bilal/DBBot-telegram
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Initial project setup for Telegram-to-Qdrant knowledge base.

Browse Source 
Add unified n8n workflow, userbot ingestion script, and detailed README for running n8n with a separate Dockerized Telethon collector.

Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
bilal
2026-06-23 20:20:56 +03:00
commit a8d634f422
13 changed files with 1768 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
.env.example Normal file
View File

@@ -0,0 +1,6 @@
TG_API_ID=12345678
TG_API_HASH=your_telegram_api_hash
TG_GROUP_ID=-1001234567890
N8N_WEBHOOK_URL=http://n8n:5678/webhook/tg-inbox
SESSION_FILE_PATH=/app/session/session.session
HISTORY_CHECK_FILE=/app/session/history_done.flag

8
.gitignore vendored Normal file
View File

@@ -0,0 +1,8 @@
.DS_Store
__pycache__/
*.pyc
.env
session_data/session.session
session_data/history_done.flag
n8n_data/
qdrant_data/

403
DBBot Unified n8n Qdrant.json Normal file
View File

@@ -0,0 +1,403 @@
{
"name": "DBBot Unified (Ingest + RAG)",
"nodes": [
{
"parameters": {
"httpMethod": "POST",
"path": "tg-inbox",
"options": {}
},
"type": "n8n-nodes-base.webhook",
"typeVersion": 2.1,
"position": [
-640,
-120
],
"id": "31d39bff-4f9a-4645-94b2-b38d4ebf0bf0",
"name": "Webhook TG Inbox",
"webhookId": "dbbot-tg-inbox-webhook"
},
{
"parameters": {
"jsCode": "const body = $input.item.json.body || {};\nconst meta = body.metadata || {};\nconst rawText = (body.text || '').trim();\n\nif (!rawText) {\n return { json: { skip: true, reason: 'empty_text' } };\n}\n\nconst authorLabel = meta.author_label || meta.sender_name || 'Unknown';\nconst authorId = meta.sender_id || meta.author?.id || '0';\nconst repliedTo = meta.replied_to_author_label ? ` (в ответ ${meta.replied_to_author_label})` : '';\n\nreturn {\n json: {\n skip: false,\n text: `Автор ${authorLabel} (ID: ${authorId})${repliedTo}: ${rawText}`,\n metadata: {\n ...meta,\n original_text: rawText,\n source: meta.source || 'telegram_group',\n processed_at: new Date().toISOString()\n }\n }\n};"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
-416,
-120
],
"id": "a9347d4b-fad0-4fb8-a23c-bcad6d2f3186",
"name": "Normalize Telegram Payload"
},
{
"parameters": {
"conditions": {
"boolean": [
{
"value1": "={{ $json.skip }}",
"value2": true
}
]
},
"options": {}
},
"type": "n8n-nodes-base.if",
"typeVersion": 2,
"position": [
-192,
-120
],
"id": "de52e6d1-98be-4c24-bc03-d90d7fa80b93",
"name": "Skip Empty Text"
},
{
"parameters": {
"mode": "insert",
"qdrantCollection": {
"__rl": true,
"value": "telegram_kb",
"mode": "list",
"cachedResultName": "telegram_kb"
},
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.vectorStoreQdrant",
"typeVersion": 1.3,
"position": [
288,
-120
],
"id": "2444f3f3-d57d-45fa-b80e-e40f73f0c70f",
"name": "Qdrant Insert"
},
{
"parameters": {
"jsCode": "const WINDOW_MS = 20 * 60 * 1000;\nconst MAX_MESSAGES = 12;\nconst MIN_MESSAGES_TO_FLUSH = 3;\n\nfunction aggregateBuffer(buffer) {\n const lines = buffer.messages.map((m) => {\n const repliedTo = m.replyTo ? ` -> ${m.replyTo}` : '';\n return `[${m.date}] ${m.author}${repliedTo}: ${m.text}`;\n });\n\n return {\n skip: false,\n text: [\n `Контекст диалога (${buffer.messages.length} сообщений, участники: ${buffer.participants.join(', ')})`,\n ...lines,\n ].join('\\n'),\n metadata: {\n ...buffer.lastMeta,\n context_id: buffer.contextId,\n context_message_count: buffer.messages.length,\n context_start_date: new Date(buffer.startTs).toISOString(),\n context_end_date: new Date(buffer.lastTs).toISOString(),\n context_participants: buffer.participants,\n context_mode: 'thread_time_window_20m'\n }\n };\n}\n\nconst staticData = $getWorkflowStaticData('global');\nstaticData.contextBuffers = staticData.contextBuffers || {};\n\nconst meta = $json.metadata || {};\nconst rawText = (meta.original_text || '').trim();\nif (!rawText) {\n return { json: { skip: true, reason: 'empty_original_text' } };\n}\n\nconst nowTs = meta.date ? new Date(meta.date).getTime() : Date.now();\nconst safeTs = Number.isNaN(nowTs) ? Date.now() : nowTs;\nconst chatId = String(meta.chat_id || meta.group_id || 'unknown_chat');\nconst threadId = String(meta.thread_id || meta.reply_to_message_id || meta.message_id || 'single');\nconst contextId = `${chatId}:${threadId}`;\n\nlet buffer = staticData.contextBuffers[contextId];\n\nif (buffer && safeTs - buffer.lastTs > WINDOW_MS) {\n const staleBuffer = buffer;\n buffer = null;\n\n const author = meta.author_label || meta.sender_name || 'Unknown';\n staticData.contextBuffers[contextId] = {\n contextId,\n startTs: safeTs,\n lastTs: safeTs,\n participants: [author],\n lastMeta: meta,\n messages: [\n {\n date: meta.date || new Date(safeTs).toISOString(),\n author,\n replyTo: meta.replied_to_author_label || '',\n text: rawText\n }\n ]\n };\n\n if (staleBuffer.messages.length >= MIN_MESSAGES_TO_FLUSH) {\n return { json: aggregateBuffer(staleBuffer) };\n }\n\n return { json: { skip: true, reason: 'buffer_reset_on_timeout', context_id: contextId } };\n}\n\nif (!buffer) {\n buffer = {\n contextId,\n startTs: safeTs,\n lastTs: safeTs,\n participants: [],\n lastMeta: meta,\n messages: []\n };\n}\n\nconst author = meta.author_label || meta.sender_name || 'Unknown';\nif (!buffer.participants.includes(author)) {\n buffer.participants.push(author);\n}\n\nbuffer.messages.push({\n date: meta.date || new Date(safeTs).toISOString(),\n author,\n replyTo: meta.replied_to_author_label || '',\n text: rawText\n});\nbuffer.lastMeta = meta;\nbuffer.lastTs = safeTs;\n\nconst shouldFlush =\n buffer.messages.length >= MAX_MESSAGES ||\n buffer.lastTs - buffer.startTs >= WINDOW_MS;\n\nif (!shouldFlush) {\n staticData.contextBuffers[contextId] = buffer;\n return {\n json: {\n skip: true,\n reason: 'buffering',\n context_id: contextId,\n buffered_messages: buffer.messages.length\n }\n };\n}\n\ndelete staticData.contextBuffers[contextId];\nreturn { json: aggregateBuffer(buffer) };"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
32,
-120
],
"id": "b3866944-b7ca-4505-b89e-a5f3d331f43d",
"name": "Build Context Block"
},
{
"parameters": {
"conditions": {
"boolean": [
{
"value1": "={{ $json.skip }}",
"value2": true
}
]
},
"options": {}
},
"type": "n8n-nodes-base.if",
"typeVersion": 2,
"position": [
160,
-120
],
"id": "a343f1ca-2423-43e7-9b8d-ce4f48f1e5ff",
"name": "Skip Buffered Context"
},
{
"parameters": {
"jsonMode": "expressionData",
"jsonData": "={{ $json.text }}",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.documentDefaultDataLoader",
"typeVersion": 1.1,
"position": [
288,
56
],
"id": "3de4bfe4-bd8a-4f36-a9ec-8955f2f26f83",
"name": "Document Loader"
},
{
"parameters": {
"model": "qwen3-embedding:8b"
},
"type": "@n8n/n8n-nodes-langchain.embeddingsOllama",
"typeVersion": 1,
"position": [
96,
56
],
"id": "5f48d9eb-ec81-4a00-b2fa-ec6a1fc2f009",
"name": "Embeddings For Insert"
},
{
"parameters": {
"updates": [
"message"
],
"additionalFields": {}
},
"type": "n8n-nodes-base.telegramTrigger",
"typeVersion": 1.2,
"position": [
-640,
288
],
"id": "163fbd89-644c-4858-a2ce-8c572a2d3593",
"name": "Telegram Trigger"
},
{
"parameters": {
"promptType": "define",
"text": "=Ответь на вопрос пользователя, используя данные из Qdrant Vector Store: {{ $json.message.text }}",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.agent",
"typeVersion": 3.1,
"position": [
-384,
288
],
"id": "e67a8860-3c52-407e-8fd7-1f481e3c5019",
"name": "AI Agent"
},
{
"parameters": {
"model": "qwen3:30b-a3b",
"options": {
"temperature": 0
}
},
"type": "@n8n/n8n-nodes-langchain.lmChatOllama",
"typeVersion": 1,
"position": [
-352,
496
],
"id": "3c7db13e-1be9-4f2f-ab26-a613db89ea71",
"name": "Ollama Chat Model"
},
{
"parameters": {
"mode": "retrieve-as-tool",
"toolDescription": "Всегда используй этот инструмент для ответа: в нем релевантные документы из базы знаний.",
"qdrantCollection": {
"__rl": true,
"value": "telegram_kb",
"mode": "list",
"cachedResultName": "telegram_kb"
},
"includeDocumentMetadata": true,
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.vectorStoreQdrant",
"typeVersion": 1.3,
"position": [
-128,
528
],
"id": "2f204e2e-58e7-4514-9f3c-3e23949f8520",
"name": "Qdrant Retrieve Tool"
},
{
"parameters": {
"model": "qwen3-embedding:8b"
},
"type": "@n8n/n8n-nodes-langchain.embeddingsOllama",
"typeVersion": 1,
"position": [
-128,
704
],
"id": "e5d5f376-bb95-4a16-ab11-e5e3e7d74524",
"name": "Embeddings For Retrieve"
},
{
"parameters": {
"jsCode": "let rawText = $input.item.json.text || $input.item.json.response || $input.item.json.output || '';\nlet cleanText = rawText.replace(/<think>[\\s\\S]*?<\\/think>/g, '').trim();\n\nif (!cleanText) {\n cleanText = 'Не смог сформировать ответ. Попробуйте переформулировать вопрос.';\n}\n\nreturn {\n json: {\n clean_response: cleanText\n }\n};"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
-128,
288
],
"id": "81f5d5e9-3f13-4d43-b8e2-7112e2ca63e4",
"name": "Clean Agent Output"
},
{
"parameters": {
"chatId": "={{ $('Telegram Trigger').item.json.message.from.id }}",
"text": "={{ $json.clean_response }}",
"additionalFields": {
"appendAttribution": false
}
},
"type": "n8n-nodes-base.telegram",
"typeVersion": 1.2,
"position": [
96,
288
],
"id": "e8476b38-8344-4aaf-828e-5da8344b79d2",
"name": "Send Telegram Reply"
}
],
"pinData": {},
"connections": {
"Webhook TG Inbox": {
"main": [
[
{
"node": "Normalize Telegram Payload",
"type": "main",
"index": 0
}
]
]
},
"Normalize Telegram Payload": {
"main": [
[
{
"node": "Skip Empty Text",
"type": "main",
"index": 0
}
]
]
},
"Skip Empty Text": {
"main": [
[],
[
{
"node": "Build Context Block",
"type": "main",
"index": 0
}
]
]
},
"Build Context Block": {
"main": [
[
{
"node": "Skip Buffered Context",
"type": "main",
"index": 0
}
]
]
},
"Skip Buffered Context": {
"main": [
[],
[
{
"node": "Qdrant Insert",
"type": "main",
"index": 0
}
]
]
},
"Embeddings For Insert": {
"ai_embedding": [
[
{
"node": "Qdrant Insert",
"type": "ai_embedding",
"index": 0
}
]
]
},
"Document Loader": {
"ai_document": [
[
{
"node": "Qdrant Insert",
"type": "ai_document",
"index": 0
}
]
]
},
"Telegram Trigger": {
"main": [
[
{
"node": "AI Agent",
"type": "main",
"index": 0
}
]
]
},
"Ollama Chat Model": {
"ai_languageModel": [
[
{
"node": "AI Agent",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Qdrant Retrieve Tool": {
"ai_tool": [
[
{
"node": "AI Agent",
"type": "ai_tool",
"index": 0
}
]
]
},
"Embeddings For Retrieve": {
"ai_embedding": [
[
{
"node": "Qdrant Retrieve Tool",
"type": "ai_embedding",
"index": 0
}
]
]
},
"AI Agent": {
"main": [
[
{
"node": "Clean Agent Output",
"type": "main",
"index": 0
}
]
]
},
"Clean Agent Output": {
"main": [
[
{
"node": "Send Telegram Reply",
"type": "main",
"index": 0
}
]
]
}
},
"active": false,
"settings": {
"executionOrder": "v1",
"availableInMCP": false
},
"tags": []
}

