A way to pull data from the Chainlink Price Feeds for analytics, algorithmic trading models, or else.
This repo uses either an RPC_URL or the Chainlink Subgraph
Install:
pip install chainlink_feeds
When you don't specify an RPC_URL, you automatically use the Chainlink subgraph.
from chainlink_feeds.chainlink_feeds import ChainlinkFeeds
cf = ChainlinkFeeds()
print(cf.get_latest_round_data(pair='ETH_USD'))
Result:
[{'assetPair': 'ETH/USD', 'blockHash': '0x141ad3c7468f4263d8b1b98a73f804b40ef1eb3a966bc2151646a08ba9872a58', 'blockNumber': '10887253', 'id': '0xf79d6afbb6da890132f9d7c355e3015f15f3406f/10887253/8', 'price': '38281000000', 'timestamp': '1600446952', 'transactionHash': '0x44e321f415e2ae236e3fbfb0df024825ff95331dca89dd25401303f0433fdb9d'}]
You can also pass:
cf.get_historical_price()
cf.get_price_feeds()
cf.get_prices()
cf.get_hourly_candle()
cf.get_daily_candle()
cf.get_weekly_candle()
This will get you all the data the subgraph can return. If you'd like to get pandas, you can just change the output format of the object.
cf = ChainlinkFeeds(output_format = 'pandas')
data = cf.get_daily_candle(pair='eth/usd')
data['closePrice'] = data['closePrice'].astype(float)
data.index = pd.to_datetime(data.index, unit='s')
data['closePrice'].plot()
resulting in:
assetPair averagePrice closePrice highPrice lowPrice medianPrice openPrice
openTimestamp
1600387200 ETH/USD 38615230654 38281000000 39190413319 38110269640 38933749501 38933749501
1600300800 ETH/USD 37891000000 39020000000 39344406296 36503000000 36503000000 36503000000
1600214400 ETH/USD 36564000000 36555000000 37276742411 35743000000 36428000000 36428000000
1600128000 ETH/USD 36983000000 36385393883 38069814258 36319641931 37744000000 37744000000
1600041600 ETH/USD 36914207489 37675318623 38270000000 35817000000 36634540717 36634540717
... ... ... ... ... ... ... ...
1586649600 ETH/USD 16022887240 15994875000 16467921975 15597479550 15864014512 15864014512
1586563200 ETH/USD 15838000000 15738831997 16108625000 15543870740 15813418305 15813418305
1586476800 ETH/USD 15868349654 15729441133 16995863008 15328500000 16994301012 16994301012
1586390400 ETH/USD 16990152130 17075750000 17151190079 16863999644 16943041228 16943041228
1586304000 ETH/USD 17277450182 17276241069 17277853301 17276241069 17277853301 17277853301
You can then run some analytics on it:
from chainlink_feeds.chainlink_feeds import ChainlinkFeeds
import matplotlib.pyplot as plt
import pandas as pd
cf = ChainlinkFeeds(output_format = 'pandas')
data = cf.get_daily_candle(pair='eth/usd')
data['closePrice'] = data['closePrice'].astype(float)
data.index = pd.to_datetime(data.index, unit='s')
data['closePrice'].plot()
Which results in:
You can also run your own GraphQL Query with:
cf.graphql_query("{enter_query_here}")
You can check out some methods on the Chainlink subgraph site.
Otherwise you can specify an RPC_URL and query the blockchain yourself.
from chainlink_feeds.chainlink_feeds import ChainlinkFeeds
cf = ChainlinkFeeds(rpc_url = "https://www.infura.com/asdfasdfasdfas)
And query the blockchain directly. You do need to know the address and the ABI of the pair. For reference, you can check the config folder in this repo. You can also load_config with your own config with addresses or ABIs. A lot of them are prepopulated in this repo, but be sure to check if you're using outdated addresses.
print(cf.get_latest_round_data(network='KOVAN', pair='ETH_USD'))
Resulting in:
{'round_id': 18446744073709562669, 'price': 382.66, 'started_at': '2020-09-18 13:30:12', 'time_stamp': '2020-09-18 13:30:12', 'answered_in_round': 18446744073709562669}
You can use the methods directly from the Chainlink Price Feeds Documentation.
- Add more query support
- Make docs other than this readme
- add more tests
- figure out how to go back more than 1000 results
Be sure to check out the Chainlink Developers Discord as well!
Consider starring this repo if you enjoyed it :)