feat(whatsapp): QWEN primary AI backend, Hermes fallback, conversation history, vehicle persistence, demo prompts

- Add QWEN (qwen3.6) as primary AI backend with short system prompt
- Hermes remains as fallback with 45s timeout
- Increase QWEN timeout to 35s, max_tokens to 4000
- Add conversation history loading from whatsapp_messages (last 4 msgs)
- Persist detected vehicle in whatsapp_sessions table
- Add 'limpiar chat' / 'nuevo chat' / 'reset' commands to clear history
- Fix CSS conflict: rename whatsapp chat-panel classes to wa-chat-panel
- Fix JS ID conflicts with chat.js widget (waChatPanel, waChatMessages, etc.)
- Improve no-stock response: conversational with alternatives
- Split search_query by | for multi-part lookups
- Add DEMO_PROMPTS.md and DEMO_PROMPTS_V2.md

pos/services/wa_quotation.py

@@ -109,11 +109,30 @@ def confirm_quotation(tenant_conn, phone):
     return qid
 
 
-# ─── In-memory last-shown-part per phone ─────────────────────────────
+# ─── Persistent last-shown-part per phone ────────────────────────────
 # Tracks what part the bot last showed so "cotizar" knows what to add.
-# Key: phone (clean, no @lid). Value: dict with inventory item info.
+# Stored in tenant DB table whatsapp_sessions so it survives restarts.
 
-_last_shown = {}
+_WHATSAPP_SESSIONS_SQL = """
+CREATE TABLE IF NOT EXISTS whatsapp_sessions (
+    phone VARCHAR(50) PRIMARY KEY,
+    last_shown JSONB,
+    vehicle JSONB,
+    updated_at TIMESTAMP DEFAULT NOW()
+);
+"""
+
+
+def _ensure_sessions_table(tenant_conn):
+    cur = tenant_conn.cursor()
+    cur.execute(_WHATSAPP_SESSIONS_SQL)
+    # Migrate: add vehicle column if table already existed without it
+    cur.execute("""
+        ALTER TABLE whatsapp_sessions
+        ADD COLUMN IF NOT EXISTS vehicle JSONB
+    """)
+    tenant_conn.commit()
+    cur.close()
 
 
 def set_last_shown_part(phone, part_info):
@@ -122,15 +141,110 @@ def set_last_shown_part(phone, part_info):
     part_info: dict with keys inventory_id, part_number, name, brand,
     price, stock, unit
     """
-    _last_shown[phone] = part_info
+    # In-memory fallback for when tenant_conn is not available
+    from tenant_db import get_tenant_conn
+    try:
+        conn = get_tenant_conn(11)
+        _ensure_sessions_table(conn)
+        cur = conn.cursor()
+        import json
+        cur.execute("""
+            INSERT INTO whatsapp_sessions (phone, last_shown, updated_at)
+            VALUES (%s, %s, NOW())
+            ON CONFLICT (phone) DO UPDATE SET last_shown = EXCLUDED.last_shown, updated_at = NOW()
+        """, (phone, json.dumps(part_info)))
+        conn.commit()
+        cur.close()
+        conn.close()
+    except Exception as e:
+        print(f"[WA-SESSION] Failed to persist last_shown for {phone}: {e}")
 
 
 def get_last_shown_part(phone):
-    return _last_shown.get(phone)
+    from tenant_db import get_tenant_conn
+    try:
+        conn = get_tenant_conn(11)
+        _ensure_sessions_table(conn)
+        cur = conn.cursor()
+        cur.execute("SELECT last_shown FROM whatsapp_sessions WHERE phone = %s", (phone,))
+        row = cur.fetchone()
+        cur.close()
+        conn.close()
+        if row and row[0]:
+            return row[0]
+    except Exception as e:
+        print(f"[WA-SESSION] Failed to read last_shown for {phone}: {e}")
+    return None
 
 
 def clear_last_shown(phone):
-    _last_shown.pop(phone, None)
+    from tenant_db import get_tenant_conn
+    try:
+        conn = get_tenant_conn(11)
+        _ensure_sessions_table(conn)
+        cur = conn.cursor()
+        cur.execute("DELETE FROM whatsapp_sessions WHERE phone = %s", (phone,))
+        conn.commit()
+        cur.close()
+        conn.close()
+    except Exception as e:
+        print(f"[WA-SESSION] Failed to clear last_shown for {phone}: {e}")
+
+
+def set_vehicle(phone, vehicle):
+    """Store the detected vehicle for this phone number.
+
+    vehicle: dict with keys brand, model, year
+    """
+    from tenant_db import get_tenant_conn
+    try:
+        conn = get_tenant_conn(11)
+        _ensure_sessions_table(conn)
+        cur = conn.cursor()
+        import json
+        cur.execute("""
+            INSERT INTO whatsapp_sessions (phone, vehicle, updated_at)
+            VALUES (%s, %s, NOW())
+            ON CONFLICT (phone) DO UPDATE SET vehicle = EXCLUDED.vehicle, updated_at = NOW()
+        """, (phone, json.dumps(vehicle)))
+        conn.commit()
+        cur.close()
+        conn.close()
+    except Exception as e:
+        print(f"[WA-SESSION] Failed to persist vehicle for {phone}: {e}")
+
+
+def get_vehicle(phone):
+    """Retrieve the stored vehicle for this phone number."""
+    from tenant_db import get_tenant_conn
+    try:
+        conn = get_tenant_conn(11)
+        _ensure_sessions_table(conn)
+        cur = conn.cursor()
+        cur.execute("SELECT vehicle FROM whatsapp_sessions WHERE phone = %s", (phone,))
+        row = cur.fetchone()
+        cur.close()
+        conn.close()
+        if row and row[0]:
+            return row[0]
+    except Exception as e:
+        print(f"[WA-SESSION] Failed to read vehicle for {phone}: {e}")
+    return None
+
+
+def clear_session(phone):
+    """Clear all session data (last_shown + vehicle) for this phone."""
+    from tenant_db import get_tenant_conn
+    try:
+        conn = get_tenant_conn(11)
+        _ensure_sessions_table(conn)
+        cur = conn.cursor()
+        cur.execute("DELETE FROM whatsapp_sessions WHERE phone = %s", (phone,))
+        conn.commit()
+        cur.close()
+        conn.close()
+    except Exception as e:
+        print(f"[WA-SESSION] Failed to clear session for {phone}: {e}")
 
 
 # ─── Quotation CRUD ─────────────────────────────────────────────────