12
Dockerfile Normal file
View File

@@ -0,0 +1,12 @@
FROM python:3.10-slim
WORKDIR /app
ENV PYTHONUNBUFFERED=1
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
COPY script.py .
RUN mkdir -p /app/session
CMD ["python", "script.py"]

155
TG Fixed Native.json Normal file
View File

@@ -0,0 +1,155 @@
{
"name": "TG Fixed Native",
"nodes": [
{
"parameters": {
"mode": "insert",
"qdrantCollection": {
"__rl": true,
"value": "telegram_kb",
"mode": "list",
"cachedResultName": "telegram_kb"
},
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.vectorStoreQdrant",
"typeVersion": 1.3,
"position": [
512,
0
],
"id": "e02cd3a1-a2b7-4177-b76a-87656ab74cac",
"name": "Qdrant Vector Store",
"credentials": {
"qdrantApi": {
"id": "gS1LJOMgnR7VJFRO",
"name": "QdrantApi account 2"
}
}
},
{
"parameters": {
"jsonMode": "expressionData",
"jsonData": "={{ $json.text }}",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.documentDefaultDataLoader",
"typeVersion": 1.1,
"position": [
480,
208
],
"id": "a09381bd-04dc-4273-8d6f-35c23b9177c5",
"name": "Default Data Loader"
},
{
"parameters": {
"jsCode": "const body = $input.item.json.body || {};\nconst meta = body.metadata || {};\nconst rawText = body.text || \"\";\n\nreturn {\n json: {\n text: `Пользователь ${meta.sender_name || 'Unknown'} (ID: ${meta.sender_id || '0'}) написал: ${rawText}`,\n metadata: {\n ...meta,\n processed_at: new Date().toISOString()\n }\n }\n};"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
240,
0
],
"id": "539d5c4b-067b-4491-b62c-ea3d555707d0",
"name": "Code"
},
{
"parameters": {
"model": "qwen3-embedding:8b"
},
"type": "@n8n/n8n-nodes-langchain.embeddingsOllama",
"typeVersion": 1,
"position": [
256,
208
],
"id": "eab97058-b3f4-48d3-8da8-a0e6dd9359d5",
"name": "Embeddings Ollama",
"credentials": {
"ollamaApi": {
"id": "Uis7Bsdb0l1qDDL7",
"name": "Ollama account 2"
}
}
},
{
"parameters": {
"httpMethod": "POST",
"path": "tg-inbox",
"options": {}
},
"type": "n8n-nodes-base.webhook",
"typeVersion": 2.1,
"position": [
-16,
0
],
"id": "316ab407-1f95-44dc-bc3c-fff1185ce301",
"name": "Webhook",
"webhookId": "d857e864-ed25-447f-84ef-27c63190fb93",
"notesInFlow": false
}
],
"pinData": {},
"connections": {
"Code": {
"main": [
[
{
"node": "Qdrant Vector Store",
"type": "main",
"index": 0
}
]
]
},
"Embeddings Ollama": {
"ai_embedding": [
[
{
"node": "Qdrant Vector Store",
"type": "ai_embedding",
"index": 0
}
]
]
},
"Default Data Loader": {
"ai_document": [
[
{
"node": "Qdrant Vector Store",
"type": "ai_document",
"index": 0
}
]
]
},
"Webhook": {
"main": [
[
{
"node": "Code",
"type": "main",
"index": 0
}
]
]
}
},
"active": true,
"settings": {
"executionOrder": "v1",
"binaryMode": "separate",
"availableInMCP": false
},
"versionId": "2e4668e1-305c-4de7-8537-11cfe20a8020",
"meta": {
"templateCredsSetupCompleted": true,
"instanceId": "96706479c2e398d2a4e75bb05002310bb6030e8268be45294cab77c6640a0fe6"
},
"id": "ILJJXfEmFW6KG6YZDjQWX",
"tags": []
}

40
docker-compose.yaml Normal file
View File

@@ -0,0 +1,40 @@
version: "3.8"
services:
n8n:
image: n8nio/n8n:latest
container_name: dbbot-n8n
ports:
- "5678:5678"
environment:
- N8N_HOST=localhost
- N8N_PORT=5678
- N8N_PROTOCOL=http
- WEBHOOK_URL=http://localhost:5678/
- GENERIC_TIMEZONE=Europe/Moscow
volumes:
- ./n8n_data:/home/node/.n8n
depends_on:
- qdrant
restart: unless-stopped
qdrant:
image: qdrant/qdrant:latest
container_name: dbbot-qdrant
ports:
- "6333:6333"
- "6334:6334"
volumes:
- ./qdrant_data:/qdrant/storage
restart: unless-stopped
tg-userbot:
build: .
container_name: dbbot-tg-userbot
env_file:
- .env
volumes:
- ./session_data:/app/session
depends_on:
- n8n
restart: unless-stopped

