A Telegram chat bot made with ❤️ and ☕
- Webhooks
- Integrated web server
- Add custom API
- Built to run inside docker
- Light Docker image ~20MB (no trained model)
- Read to deploy on Caprover
- Machine learning based
- Json Input files easy to read and customize
- Easy config with env vars (list below)
- Reply on: quote, name found in message, tag
- Docker
- (Recommended) Nginx or Apache as reverse proxy
| Variable | Type | Default Value | Description |
|---|---|---|---|
| BOT_PATH | string | /tgdev | Bot request path (eg example.com/tgdev) |
| SERVER_PORT | int | 8080 | Server listen port. Changing this will break builtin Dockerfiles! |
| BOT_NAME | string | NanaoChanBot | Bot username, must match Telegram bot username |
| BOT_NAME_REGEX | string | \b@?nanao(chanbot)?\b | Regex used to find bot name in message |
| BOT_MAX_MESSAGE_SIZE | int | 256 | Max message size, if message is longer appropriate event is triggered |
| ML_MIN_SCORE | double | 0.5 | Minimum score to show ml result for message |
| SHOW_ML_SCORE | string | false | Add ml score at the end of the message |
Run: sh docker-build.sh
Note: this won't train your model!
To build and train the model run:
sh docker-train.sh
If you want to create a ready to deploy packet for your caprover node run:
sh captain-build.sh
This will build the container and create the .tar to upload!
(Note: remember to train your model before uploading the build!)