commit 08db196ee039ec6845bc84e15eee31806945bec9 · edavis.dev/bsky-tools

+46
firehose.py
···

       1
       1
       +
       #!/usr/bin/env python3

     

       2
       2
       +
       

     

       3
       3
       +
       import asyncio

     

       4
       4
       +
       import dag_cbor

     

       5
       5
       +
       import redis

     

       6
       6
       +
       import sys

     

       7
       7
       +
       import websockets

     

       8
       8
       +
       from datetime import datetime, timezone

     

       9
       9
       +
       from io import BytesIO

     

       10
       10
       +
       

     

       11
       11
       +
       async def main():

     

       12
       12
       +
           redis_cnx = redis.Redis()

     

       13
       13
       +
           relay_url = 'wss://bsky.network/xrpc/com.atproto.sync.subscribeRepos'

     

       14
       14
       +
           firehose_seq = redis_cnx.get('bsky-tools:firehose:subscribe-repos:seq')

     

       15
       15
       +
           if firehose_seq:

     

       16
       16
       +
               relay_url += f'?cursor={firehose_seq.decode()}'

     

       17
       17
       +
       

     

       18
       18
       +
           sys.stdout.write(f'opening websocket to {relay_url}\n')

     

       19
       19
       +
           sys.stdout.flush()

     

       20
       20
       +
       

     

       21
       21
       +
           async with websockets.connect(relay_url, ping_timeout=None) as firehose:

     

       22
       22
       +
               current_minute = None

     

       23
       23
       +
               while True:

     

       24
       24
       +
                   message = BytesIO(await firehose.recv())

     

       25
       25
       +
                   header = dag_cbor.decode(message, allow_concat=True)

     

       26
       26
       +
                   if header['op'] != 1 or header['t'] == '#info':

     

       27
       27
       +
                       continue

     

       28
       28
       +
       

     

       29
       29
       +
                   redis_cnx.publish('bsky-tools:firehose:stream', message.getvalue())

     

       30
       30
       +
       

     

       31
       31
       +
                   # checkpoint the seq

     

       32
       32
       +
                   now = datetime.now(timezone.utc)

     

       33
       33
       +
                   if now.time().minute != current_minute:

     

       34
       34
       +
                       current_minute = now.time().minute

     

       35
       35
       +
       

     

       36
       36
       +
                       payload = dag_cbor.decode(message)

     

       37
       37
       +
                       payload_seq = payload['seq']

     

       38
       38
       +
                       payload_time = datetime.strptime(payload['time'], '%Y-%m-%dT%H:%M:%S.%fZ').replace(tzinfo=timezone.utc)

     

       39
       39
       +
                       payload_lag = now - payload_time

     

       40
       40
       +
       

     

       41
       41
       +
                       redis_cnx.set('bsky-tools:firehose:subscribe-repos:seq', payload_seq)

     

       42
       42
       +
                       sys.stdout.write(f'seq: {payload_seq}, lag: {payload_lag.total_seconds()}\n')

     

       43
       43
       +
                       sys.stdout.flush()

     

       44
       44
       +
       

     

       45
       45
       +
       if __name__ == '__main__':

     

       46
       46
       +
           asyncio.run(main())