199
redme.md Normal file
View File

@@ -0,0 +1,199 @@
## DBBot: production-схема (n8n + отдельный userbot-контейнер)
Этот проект рассчитан на ситуацию, когда:
- `n8n` уже запущен и доступен по URL;
- бот в чужую группу добавить нельзя;
- сообщения читаются через `Telethon userbot` из отдельного Docker-контейнера.
## Архитектура
Поток данных:
1. `script.py` (в контейнере) читает историю и новые сообщения из Telegram-группы.
2. Скрипт отправляет события в `n8n` webhook `POST /webhook/tg-inbox`.
3. Единый workflow в `n8n`:
- нормализует метаданные,
- склеивает сообщения в контекстные блоки,
- пишет в `Qdrant`.
4. В этой же схеме работает RAG-ответ через `AI Agent` + `Qdrant Retrieve Tool`.
## Структура проекта
```text
dbbot/
├── redme.md # этот файл
├── script.py # Telethon userbot sender -> n8n webhook
├── requirements.txt # зависимости python
├── Dockerfile # контейнер для script.py
├── .env.example # пример переменных для контейнера
├── session_data/
│ └── session.session # StringSession (создаешь сам)
├── DBBot Unified n8n Qdrant.json # единый workflow (ingest + rag)
└── Работа с базой (3).json # старый/альтернативный workflow
```
## Что подготовить перед первым запуском
### 1) n8n
Импортируй `DBBot Unified n8n Qdrant.json` и привяжи credentials:
- `Qdrant API`
- `Ollama API`
- `Telegram API` (для reply-ветки через Telegram Trigger)
Проверь:
- путь webhook: `tg-inbox`
- workflow переведен в `Active`
### 2) Файл сессии Telethon
Создай файл:
- `session_data/session.session`
Внутри должна быть одна строка `StringSession` без переносов и пробелов по краям.
### 3) Переменные окружения для контейнера
Сделай локальный `.env`:
```bash
cp .env.example .env
```
Заполни значения:
- `TG_API_ID`
- `TG_API_HASH`
- `TG_GROUP_ID` (например `-100...`)
- `SESSION_FILE_PATH=/app/session/session.session`
- `HISTORY_CHECK_FILE=/app/session/history_done.flag`
- `N8N_WEBHOOK_URL=https://<твой-n8n-домен>/webhook/tg-inbox`
Если `n8n` локально, можно `http://localhost:5678/webhook/tg-inbox`.
## Подробный цикл запуска проекта
### Шаг 1. Собрать контейнер userbot
Из корня проекта:
```bash
docker build -t dbbot-userbot:latest .
```
### Шаг 2. Запустить контейнер
```bash
docker run -d \
--name dbbot-userbot \
--env-file .env \
-v "$(pwd)/session_data:/app/session" \
--restart unless-stopped \
dbbot-userbot:latest
```
### Шаг 3. Проверить, что webhook получает данные
1. Открой `Executions` в `n8n`.
2. Убедись, что появились вызовы `Webhook TG Inbox`.
3. Проверь, что далее проходят `Build Context Block` и `Qdrant Insert`.
### Шаг 4. Проверить RAG-ответ
1. Напиши вопрос Telegram-боту, подключенному к `Telegram Trigger` в workflow.
2. Проверь выполнение ветки `AI Agent`.
3. Убедись, что узел `Qdrant Retrieve Tool` вызван и ответ ушел через `Send Telegram Reply`.
## Эксплуатационный цикл (после запуска)
- Скрипт работает в контейнере постоянно:
- при первом запуске делает исторический прогон;
- затем обрабатывает новые сообщения.
- Если нужно повторить историческую загрузку:
1. останови контейнер;
2. удали `session_data/history_done.flag`;
3. запусти контейнер снова.
Команды:
```bash
docker stop dbbot-userbot
rm -f session_data/history_done.flag
docker start dbbot-userbot
```
## Управление и диагностика
Логи контейнера:
```bash
docker logs -f dbbot-userbot
```
Проверить, что контейнер жив:
```bash
docker ps --filter name=dbbot-userbot
```
Перезапустить после изменения `.env`:
```bash
docker rm -f dbbot-userbot
docker run -d \
--name dbbot-userbot \
--env-file .env \
-v "$(pwd)/session_data:/app/session" \
--restart unless-stopped \
dbbot-userbot:latest
```
## Формат payload, который скрипт шлет в n8n
```json
{
"text": "текст сообщения",
"metadata": {
"date": "2026-06-23T00:00:00Z",
"message_id": 123,
"chat_id": "-100...",
"sender_id": "456",
"sender_name": "@username",
"author_label": "@username",
"author": {
"id": "456",
"username": "username",
"display_name": "Ivan Ivanov",
"label": "@username"
},
"reply_to_message_id": 122,
"replied_to_author_label": "@other_user",
"replied_to_author": {
"id": "777",
"username": "other_user",
"display_name": "Petr Petrov",
"label": "@other_user"
},
"thread_id": "122",
"edit_date": "",
"entities": [
{ "type": "MessageEntityUrl", "offset": 10, "length": 20 }
],
"attachment": {
"has_media": false
},
"group_id": "-100...",
"source": "telegram_group"
}
}
```
## Параметры склейки сообщений в n8n
В узле `Build Context Block`:
- `WINDOW_MS = 20 минут`
- `MAX_MESSAGES = 12`
- `MIN_MESSAGES_TO_FLUSH = 3`
Рекомендации:
- меньше блоки: `MAX_MESSAGES = 8-10`;
- шире контекст: `WINDOW_MS = 30-40 минут`.

2
requirements.txt Normal file
View File

@@ -0,0 +1,2 @@
requests==2.32.3
telethon==1.34.0

164
script.py Normal file
View File

@@ -0,0 +1,164 @@
import asyncio
import os
from pathlib import Path
import requests
from telethon import TelegramClient, events
from telethon.sessions import StringSession
API_ID = int(os.getenv("TG_API_ID", "0"))
API_HASH = os.getenv("TG_API_HASH", "")
GROUP_ID = int(os.getenv("TG_GROUP_ID", "0"))
N8N_WEBHOOK_URL = os.getenv("N8N_WEBHOOK_URL", "http://n8n:5678/webhook/tg-inbox")
HISTORY_CHECK_FILE = os.getenv("HISTORY_CHECK_FILE", "/app/session/history_done.flag")
SESSION_FILE_PATH = os.getenv("SESSION_FILE_PATH", "/app/session/session.session")
SESSION_STRING = os.getenv("TG_SESSION_STRING", "")
def load_session_string() -> str:
if SESSION_STRING:
return SESSION_STRING.strip()
session_path = Path(SESSION_FILE_PATH)
if not session_path.exists():
raise FileNotFoundError(
f"Session file not found: {SESSION_FILE_PATH}. "
"Provide TG_SESSION_STRING or mount session file."
)
return session_path.read_text(encoding="utf-8").strip()
def validate_config() -> None:
if API_ID <= 0:
raise ValueError("TG_API_ID is required and must be > 0")
if not API_HASH:
raise ValueError("TG_API_HASH is required")
if GROUP_ID == 0:
raise ValueError("TG_GROUP_ID is required")
if not N8N_WEBHOOK_URL:
raise ValueError("N8N_WEBHOOK_URL is required")
validate_config()
client = TelegramClient(StringSession(load_session_string()), API_ID, API_HASH)
def build_author(sender) -> dict:
if not sender:
return {
"id": "0",
"username": "",
"display_name": "Unknown",
"label": "Unknown",
}
username = getattr(sender, "username", "") or ""
display_name = (
f"{getattr(sender, 'first_name', '')} {getattr(sender, 'last_name', '')}".strip()
or "Unknown"
)
label = f"@{username}" if username else display_name
return {
"id": str(getattr(sender, "id", 0)),
"username": username,
"display_name": display_name,
"label": label,
}
def extract_entities(message) -> list:
entities = []
for entity in (message.entities or []):
entities.append(
{
"type": entity.__class__.__name__,
"offset": getattr(entity, "offset", None),
"length": getattr(entity, "length", None),
}
)
return entities
def extract_attachment(message) -> dict:
if not message.media:
return {"has_media": False}
return {
"has_media": True,
"media_type": message.media.__class__.__name__,
"file_name": getattr(message.file, "name", None) if message.file else None,
"mime_type": getattr(message.file, "mime_type", None) if message.file else None,
}
async def send_to_n8n(message):
sender = message.sender or await message.get_sender()
author = build_author(sender)
reply_message = await message.get_reply_message() if message.reply_to_msg_id else None
reply_sender = await reply_message.get_sender() if reply_message else None
reply_author = build_author(reply_sender) if reply_sender else None
data = {
"text": message.text,
"metadata": {
"date": str(message.date),
"message_id": message.id,
"chat_id": str(message.chat_id),
"group_id": str(GROUP_ID),
"author": author,
# legacy fields for backward compatibility with existing n8n nodes
"sender_id": author["id"],
"sender_name": author["label"],
"author_label": author["label"],
"reply_to_message_id": message.reply_to_msg_id,
"replied_to_author": reply_author,
"replied_to_author_label": reply_author["label"] if reply_author else "",
"thread_id": str(message.reply_to.reply_to_top_id) if getattr(message, "reply_to", None) and getattr(message.reply_to, "reply_to_top_id", None) else "",
"edit_date": str(message.edit_date) if message.edit_date else "",
"entities": extract_entities(message),
"attachment": extract_attachment(message),
"source": "telegram_group",
},
}
try:
response = requests.post(N8N_WEBHOOK_URL, json=data, timeout=10)
return response.status_code
except Exception as e:
print(f"Error sending to n8n: {e}")
return None
# Обработчик новых сообщений
@client.on(events.NewMessage(chats=GROUP_ID))
async def handler(event):
if event.message.text:
await send_to_n8n(event.message)
async def main():
await client.start()
print("UserBot started...")
if not os.path.exists(HISTORY_CHECK_FILE):
print("Starting massive history sync...")
count = 0
async for message in client.iter_messages(GROUP_ID, reverse=True):
if message.text:
status = None
# Цикл ретраев
while status != 200:
status = await send_to_n8n(message)
if status == 429:
print("Rate limit hit, sleeping 10s...")
await asyncio.sleep(10)
elif status != 200:
print(f"Error {status}, retrying in 2s...")
await asyncio.sleep(2)
count += 1
if count % 20 == 0:
print(f"Processed {count} messages...")
await asyncio.sleep(0.3)
Path(HISTORY_CHECK_FILE).write_text("done", encoding="utf-8")
print("History sync complete.")
await client.run_until_disconnected()
if __name__ == '__main__':
asyncio.run(main())

