Get ib data feed hackzorz workin
ib has a throttle limit for "hft" bars but contained in here is some hackery using ``xdotool`` to reset data farms auto-magically B) This copies the working script into the ib backend mod as a routine and now uses `trio.run_process()` and calls into it from the `get_bars()` history retriever and then waits for "data re-established" events to be received from the client before making more history queries. TL;DR summary of changes: - relay ib's "system status" events (like for data farm statuses) as a new "event" msg that can be processed by registers of `Client.inline_errors()` (though we should probably make a new method for this). - add `MethodProxy.status_event()` which allows a proxy user to register for a particular "system event" (as mentioned above), which puts a `trio.Event` entry in a small table can be set by an relay task if there are any detected waiters. - start a "msg relay task" when opening the method proxy which does the event setting mentioned above in the background. - drop the request error handling around the proxy creation, doesn't seem necessary any more now that we have better error propagation from `asyncio`. - add event waiting logic around the data feed reset hackzorin. - change the order relay task to only log system events for now (though we need to do some better parsing/logic to get tws-external order updates to work again..broker_bumpz
parent
874374af06
commit
b579d4b1f5
|
@ -294,7 +294,8 @@ class Client:
|
|||
bars_kwargs = {'whatToShow': 'TRADES'}
|
||||
|
||||
global _enters
|
||||
print(f'ENTER BARS {_enters} @ end={end_dt}')
|
||||
# log.info(f'REQUESTING BARS {_enters} @ end={end_dt}')
|
||||
print(f'REQUESTING BARS {_enters} @ end={end_dt}')
|
||||
_enters += 1
|
||||
|
||||
contract = await self.find_contract(fqsn)
|
||||
|
@ -304,6 +305,7 @@ class Client:
|
|||
bars = await self.ib.reqHistoricalDataAsync(
|
||||
contract,
|
||||
endDateTime=end_dt,
|
||||
formatDate=2,
|
||||
|
||||
# time history length values format:
|
||||
# ``durationStr=integer{SPACE}unit (S|D|W|M|Y)``
|
||||
|
@ -738,25 +740,38 @@ class Client:
|
|||
def inline_errors(
|
||||
self,
|
||||
to_trio: trio.abc.SendChannel,
|
||||
) -> None:
|
||||
|
||||
# connect error msgs
|
||||
) -> None:
|
||||
'''
|
||||
Setup error relay to the provided ``trio`` mem chan such that
|
||||
trio tasks can retreive and parse ``asyncio``-side API request
|
||||
errors.
|
||||
|
||||
'''
|
||||
def push_err(
|
||||
reqId: int,
|
||||
errorCode: int,
|
||||
errorString: str,
|
||||
contract: Contract,
|
||||
|
||||
) -> None:
|
||||
|
||||
log.error(errorString)
|
||||
reason = errorString
|
||||
|
||||
if reqId == -1:
|
||||
# it's a general event?
|
||||
key = 'event'
|
||||
else:
|
||||
key = 'error'
|
||||
|
||||
try:
|
||||
to_trio.send_nowait((
|
||||
'error',
|
||||
key,
|
||||
|
||||
# error "object"
|
||||
{'reqid': reqId,
|
||||
'reason': errorString,
|
||||
'reason': reason,
|
||||
'contract': contract}
|
||||
))
|
||||
except trio.BrokenResourceError:
|
||||
|
@ -1123,9 +1138,11 @@ class MethodProxy:
|
|||
def __init__(
|
||||
self,
|
||||
chan: to_asyncio.LinkedTaskChannel,
|
||||
event_table: dict[str, trio.Event],
|
||||
|
||||
) -> None:
|
||||
self.chan = chan
|
||||
self.event_table = event_table
|
||||
|
||||
async def _run_method(
|
||||
self,
|
||||
|
@ -1140,19 +1157,44 @@ class MethodProxy:
