npaperbot-telegram

What

Search C++ proposals directly in Telegram.

Dependencies

Rust 1.64 or newer
Cargo

Older Rust compiler versions possibly able to build the project, but I didn't test it.

How to build

Clone this repository
cargo build --release

How to run

I recommend running this bot as a service(e.g. as systemd service) on a machine. Also, Docker images are available here: https://hub.docker.com/repository/docker/zamazan4ik/npaperbot-telegram

Configuration

The bot can be configured only with environment variables. For now there are we support the following variables:

Name	Description	Values	Default value	Required
TELOXIDE_TOKEN	Telegram bot token	Any valid and registered Telegram bot token	None	All mods
WEBHOOK_MODE	Run bot in webhook mode or long-polling mode	`true` for webhook, 'false' for long-polling	`false`	All mods
PAPERS_DATABASE_URI	HTTP(S) URI with C++ proposals JSON file	Any valid URI	`https://wg21.link/index.json`	All mods
MAX_RESULTS_PER_REQUEST	Number of at most permitted results per request. Other results will be truncated	Unsigned 8-bit integer	`20`	All mods
DATABASE_UPDATE_PERIODICITY_IN_HOURS	Papers database update periodicity in hours	Any reasonable positive i64 integer	`1`	All mods
BIND_ADDRESS	Address for binding the web-service	Any valid IP address	`0.0.0.0`	Webhook mode
BIND_PORT	Port for binding the web-service	Any valid port	`8080`	Webhook mode
HOST	Host, where Telegram will send updates in webhook mode	Any valid host address	None	Webhook mode
BOT_NAME	Telegram bot name	Any string	None	All mods

If for any variable there is no default value, and you didn't provide any value - the bot won't start. Bot automatically registers webhook (if is launched in webhook mode) with address https://$HOST/$TELOXIDE_TOKEN/api/v1/message.

How to use

Inline mode. Write any C++ proposal number (like p1000) in any paired brackets (e.g. [p1000] or {p1000}) and the bot will return all corresponding results.
Search command. Type /search pattern and bot will try to find corresponding paper. Pattern shall be a paper number or a title part or an author. Currently, search is case-insensitive (but without fuzzy search support).

Feedback

If you have any suggestions or want to report a bug - feel free to create in issue in this repo. Thank you!