func_connections.py - connect_dydx()

func_private.py - place_market_order()

• Get Position Id

• Get expiraion time

• Place an order

For testing the code repetitively

func_private.py - abort_all_positions()

• Cancel all orders

• Cancel all positions

  • Get "markets" to access "tickSize" to find out correct decimals lateron

  • Get all open positions

  • Loop through each position

    • Determine the crypto (market), side, worst price, decimals (func_utils.py - format_number())

    • Place order to close: place_market_order()

    • Append the result

Trigger by FIND_COINTEGRATED in constants.py Stage 2 is only done once a day.

Build a table with all of the prices across all the cryptos

Functions:

• Get ISO Times

  We can get market candles at max. 100 on time, but we will need 300 or 400 in total. So we will make separate requests with different "timeframes" and join them together.

  get_ISO_times() in func_utils.py

• Get Historical Candles

  get_candles_historical() in func_public.py

• Construct Market Prices

  construct_market_prices() in func_public.py

• Write Cointegration Functions in func_cointegration.py

  • calculate_half_lif()

  • calculate_zscore() -> Will be used in Stage 3

  • calculate_cointegration()

• Find and Store Cointegrated Pairs

  We don't need to run "Stage 2" everytime. Instead, we just refer to a csv file.

  • store_cointegration_result() in func_cointegration.py

While True + CRON

• Create BotAgent Class

  we are going to opening potentially many positions in one go, so use Python class.

  func_bot_agent.py:

  • Attributes

  • order_dict

  • check_order_status_by_id()

    we will need to check order status before opening trades -> create check_order_status() in func_private.py

  • open_trades()

• Get Recent Candles (close prices)

  Decide when to enter and exit -> get_candles_recent() in func_public.py

• Check Open Positions

  Avoild double orders -> is_open_positions() in func_private.py

• Place and Save Trades

  open_positions() in func_entry_pairs.py

  • Load cointegrated pairs from csv

  • Use the market pairs in csv to get recent prices

  • Use the prices to calculate Z-score

  • If z-score > THRESHOLD

  • check if base or quote is already open

  • Determine all the args for placing order

  • Create Bot Agent

  • Open trades

    Call open_trades() on the instance of BotAgent

  • Save the dicts to a csv

manage_trade_exits() in func_exit_pairs.py

• Open JSON file

• Get all open positions per trading platform -> market_live

• Check all saved positions in saved cvs match order record on DYDX as well as in market_live

• Get price for series_1 and series_2

• Determine all the args

• CLOSE_AT_ZSCORE_CROSS (Trigger) -> calculate current z-score -> is_close

• is_close (Trigger) -> determine args -> place opposite order

1. Create a bot @BotFather

• Get token of the created bot (save to .env)

• Get your personal ChatID (save to .env)

• Send 3 test messages in the bot

• Send a mesage through API via the url: "https://api.telegram.org/bot{bot_token}/sendMessage?chat_id={chat_id}&text={message}"

2. Send message from Python

• pip3 install requests

• send_messa() in func_messaging.py

• AWS (Free tier: EC2 compute - 750 hrs)
  • VPC
    • Public Subnet
      • EC2
        • Code (from GitHub)
        • .env
        • venv (from txt file)
        • CRON
        • Set Inbound SSH & Outbound Requests

1. Create Security group

• Inbound Rule - SSH

2. Launch EC2 instance

• Ubuntu

• Architecture: Arm for M1 chip but x86 for free tier

• Select existing security group

3. Install Python

• Connect the instance

• Ubuntu terminal

sudo apt-get update

sudo apt-get install python3.11

sudo git clone https://...

pip3 instll -r requirements.txt

sudo apt install python3-pip

We don't set venv on EC2 for easier access later on. But it can be done.

touch .env

sudo nano .env

# Paste the variables

control + x

y

Enter

After making updates on Github,

# Since we have added .env here
git reset --hard HEAD

git pull origin --rebase

In case you want to pull a fresh new git from GitHub, you can delete the whole directory, and then pull from Github.

rm -rf <folderName>

https://crontab.guru/

crontab -e

0 12 * * * /bin/timeout -s 2 86330 python3 dydx_bot/program/main.py > output.txt  2>&1

*/5 * * * * /bin/timeout -s 2 290 python3 dydx_bot/program/main.py > output.txt  2>&1

crontab -l

• */5 * * * *: runs the job every 5 mins == 300 seconds

• /bin/timeout -s 2 290: kill the job after 290 seconds

• python3 dydx_bot/program/main.py: path to run the code (job)

• > output.txt: print output to file

• 2>&1: end