-
-
Save qharlie/ef43db549629ded20103d8cf853f14d3 to your computer and use it in GitHub Desktop.
These are just some sample scripts to dump the nats stream and replay it to a local NATs server.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/usr/bin/env python | |
from nats.aio.client import Client as NATS | |
import json | |
import asyncio | |
import logging | |
import argparse | |
import gzip | |
import glob | |
import time | |
import os | |
import re | |
async def main(loop, nc, datafiles): | |
await nc.connect(servers=opt.nats, loop=loop) | |
map = { | |
"sym": "symbol", | |
"a": "average", | |
"c": "close", | |
"h": "high", | |
"k": "transactions", | |
"l": "low", | |
"o": "open", | |
"t": "totalvalue", | |
"x": "exchange", | |
"v": "volume", | |
"s": "start", | |
"e": "end", | |
"vw": "vwap", | |
"av": "totalvolume", | |
"op": | |
"dailyopen", # depricated? stream often has 0 for op | |
} | |
needed_keys = {'o', 'h', 'l', 'c', 'v'} | |
pam = {v: k for k, v in map.items()} | |
prevts = None | |
for file in datafiles: | |
print(f"Streaming contents of {file}...") | |
try: | |
with gzip.GzipFile(file, 'rb') as feed: | |
for line in feed: | |
event = json.loads(line) | |
if (not opt.all) and 'A.' in event['channel']: | |
continue | |
if (opt.match) and not opt.match.search(event['channel']): | |
continue | |
# print(event) | |
ts = int(event['ts']) | |
if needed_keys.intersection(list(event['data'].keys())): | |
data = event['data'] | |
else: | |
data = ({ | |
pam[k]: v | |
for k, v in event['data'].items() if k in pam | |
}) | |
print( | |
f'{time.strftime("%m/%d %H:%M:%S", time.localtime(ts/1000))} {event["channel"]}' | |
f' {data["o"]} {data["h"]} {data["l"]} {data["c"]} {data["v"]}' | |
f' {time.strftime("%m/%d %H:%M:%S", time.localtime(data['s']/1000))} - ' | |
f' {time.strftime("%m/%d %H:%M:%S", time.localtime(data['e']/1000))}' | |
) | |
if opt.speedup > 0 and prevts and prevts < ts: | |
sleep = ((ts - prevts) / 100) * opt.speedup | |
if sleep > opt.maxsleep: | |
sleep = opt.maxsleep | |
print(f' - sleep {sleep:0.2f}') | |
await asyncio.sleep(sleep, loop=loop) | |
# print(data) | |
await nc.publish(event['channel'], | |
json.dumps(data).encode()) | |
await nc.flush() | |
# msg = await asyncio.wait_for(future, 1) | |
prevts = ts | |
except Exception as e: | |
print(e) | |
await nc.close() | |
if __name__ == '__main__': | |
nc = NATS() | |
logging.basicConfig() | |
parser = argparse.ArgumentParser() | |
parser.add_argument( | |
"--datadir", | |
help="Specify a directory containing *.json.gz files to replay", | |
type=str, | |
nargs='*', | |
default='data/') | |
parser.add_argument( | |
"--datafiles", | |
help="Specify the file(s) to replay", | |
type=str, | |
nargs='*') | |
parser.add_argument( | |
"--speedup", | |
help="Change the time multiplier to speedup (or slow down >1) the event flow", | |
default=0.002, | |
type=float) | |
parser.add_argument( | |
"--maxsleep", | |
help="Max time to sleep between events", | |
default=2.0, | |
type=float) | |
parser.add_argument( | |
"--nosleep", "--fast", | |
help="Stream the data as fast as possible (no speedup multiplier used)", | |
const=0, | |
dest='speedup', | |
nargs='?', | |
type=float) | |
parser.add_argument( | |
"--match", | |
help="Include events wtih channel matching supplied regular expression", | |
type=re.compile) | |
parser.add_argument( | |
"--nats", | |
help="Specify the destination NATS server(s)", | |
type=str, | |
nargs='*', | |
default=["nats://192.168.0.100:4222"]) | |
parser.add_argument( | |
"--all", | |
help="Include A.* events as well as AM.*", | |
action='store_true') | |
opt = parser.parse_args() | |
if not opt.datafiles: | |
opt.datafiles = glob.glob(os.path.join(opt.datadir, '*.json.gz')) | |
future = asyncio.Future() | |
loop = asyncio.get_event_loop() | |
loop.run_until_complete(main(loop, nc, opt.datafiles)) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/usr/bin/env python | |
from alpaca_trade_api.stream2 import StreamConn | |
import time | |
import json | |
conn = StreamConn() | |
map = { | |
"sym": "symbol", | |
"a": "average", | |
"c": "close", | |
"h": "high", | |
"k": "transactions", | |
"l": "low", | |
"o": "open", | |
"t": "totalvalue", | |
"x": "exchange", | |
"v": "volume", | |
"s": "start", | |
"e": "end", | |
"vw": "vwap", | |
"av": "totalvolume", | |
"op": "dailyopen", | |
} | |
reverse_pam = {v: k for k,v in map.items()} | |
# revert the mapping done in polygon/stream, we could avoid using that module | |
# but it does the heavy asyncio lifting to make this a simple script. | |
def reverse_map(data): | |
data = ({reverse_pam[k]: v for k, v in data.items() if k in reverse_pam}) | |
return data | |
@conn.on(r'.*') | |
async def on_data(conn, channel, data): | |
print(json.dumps({'ts': round(time.time_ns()/1000000), 'channel': channel, 'data': reverse_map(data._raw)})) | |
if __name__ == '__main__': | |
conn.run(['AM.*', 'C.*']) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment