a Binance crypto trading bot which buys on news
- requests
- BeautifulSoup
- Binance API (pip install python-binance)
import requests
from bs4 import BeautifulSoup as BS
from binance.client import Client
client = Client(' ',' ')
in the first and second line, we import requests and bs4 from BeautifulSoup to be able to extract the HTML syntax of Binance news webpage.
After that, you need to connect your Binance.com account to the bot (i suggest to use a demo account, not your main trading account): client = Client(' ',' ') . inside the single/double quote you input your keys. To generate the API keys and assign relevant permissions, tap HERE
I have avoided the integration of any NLP or ML related keywords analysis methods to make the code's execution faster. Yes it may have weak points specially if Binanace changes its most used words in positive news, so if you used NLP, the analysis of news might be more precise, but it's up to you. You can kepp following Binance announcements and upgrade the `trigger` list's keywords as you wish.
Regarding the coins list, i have added the most recent added tokens and coins (before the DeFi hype). Again, it's up to you to keep this list, or reduce it to watch less pairs (like your favorite pairs where you may automate trades basing on its news)
the code below loops over all ` ` HTML elements in Binance announcement page:
for link in html.find_all('a'):
for link in html.find_all('a'):
After that, using BeautifulSoup again, we get the href or links inserted in the <a> </a> element which result in extracting the announcement title:
link_text = link.get('href')
The next step is to use a condition statement if to check if the announcement's link start with /en/support/articles/:
if str(link_text).startswith('/en/support/articles/'):
if True , we append to to_fetch list the 'clean' announcement title as follow:
- /en/support/articles/ is removed from the title by replacing it with an empty string: link_text.replace('/en/support/articles/', '')
- all title's words are added to a list then splitted on - : list(link_text.replace('/en/support/articles/', '').split('-'))
- finally the result (a list) is added to to_fetch list: to_fetch.append(list(link_text.replace('/en/support/articles/', '').split('-')))
the trick in this bot is to analyse the announcement's title instead of the whole text. Because Binance is not a news magazine or something like that (it means there are no "clickbait" titles), so because of titles clarity, we can use the title's words to analyse and predict if it's a positive or negative news.
This bot is only "Buy on news", you can work on it more to add the ability to buy (+ve news) or sell (-ve news) according the announcement. Or even that, you can splitthe positive and negative news trading into 2 bots to make the execution faster.
Now let's dive into how the analysis work. First of all, a for loop iterate over each list in the to_fetch list (which is a list of lists):
for x in to_fetch:
...
for i in x:
After getting the list (news) from to_fetch , we use a conditional statement to check if the i (each element of the news, title's words) exist in trigger list. Then we iterate over coins tickers in coins list to check if it exist in the announcement words, if True , we assign the coin ticker to coin_ticker variable.
Next, it checks if the announcemet words splitted list exist in saved_data -which is a global listed used as temporar memory as long as the .py is ruuning in your IDE- if True , it passes because it means that the news already exist. if False , it appends the news list to the saved data list and fetch pairs askPrice (which may be your entry price):
if x in saved_data:
pass
else:
saved_data.append(x)
print('[*] Possible Trade: ', ' '.join(x), f'Trade {coin_ticker}')
print(client.get_ticker(symbol=(f'{coin_ticker}BTC'))['askPrice'])
`