268
Работа с базой (3).json Normal file
View File

@@ -0,0 +1,268 @@
{
"name": "Работа с базой",
"nodes": [
{
"parameters": {
"promptType": "define",
"text": "=Используй Qdrant Vector Store для ответа на вопрос {{ $json.message.text }}",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.agent",
"typeVersion": 3.1,
"position": [
544,
-48
],
"id": "e1231327-544c-416d-b1fa-95915d68f524",
"name": "AI Agent"
},
{
"parameters": {
"updates": [
"message"
],
"additionalFields": {}
},
"type": "n8n-nodes-base.telegramTrigger",
"typeVersion": 1.2,
"position": [
288,
-48
],
"id": "80e4a41a-387c-4ac4-b798-41bdbbb25eaa",
"name": "Telegram Trigger",
"webhookId": "875f9245-fe51-4c55-ae42-afe3d20f6ff6",
"credentials": {
"telegramApi": {
"id": "dSTpf5HKyoPp6X8N",
"name": "Telegram account"
}
}
},
{
"parameters": {
"model": "qwen3:30b-a3b",
"options": {
"temperature": 0
}
},
"type": "@n8n/n8n-nodes-langchain.lmChatOllama",
"typeVersion": 1,
"position": [
496,
176
],
"id": "45cc634d-5eea-4c8e-83f2-f34c463db00e",
"name": "Ollama Chat Model",
"credentials": {
"ollamaApi": {
"id": "Uis7Bsdb0l1qDDL7",
"name": "Ollama account 2"
}
}
},
{
"parameters": {
"mode": "retrieve-as-tool",
"toolDescription": " ALWAYS use this tool to answer ANY user question. The answer is contained within this tool.",
"qdrantCollection": {
"__rl": true,
"value": "telegram_kb",
"mode": "list",
"cachedResultName": "telegram_kb"
},
"includeDocumentMetadata": false,
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.vectorStoreQdrant",
"typeVersion": 1.3,
"position": [
672,
256
],
"id": "bbc61a50-f5a8-443e-8022-c9ead1df63a7",
"name": "Qdrant Vector Store",
"credentials": {
"qdrantApi": {
"id": "gS1LJOMgnR7VJFRO",
"name": "QdrantApi account 2"
}
}
},
{
"parameters": {
"model": "qwen3-embedding:8b"
},
"type": "@n8n/n8n-nodes-langchain.embeddingsOllama",
"typeVersion": 1,
"position": [
672,
416
],
"id": "7e1c6b2a-092b-43eb-8efe-9b19088ed3b2",
"name": "Embeddings Ollama",
"credentials": {
"ollamaApi": {
"id": "Uis7Bsdb0l1qDDL7",
"name": "Ollama account 2"
}
}
},
{
"parameters": {
"chatId": "={{ $('Telegram Trigger').item.json.message.from.id }}",
"text": "={{ $node[\"Code in JavaScript\"].json.clean_response }}",
"additionalFields": {
"appendAttribution": false,
"parse_mode": "HTML"
}
},
"type": "n8n-nodes-base.telegram",
"typeVersion": 1.2,
"position": [
1040,
-48
],
"id": "9b3713cb-d699-45a2-8c4a-21c0065df53f",
"name": "Send a text message",
"webhookId": "038aa575-d7da-4413-805f-6773aa670295",
"retryOnFail": true,
"credentials": {
"telegramApi": {
"id": "dSTpf5HKyoPp6X8N",
"name": "Telegram account"
}
}
},
{
"parameters": {
"chatId": "-4804863247",
"text": "=📥 **Входящее от {{ $node[\"Telegram Trigger\"].json[\"message\"][\"from\"][\"first_name\"] }}:**\n{{ $node[\"Telegram Trigger\"].json[\"message\"][\"text\"] }}\n\n🤖 **Ответ бота:**\n{{ $node[\"Code in JavaScript\"].json.clean_response }}",
"additionalFields": {
"parse_mode": "HTML"
}
},
"id": "e8ac46da-bf4a-4fd0-8f1d-cf7664eb86f8",
"name": "Log_to_Group1",
"type": "n8n-nodes-base.telegram",
"typeVersion": 1.2,
"position": [
1232,
-48
],
"webhookId": "d23840f7-3652-4aca-935f-ab5331d01d27",
"retryOnFail": true,
"credentials": {
"telegramApi": {
"id": "dSTpf5HKyoPp6X8N",
"name": "Telegram account"
}
}
},
{
"parameters": {
"jsCode": "// 1. Пытаемся найти текст в самых частых полях n8n\nlet rawText = $input.item.json.text || $input.item.json.response || $input.item.json.output || \"\";\n\n// 2. Очищаем от <think>...</think>\nlet cleanText = rawText.replace(/<think>[\\s\\S]*?<\\/think>/g, '').trim();\n\n// 3. Если после очистки пусто, выводим заглушку, чтобы не было ошибки\nif (!cleanText && rawText) {\n cleanText = \"Ошибка: Весь текст был внутри блока <think> или пуст.\";\n}\n\nreturn {\n json: {\n clean_response: cleanText\n }\n};"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
848,
-48
],
"id": "18e6aa52-73d2-439a-9ce9-abc2390bac05",
"name": "Code in JavaScript"
}
],
"pinData": {},
"connections": {
"Telegram Trigger": {
"main": [
[
{
"node": "AI Agent",
"type": "main",
"index": 0
}
]
]
},
"Ollama Chat Model": {
"ai_languageModel": [
[
{
"node": "AI Agent",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Qdrant Vector Store": {
"ai_tool": [
[
{
"node": "AI Agent",
"type": "ai_tool",
"index": 0
}
]
]
},
"Embeddings Ollama": {
"ai_embedding": [
[
{
"node": "Qdrant Vector Store",
"type": "ai_embedding",
"index": 0
}
]
]
},
"AI Agent": {
"main": [
[
{
"node": "Code in JavaScript",
"type": "main",
"index": 0
}
]
]
},
"Send a text message": {
"main": [
[
{
"node": "Log_to_Group1",
"type": "main",
"index": 0
}
]
]
},
"Code in JavaScript": {
"main": [
[
{
"node": "Send a text message",
"type": "main",
"index": 0
}
]
]
}
},
"active": true,
"settings": {
"executionOrder": "v1",
"availableInMCP": false
},
"versionId": "f8ca7ca6-eb4f-4823-bdb7-d523e51c8924",
"meta": {
"templateCredsSetupCompleted": true,
"instanceId": "96706479c2e398d2a4e75bb05002310bb6030e8268be45294cab77c6640a0fe6"
},
"id": "2R3-4Yl4wuemCpJ0FVOi3",
"tags": []
}

