      if (act.agent) {
+        byAgent[act.agent] = (byAgent[act.agent] || 0) + 1;
+      }
+      if (act.status === "error") {
+        errors.push(`[${act.type}] ${act.description}`);
+      }
+      // Notable: security, build, or high-duration actions
+      if (act.type === "security" || act.type === "build" || (act.duration_ms && act.duration_ms > 60000)) {
+        notableActions.push(`${act.type}: ${act.description}`);
+      }
+    }
+
+    const successCount = (byStatus["success"] || 0) + (byStatus["approved"] || 0);
+    const errorCount = (byStatus["error"] || 0) + (byStatus["rejected"] || 0);
+    const successRate = total > 0 ? Math.round((successCount / total) * 100) : 100;
+    const agents = Object.entries(byAgent).sort((a, b) => b[1] - a[1]);
+    const types = Object.entries(byType).sort((a, b) => b[1] - a[1]);
+
+    // Compose narrative
+    const lines: string[] = [];
+    lines.push(`Resumen de operaciones del ${date}: ${total} actividades registradas con ${successRate}% de tasa de éxito.`);
+
+    if (agents.length > 0) {
+      const agentSummary = agents.map(([name, count]) => `${name} (${count})`).join(", ");
+      lines.push(`Agentes activos: ${agentSummary}.`);
+    }
+
+    if (types.length > 0) {
+      const typeSummary = types.slice(0, 5).map(([t, c]) => `${t}: ${c}`).join(", ");
+      lines.push(`Distribución: ${typeSummary}.`);
+    }
+
+    if (errorCount > 0) {
+      lines.push(`⚠️ ${errorCount} errores detectados.`);
+    }
+
+    if (notableActions.length > 0) {
+      lines.push(`Acciones destacadas: ${notableActions.slice(0, 5).join("; ")}.`);
+    }
+
+    // Total tokens if available
+    const totalTokens = activities.reduce((sum, a) => sum + (a.tokens_used || 0), 0);
+    if (totalTokens > 0) {
+      lines.push(`Tokens consumidos: ${totalTokens.toLocaleString()}.`);
+    }
+
+    const narrative = lines.join(" ");
+
+    // Build highlights
+    const highlights: string[] = [];
+    if (successRate === 100) highlights.push("✅ 100% tasa de éxito");
+    if (errorCount > 0) highlights.push(`❌ ${errorCount} errores`);
+    if (total > 50) highlights.push(`🔥 ${total} actividades (día intenso)`);
+    if (agents.length > 1) highlights.push(`🤖 ${agents.length} agentes activos`);
+    if (totalTokens > 10000) highlights.push(`📊 ${Math.round(totalTokens / 1000)}K tokens`);
+    highlights.push(`📈 ${successRate}% éxito`);
+
+    // Create the entry
+    const entry = createJournalEntry({
+      date,
+      narrative,
+      highlights: highlights.slice(0, 10),
+    });
+
+    return NextResponse.json({ entry, generated: true }, { status: 201 });
+  } catch (error) {
+    console.error("Failed to auto-generate journal entry:", error);
+    return NextResponse.json(
+      { error: "Failed to auto-generate journal entry" },
+      { status: 500 }
+    );
+  }
+}

src/app/api/pipeline/auto-sync/route.ts

@@ -7,6 +7,7 @@ import {
   getOpportunity,
   updateOpportunity,
   getPipelineKPIs,
+  findOpportunityByCompany,
   type CreateOpportunityInput,
 } from "@/lib/pipeline-db";
 import {
@@ -139,10 +140,17 @@ function syncReportsToPipeline(): SyncResult {
           continue;
         }
 
-        const existingOpp = getOpportunity(parsed.reportId);
+        // Dedup: find existing by company (same company = same deal)
+        // Previously used getOpportunity(parsed.reportId) which NEVER matched
+        // because createOpportunity generates a random UUID, not the reportId.
+        const existingOpp = findOpportunityByCompany(parsed.target);
 
         if (existingOpp) {
-          updateOpportunity(parsed.reportId, {
+          // UPDATE: keep max value, append notes
+          const newValue = getEstimatedValue(parsed.serviceType);
+          const maxValue = Math.max(existingOpp.value, newValue);
+          updateOpportunity(existingOpp.id, {
+            value: maxValue,
             notes: [
               `Updated from report: ${file}`,
               `Confidence: ${parsed.confidence}%`,