Skip to content

README.md

Latest commit

 

History

History
149 lines (111 loc) · 5.8 KB

README.md

File metadata and controls

149 lines (111 loc) · 5.8 KB
Raw

MUCTR-BOT

A Telegram bot for students of the Mendeleev University of Chemical Technology of Russia (MUCTR) — check your grades, share homework with your group, and find your way around campus without leaving Telegram.

The bot's interface is in Russian, since that is what its users speak. This README is in English.

About

MUCTR runs a student portal called EIOS, which holds your grades, checkpoint scores and exam results. Reading them means logging into a web portal — awkward on a phone, between classes, which is exactly when you want to check them.

MUCTR-BOT moves that into Telegram. You register once with your EIOS credentials, and from then on your grades are a couple of taps away. Around that core it grew the other things a group of students kept needing: somewhere to put the homework, a way to tell everyone what the canteen is serving, and floor plans for buildings that are genuinely easy to get lost in.

This is my first programming project. I built it while learning Python — partly from tutorials, and partly by watching what the university portal sent over the network and working out how to ask it the same questions.

Features

Command What it does
/start Registration: consent screen, then your EIOS login and password
/journal Your grades from EIOS — pick a course, a semester, then a subject
/sethomework · /viewhomework Shared homework for your study group, with photos, documents and voice notes
/menu · /setmenu The dorm canteen menu, kept up to date by whoever gets there first
/schedule Today's timetable
/map Floor plans for the Miusskaya and Tushino buildings
/an Broadcast an announcement to everyone in your group
/loggs · /DROPDATABASE Admin only — fetch the log files, or wipe the database

How it works

Talking to the EIOS API

EIOS has no public API and no documentation for one. The endpoints below were worked out by inspecting the requests the portal's own web interface makes:

  1. POST /accounts/authenticate/login/ with your username and password returns a token.
  2. Every subsequent request carries Authorization: Token <token>.
  3. GET /education/students/student/ returns your enrolment — from it the bot takes your EIOS education ID and your study group.
  4. GET /education/students/education/{id}/journal/full/ returns the entire journal as JSON.
  5. The bot walks that JSON — courses → semesters → subjects — and pulls out the semester rating, exam score, checkpoint values and control type for whichever subject you picked.

Requests go through a single requests.Session() with a randomised user agent.

Storing credentials

Fetching your grades means holding your EIOS password, so the bot encrypts it. Each user gets their own key from cryptography.fernet, generated at registration; the password is encrypted before it reaches the database and decrypted only when a request to EIOS is actually made.

Registration begins with an explicit consent step that links to the bot's privacy page. Declining is a valid choice — /map, /menu and /setmenu keep working without an account, since none of them touch EIOS.

Conversation state

Most of what the bot does is a conversation rather than a single command: /sethomework asks for a subject, then a caption, then the task itself. Each of these flows is a finite state machine built on aiogram's FSM — SignUpState, JournalState, HomeworkState, MapState, СomboState and AnnounceState in states/states.py — so the bot always knows which answer it is waiting for, per user.

Database access

A middleware (middlewares/db.py) opens an async SQLAlchemy session for every incoming update and hands it to the handler, so no handler has to manage a connection itself. All the actual queries live in one place, database/orm_querry.py, rather than being scattered through the handlers.

Three tables (database/models.py): UserInfo for registered users, homework for tasks posted per group, and combo for the canteen menu.

Project structure

main.py              Entry point — creates the bot, registers routers and middleware
common/              Bot commands, the EIOS API client, the timetable
handlers/            One router per feature area
  Journal/           Grades: menus and journal parsing
  Homework/          Posting and reading homework
  UserSave/          Registration, encryption, group announcements
  admin.py           Admin-only commands
  menu.py            Canteen menu
  simple.py          Maps and timetable
  ExceptionHandler.py
database/            Engine, models, and all ORM queries
middlewares/         Database session injection
states/              FSM state groups
filters/             Admin check
kbds/                Inline keyboards
LOGGING/             Logger configuration

Tech stack

  • Python 3.12
  • aiogram 3.4.1 — Telegram bot framework, using routers and FSM
  • SQLAlchemy 2.0.27 (async) with aiosqlite — ORM over SQLite
  • cryptography 42.0.1 — Fernet encryption for stored passwords
  • requests + fake-useragent — EIOS API client
  • python-dotenv — configuration

Running it locally

You need Python 3.12 and a bot token from @BotFather.

git clone https://github.com/sasha720709/MUCTR-BOT.git
cd MUCTR-BOT

python -m venv venv
venv\Scripts\activate        # Linux/macOS: source venv/bin/activate

pip install -r requirements.txt

Create a .env file next to main.py:

TOKEN=your-token-from-botfather

Then start it:

python main.py

The SQLite database is created automatically on first run, at database/dbase.db.

Note that /journal and /schedule only do anything useful against real MUCTR credentials — the rest of the bot works for anyone.

License

MIT