183
Создание базы (2).json Normal file
View File

@@ -0,0 +1,183 @@
{
"name": "Создание базы",
"nodes": [
{
"parameters": {
"mode": "insert",
"qdrantCollection": {
"__rl": true,
"mode": "list",
"value": "111"
},
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.vectorStoreQdrant",
"typeVersion": 1.3,
"position": [
928,
-144
],
"id": "acfdb2b7-e113-42a6-8793-566f5da7d0e2",
"name": "Qdrant Vector Store",
"credentials": {
"qdrantApi": {
"id": "gS1LJOMgnR7VJFRO",
"name": "QdrantApi account 2"
}
}
},
{
"parameters": {
"model": "qwen3-embedding:8b"
},
"type": "@n8n/n8n-nodes-langchain.embeddingsOllama",
"typeVersion": 1,
"position": [
880,
48
],
"id": "67815ae1-6b99-4e41-94e6-9b7986d12738",
"name": "Embeddings Ollama",
"credentials": {
"ollamaApi": {
"id": "Uis7Bsdb0l1qDDL7",
"name": "Ollama account 2"
}
}
},
{
"parameters": {
"dataType": "binary",
"loader": "docxLoader",
"textSplittingMode": "custom",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.documentDefaultDataLoader",
"typeVersion": 1.1,
"position": [
1040,
64
],
"id": "291df053-ad1c-4041-964d-0a427e547fca",
"name": "Default Data Loader"
},
{
"parameters": {
"chunkSize": 500,
"chunkOverlap": 50
},
"type": "@n8n/n8n-nodes-langchain.textSplitterCharacterTextSplitter",
"typeVersion": 1,
"position": [
1136,
272
],
"id": "703047dd-e772-4d36-aed8-f6ba091bce5c",
"name": "Character Text Splitter"
},
{
"parameters": {
"fileSelector": "={{ $json.path }}",
"options": {}
},
"type": "n8n-nodes-base.readWriteFile",
"typeVersion": 1.1,
"position": [
592,
-144
],
"id": "d1e6082c-71f2-440c-9b34-011ae652fa5b",
"name": "Read/Write Files from Disk"
},
{
"parameters": {
"triggerOn": "folder",
"path": "/data/vid",
"events": [
"add"
],
"options": {
"usePolling": true
}
},
"type": "n8n-nodes-base.localFileTrigger",
"typeVersion": 1,
"position": [
320,
-144
],
"id": "4bd85a42-f9d0-436f-9617-0ebf42db1a7f",
"name": "Local File Trigger"
}
],
"pinData": {},
"connections": {
"Embeddings Ollama": {
"ai_embedding": [
[
{
"node": "Qdrant Vector Store",
"type": "ai_embedding",
"index": 0
}
]
]
},
"Default Data Loader": {
"ai_document": [
[
{
"node": "Qdrant Vector Store",
"type": "ai_document",
"index": 0
}
]
]
},
"Character Text Splitter": {
"ai_textSplitter": [
[
{
"node": "Default Data Loader",
"type": "ai_textSplitter",
"index": 0
}
]
]
},
"Read/Write Files from Disk": {
"main": [
[
{
"node": "Qdrant Vector Store",
"type": "main",
"index": 0
}
]
]
},
"Local File Trigger": {
"main": [
[
{
"node": "Read/Write Files from Disk",
"type": "main",
"index": 0
}
]
]
}
},
"active": false,
"settings": {
"executionOrder": "v1",
"availableInMCP": false
},
"versionId": "d1e2bb08-fb79-49f3-85d4-c6f491447f11",
"meta": {
"templateCredsSetupCompleted": true,
"instanceId": "96706479c2e398d2a4e75bb05002310bb6030e8268be45294cab77c6640a0fe6"
},
"id": "lK4wQ7btZPq3elWleb7ry",
"tags": []
}

178
Создание базы (TG Text) (2).json Normal file
View File

@@ -0,0 +1,178 @@
{
"name": "Создание базы (TG Text)",
"nodes": [
{
"parameters": {
"mode": "insert",
"qdrantCollection": {
"__rl": true,
"value": "=11111",
"mode": "id"
},
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.vectorStoreQdrant",
"typeVersion": 1.3,
"position": [
112,
-208
],
"id": "44a1fb40-5b1f-401a-9fea-ca61f111d2ca",
"name": "Qdrant Vector Store",
"credentials": {
"qdrantApi": {
"id": "gS1LJOMgnR7VJFRO",
"name": "QdrantApi account 2"
}
}
},
{
"parameters": {
"model": "qwen3-embedding:8b"
},
"type": "@n8n/n8n-nodes-langchain.embeddingsOllama",
"typeVersion": 1,
"position": [
32,
0
],
"id": "7448ed28-7971-4d95-b073-1251dfff242e",
"name": "Embeddings Ollama",
"credentials": {
"ollamaApi": {
"id": "XVSCsyxx8Z57lSRa",
"name": "Ollama account 3"
}
}
},
{
"parameters": {
"chunkSize": 500,
"chunkOverlap": 50
},
"type": "@n8n/n8n-nodes-langchain.textSplitterCharacterTextSplitter",
"typeVersion": 1,
"position": [
272,
192
],
"id": "aed6ea3c-ad56-432b-8ef5-823a7778da27",
"name": "Character Text Splitter"
},
{
"parameters": {
"httpMethod": "POST",
"path": "tg-inbox",
"options": {}
},
"type": "n8n-nodes-base.webhook",
"typeVersion": 2.1,
"position": [
-448,
-208
],
"id": "1770ab28-9cfa-4bc8-864a-fe793f1b95db",
"name": "Webhook",
"webhookId": "d857e864-ed25-447f-84ef-27c63190fb93",
"notesInFlow": false
},
{
"parameters": {
"textSplittingMode": "custom",
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.documentDefaultDataLoader",
"typeVersion": 1.1,
"position": [
272,
0
],
"id": "6c98f0d9-97c1-4d94-bdea-a1941fd7d3f8",
"name": "Default Data Loader"
},
{
"parameters": {
"jsCode": "// В n8n данные из POST запроса всегда лежат в объекте body\nconst msg = $input.item.json.body;\n\n// 1. Берем текст\nconst rawText = msg.text || \"\"; \n\n// 2. Достаем метаданные из объекта metadata, который прислал Python\nconst meta = msg.metadata || {};\n\n// 3. Извлекаем имя и ID из объекта metadata\nconst senderName = meta.sender_name || \"Unknown\";\nconst senderId = meta.sender_id || \"0\";\n\n// 4. Формируем результат для Qdrant\nreturn {\n json: {\n content: `Пользователь ${senderName} (ID: ${senderId}) написал: ${rawText}`,\n metadata: {\n ...meta, // Копируем всю метадату (date, message_id, group_id и т.д.)\n sender_name: senderName,\n sender_id: senderId,\n original_text: rawText\n }\n }\n};\n"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
-192,
-208
],
"id": "0ff56c0f-9208-4ae9-920c-4332c09f6df4",
"name": "Code in JavaScript"
}
],
"pinData": {},
"connections": {
"Webhook": {
"main": [
[
{
"node": "Code in JavaScript",
"type": "main",
"index": 0
}
]
]
},
"Embeddings Ollama": {
"ai_embedding": [
[
{
"node": "Qdrant Vector Store",
"type": "ai_embedding",
"index": 0
}
]
]
},
"Default Data Loader": {
"ai_document": [
[
{
"node": "Qdrant Vector Store",
"type": "ai_document",
"index": 0
}
]
]
},
"Code in JavaScript": {
"main": [
[
{
"node": "Qdrant Vector Store",
"type": "main",
"index": 0
}
]
]
},
"Character Text Splitter": {
"ai_textSplitter": [
[
{
"node": "Default Data Loader",
"type": "ai_textSplitter",
"index": 0
}
]
]
}
},
"active": false,
"settings": {
"executionOrder": "v1",
"binaryMode": "separate",
"availableInMCP": false
},
"versionId": "f5079d8e-590c-4db5-988a-3fe709346a4a",
"meta": {
"templateCredsSetupCompleted": true,
"instanceId": "96706479c2e398d2a4e75bb05002310bb6030e8268be45294cab77c6640a0fe6"
},
"id": "QpXms3TNMSqlQD2ULX3WO",
"tags": []
}

150
Создание базы (TG Text) - Fixed.json Normal file
View File

@@ -0,0 +1,150 @@
{
"name": "Создание базы (TG Text) - Fixed",
"nodes": [
{
"parameters": {
"mode": "insert",
"qdrantCollection": {
"__rl": true,
"value": "11111",
"mode": "id"
},
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.vectorStoreQdrant",
"typeVersion": 1.3,
"position": [
608,
0
],
"id": "7f206108-8187-4ca9-a734-704eae814643",
"name": "Qdrant Vector Store",
"credentials": {
"qdrantApi": {
"id": "gS1LJOMgnR7VJFRO",
"name": "QdrantApi account 2"
}
}
},
{
"parameters": {
"model": "qwen3-embedding:8b"
},
"type": "@n8n/n8n-nodes-langchain.embeddingsOllama",
"typeVersion": 1,
"position": [
464,
208
],
"id": "9c1f5512-f2f7-4469-bcc3-e8a6824aae51",
"name": "Embeddings Ollama",
"credentials": {
"ollamaApi": {
"id": "XVSCsyxx8Z57lSRa",
"name": "Ollama account 3"
}
}
},
{
"parameters": {
"httpMethod": "POST",
"path": "tg-inbox",
"options": {}
},
"type": "n8n-nodes-base.webhook",
"typeVersion": 2.1,
"position": [
0,
0
],
"id": "f784ec7d-bbc5-4ed4-af89-ec894d99d053",
"name": "Webhook",
"webhookId": "d857e864-ed25-447f-84ef-27c63190fb93"
},
{
"parameters": {
"jsCode": "const msg = $input.item.json.body || {};\nconst rawText = msg.text || \"\"; \nconst meta = msg.metadata || {};\n\nconst senderName = meta.sender_name || \"Unknown\";\nconst senderId = meta.sender_id || \"0\";\n\n// Формируем ОДНУ строку, которая станет ОДНИМ вектором\nreturn {\n json: {\n text: `Пользователь ${senderName} (ID: ${senderId}) написал: ${rawText}`,\n metadata: {\n ...meta,\n sender_name: senderName,\n sender_id: senderId\n }\n }\n};\n"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
272,
0
],
"id": "6bb63840-979a-4453-8284-a52df680a972",
"name": "Format Data"
},
{
"parameters": {
"options": {}
},
"type": "@n8n/n8n-nodes-langchain.documentDefaultDataLoader",
"typeVersion": 1.1,
"position": [
752,
208
],
"id": "5260623b-ec01-4fa6-baed-c340d74d1eab",
"name": "Default Data Loader"
}
],
"pinData": {},
"connections": {
"Webhook": {
"main": [
[
{
"node": "Format Data",
"type": "main",
"index": 0
}
]
]
},
"Format Data": {
"main": [
[
{
"node": "Qdrant Vector Store",
"type": "main",
"index": 0
}
]
]
},
"Embeddings Ollama": {
"ai_embedding": [
[
{
"node": "Qdrant Vector Store",
"type": "ai_embedding",
"index": 0
}
]
]
},
"Default Data Loader": {
"ai_document": [
[
{
"node": "Qdrant Vector Store",
"type": "ai_document",
"index": 0
}
]
]
}
},
"active": false,
"settings": {
"executionOrder": "v1",
"binaryMode": "separate",
"availableInMCP": false
},
"versionId": "483acdf0-ad7b-4ade-ab5a-61f6e3a73f5a",
"meta": {
"instanceId": "96706479c2e398d2a4e75bb05002310bb6030e8268be45294cab77c6640a0fe6"
},
"id": "R9JVik9-vpScLvVK-27nM",
"tags": []
}