russian-laws-bot

Before you begin, make sure you have Docker and Python installed.

elastic-search

Model used in vector_search: https://huggingface.co/sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2

A script in Python that loads strings to Elasticsearch Docker container and then allow user to search from CLI. Make sure Docker and python are installed on your machine.

Install and start the Elasticsearch Docker container

Start container without memory limit:

docker run --restart=always -d -p 9200:9200 -p 9300:9300 -e "discovery.type=single-node" elasticsearch:7.13.4

Heap size for elastic search should be half of the container memory limit, source: https://www.elastic.co/guide/en/elasticsearch/guide/current/heap-sizing.html Swap should be disabled for better performance.

Here swap is disabled, memory limit is set to 2g, and minimum and maximum heap size is set to 1g (resize should be prevented for better performance).

docker run --restart=always -d -p 9200:9200 -p 9300:9300 -e "discovery.type=single-node" --memory="2g" --memory-swap="2g" -e "ES_JAVA_OPTS=-Xms1g -Xmx1g" elasticsearch:7.13.4

Start Elasticsearch if it was stopped

docker start $(docker ps -a -q --filter "ancestor=elasticsearch:7.13.4")

Clear elastic search index

curl -X DELETE 'http://localhost:9200/_all'
curl -X DELETE 'http://localhost:9200/text_index'

telegram-gpt-bot

Creating a bot in Telegram

• Open @BotFather.
• Type /newbot and follow the instructions to create a new bot.
• @BotFather will provide a token for the new bot.

Clone the repository

git clone https://github.com/deep-foundation/russian-laws-bot

Login to the code folder

cd russian-laws-bot

Setting environment variables

After entering the code folder, you will need to set up environment variables. To do this, create a .env file. In this file you can change the values as follows:

OPENAI_API_KEY=your_openai_api_key
TELEGRAM_TOKEN=your_telegram_bot_token

Install dependencies

pip install --upgrade pip
pip install -r requirements.txt

Run locally

Without logs:

python ./main.py

With logs:

python main.py 2>&1 | tee log.txt

And if python is not available, but python3 is.

python3 main.py 2>&1 | tee log.txt

Building a Docker image

docker build -t russian-laws-bot .

Running a Docker container

docker run -p 8080:8080 --name container-russian-laws-bot russian-laws-bot