|
|||
|
||||
'''
|
||||
chan = self.chan
|
||||
# send through method + ``kwargs: dict`` as pair
|
||||
await chan.send((meth, kwargs))
|
||||
|
||||
while not chan.closed():
|
||||
# send through method + ``kwargs: dict`` as pair
|
||||
msg = await chan.receive()
|
||||
# print(f'NEXT MSG: {msg}')
|
||||
|
||||
# TODO: py3.10 ``match:`` syntax B)
|
||||
if 'result' in msg:
|
||||
res = msg.get('result')
|
||||
if res:
|
||||
return res
|
||||
|
||||
err = msg.get('error')
|
||||
if not err:
|
||||
raise ValueError(f'Received unexpected asyncio msg {msg}')
|
||||
|
||||
elif 'exception' in msg:
|
||||
err = msg.get('exception')
|
||||
raise err
|
||||
|
||||
elif 'error' in msg:
|
||||
etype, emsg = msg
|
||||
log.warning(f'IB error relay: {emsg}')
|
||||
continue
|
||||
|
||||
else:
|
||||
log.warning(f'UNKNOWN IB MSG: {msg}')
|
||||
|
||||
def status_event(
|
||||
self,
|
||||
pattern: str,
|
||||
|
||||
) -> Union[dict[str, Any], trio.Event]:
|
||||
|
||||
ev = self.event_table.get(pattern)
|
||||
|
||||
if not ev or ev.is_set():
|
||||
# print(f'inserting new data reset event item')
|
||||
ev = self.event_table[pattern] = trio.Event()
|
||||
|
||||
return ev
|
||||
|
||||
async def wait_for_data_reset(self) -> None:
|
||||
'''
|
||||
Send hacker hot keys to ib program and wait
|
||||
|
@ -1166,6 +1208,7 @@ class MethodProxy:
|
|||
async def open_aio_client_method_relay(
|
||||
from_trio: asyncio.Queue,
|
||||
to_trio: trio.abc.SendChannel,
|
||||
event_consumers: dict[str, trio.Event],
|
||||
|
||||
) -> None:
|
||||
|
||||
|
@ -1177,7 +1220,7 @@ async def open_aio_client_method_relay(
|
|||
to_trio.send_nowait(client)
|
||||
|
||||
# TODO: separate channel for error handling?
|
||||
# client.inline_errors(to_trio)
|
||||
client.inline_errors(to_trio)
|
||||
|
||||
# relay all method requests to ``asyncio``-side client and
|
||||
# deliver back results
|
||||
|
@ -1188,8 +1231,8 @@ async def open_aio_client_method_relay(
|
|||
break
|
||||
|
||||
meth_name, kwargs = msg
|
||||
|
||||
meth = getattr(client, meth_name)
|
||||
|
||||
try:
|
||||
resp = await meth(**kwargs)
|
||||
# echo the msg back
|
||||
|
@ -1201,19 +1244,25 @@ async def open_aio_client_method_relay(
|
|||
# TODO: relay all errors to trio?
|
||||
# BaseException,
|
||||
) as err:
|
||||
to_trio.send_nowait({'error': err})
|
||||
to_trio.send_nowait({'exception': err})
|
||||
|
||||
|
||||
@acm
|
||||
async def open_client_proxy() -> MethodProxy:
|
||||
|
||||
try:
|
||||
async with to_asyncio.open_channel_from(
|
||||
# try:
|
||||
event_table = {}
|
||||
|
||||
async with (
|
||||
to_asyncio.open_channel_from(
|
||||
open_aio_client_method_relay,
|
||||
) as (first, chan):
|
||||
event_consumers=event_table,
|
||||
) as (first, chan),
|
||||
trio.open_nursery() as relay_n,
|
||||
):
|
||||
|
||||
assert isinstance(first, Client)
|
||||
proxy = MethodProxy(chan)
|
||||
proxy = MethodProxy(chan, event_table)
|
||||
|
||||
# mock all remote methods on ib ``Client``.
|
||||
for name, method in inspect.getmembers(
|
||||
|
@ -1223,33 +1272,34 @@ async def open_client_proxy() -> MethodProxy:
|
|||
continue
|
||||
setattr(proxy, name, partial(proxy._run_method, meth=name))
|
||||
|
||||
async def relay_events():
|
||||
|
||||
async with chan.subscribe() as msg_stream:
|
||||
|
||||
async for msg in msg_stream:
|
||||
if 'event' not in msg:
|
||||
continue
|
||||
|
||||
# if 'event' in msg:
|
||||
# wake up any system event waiters.
|
||||
etype, status_msg = msg
|
||||
reason = status_msg['reason']
|
||||
|
||||
ev = proxy.event_table.pop(reason, None)
|
||||
|
||||
if ev and ev.statistics().tasks_waiting:
|
||||
log.info(f'Relaying ib status message: {msg}')
|
||||
ev.set()
|
||||
|
||||
continue
|
||||
|
||||
relay_n.start_soon(relay_events)
|
||||
|
||||
yield proxy
|
||||
|
||||
# terminate asyncio side task
|
||||
await chan.send(None)
|
||||
|
||||
except (
|
||||
RequestError,
|
||||
# BaseException,
|
||||
)as err:
|
||||
code = getattr(err, 'code', None)
|
||||
if code:
|
||||
msg = err.message
|
||||
|
||||
# TODO: retreive underlying ``ib_insync`` error?
|
||||
if (
|
||||
code == 162 and (
|
||||
'HMDS query returned no data' in msg
|
||||
or 'No market data permissions for' in msg
|
||||
)
|
||||
or code == 200
|
||||
):
|
||||
# these cases should not cause a task crash
|
||||
log.warning(msg)
|
||||
|
||||
else:
|
||||
raise
|
||||
|
||||
|
||||
@acm
|
||||
async def get_client(
|
||||
|
@ -1378,6 +1428,12 @@ def normalize(
|
|||
return data
|
||||
|
||||
|
||||
_pacing: str = (
|
||||
'Historical Market Data Service error '
|
||||
'message:Historical data request pacing violation'
|
||||
)
|
||||
|
||||
|
||||
async def get_bars(
|
||||
|
||||
proxy: MethodProxy,
|
||||
|
@ -1396,14 +1452,13 @@ async def get_bars(
|
|||
|
||||
fails = 0
|
||||
bars: Optional[list] = None
|
||||
in_throttle: bool = False
|
||||
first_dt: datetime = None
|
||||
last_dt: datetime = None
|
||||
|
||||
if end_dt:
|
||||
last_dt = pendulum.from_timestamp(end_dt.timestamp())
|
||||
|
||||
for _ in range(10):
|
||||
for _ in range(2):
|
||||
try:
|
||||
bars, bars_array = await proxy.bars(
|
||||
fqsn=fqsn,
|
||||
|
@ -1449,26 +1504,43 @@ async def get_bars(
|
|||
|
||||
continue
|
||||
|
||||
else:
|
||||
log.exception(
|
||||
"Data query rate reached: Press `ctrl-alt-f`"
|
||||
"in TWS"
|
||||
elif _pacing in msg:
|
||||
|
||||
log.warning(
|
||||
'History throttle rate reached!\n'
|
||||
'Resetting farms with `ctrl-alt-f` hack\n'
|
||||
)
|
||||
# TODO: we might have to put a task lock around this
|
||||
# method..
|
||||
hist_ev = proxy.status_event(
|
||||
'HMDS data farm connection is OK:ushmds'
|
||||
)
|
||||
# live_ev = proxy.status_event(
|
||||
# # 'Market data farm connection is OK:usfuture'
|
||||
# 'Market data farm connection is OK:usfarm'
|
||||
# )
|
||||
# TODO: some kinda resp here that indicates success
|
||||
# otherwise retry?
|
||||
await data_reset_hack()
|
||||
|
||||
# TODO: should probably create some alert on screen
|
||||
# and then somehow get that to trigger an event here
|
||||
# that restarts/resumes this task?
|
||||
if not in_throttle:
|
||||
await tractor.breakpoint()
|
||||
# TODO: a while loop here if we timeout?
|
||||
for name, ev in [
|
||||
('history', hist_ev),
|
||||
# ('live', live_ev),
|
||||
]:
|
||||
with trio.move_on_after(22) as cs:
|
||||
await ev.wait()
|
||||
log.info(f"{name} DATA RESET")
|
||||
|
||||
# TODO: wait on data con reset event
|
||||
# then begin backfilling again.
|
||||
# await proxy.wait_for_data()
|
||||
if cs.cancelled_caught:
|
||||
log.warning("reset hack failed on first try?")
|
||||
# await tractor.breakpoint()
|
||||
|
||||
in_throttle = True
|
||||
fails += 1
|
||||
continue
|
||||
|
||||
else:
|
||||
raise
|
||||
|
||||
return None, None
|
||||
# else: # throttle wasn't fixed so error out immediately
|
||||
|
@ -1520,8 +1592,7 @@ async def backfill_bars(
|
|||
# on that until we have the `marketstore` daemon in place in which
|
||||
# case the shm size will be driven by user config and available sys
|
||||
# memory.
|
||||
# count: int = 120,
|
||||
count: int = 36,
|
||||
count: int = 65,
|
||||
|
||||
task_status: TaskStatus[trio.CancelScope] = trio.TASK_STATUS_IGNORED,
|
||||
|
||||
|
@ -1566,9 +1637,6 @@ async def backfill_bars(
|
|||
|
||||
out, fails = await get_bars(proxy, fqsn, end_dt=first_dt)
|
||||
|
||||
if fails is None or fails > 1:
|
||||
break
|
||||
|
||||
if out == (None, None):
|
||||
# could be trying to retreive bars over weekend
|
||||
# TODO: add logic here to handle tradable hours and
|
||||
|
@ -2222,18 +2290,26 @@ async def deliver_trade_events(
|
|||
msg = pack_position(item)
|
||||
msg.account = accounts_def.inverse[msg.account]
|
||||
|
||||
if getattr(msg, 'reqid', 0) < -1:
|
||||
elif event_name == 'event':
|
||||
|
||||
# it's a trade event generated by TWS usage.
|
||||
log.info(f"TWS triggered trade\n{pformat(msg.dict())}")
|
||||
# it's either a general system status event or an external
|
||||
# trade event?
|
||||
log.info(f"TWS system status: \n{pformat(item)}")
|
||||
|
||||
msg.reqid = 'tws-' + str(-1 * msg.reqid)
|
||||
# TODO: support this again but needs parsing at the callback
|
||||
# level...
|
||||
# reqid = item.get('reqid', 0)
|
||||
# if getattr(msg, 'reqid', 0) < -1:
|
||||
# log.info(f"TWS triggered trade\n{pformat(msg.dict())}")
|
||||
|
||||
continue
|
||||
|
||||
# msg.reqid = 'tws-' + str(-1 * reqid)
|
||||
|
||||
# mark msg as from "external system"
|
||||
# TODO: probably something better then this.. and start
|
||||
# considering multiplayer/group trades tracking
|
||||
msg.broker_details['external_src'] = 'tws'
|
||||
continue
|
||||
# msg.broker_details['external_src'] = 'tws'
|
||||
|
||||
# XXX: we always serialize to a dict for msgpack
|
||||
# translations, ideally we can move to an msgspec (or other)
|
||||
|
@ -2336,3 +2412,104 @@ async def open_symbol_search(
|
|||
|
||||
log.debug(f"sending matches: {matches.keys()}")
|
||||
await stream.send(matches)
|
||||
|
||||
|
||||
async def data_reset_hack(
|
||||
reset_type: str = 'data',
|
||||
|
||||
) -> None:
|
||||
'''
|
||||
Run key combos for resetting data feeds and yield back to caller
|
||||
when complete.
|
||||
|
||||
This is a linux-only hack around:
|
||||
|
||||
https://interactivebrokers.github.io/tws-api/historical_limitations.html#pacing_violations
|
||||
|
||||
TODOs:
|
||||
- a return type that hopefully determines if the hack was
|
||||
successful.
|
||||
- other OS support?
|
||||
- integration with ``ib-gw`` run in docker + Xorg?
|
||||
|
||||
'''
|
||||
# TODO: try out this lib instead, seems to be the most modern
|
||||
# and usess the underlying lib:
|
||||
# https://github.com/rshk/python-libxdo
|
||||
|
||||
# TODO: seems to be a few libs for python but not sure
|
||||
# if they support all the sub commands we need, order of
|
||||
# most recent commit history:
|
||||
# https://github.com/rr-/pyxdotool
|
||||
# https://github.com/ShaneHutter/pyxdotool
|
||||
# https://github.com/cphyc/pyxdotool
|
||||
|
||||
try:
|
||||
import i3ipc
|
||||
except ImportError:
|
||||
return False
|
||||
log.warning('IB data hack no-supported on ur platformz')
|
||||
|
||||
i3 = i3ipc.Connection()
|
||||
t = i3.get_tree()
|
||||
|
||||
orig_win_id = t.find_focused().window
|
||||
|
||||
# for tws
|
||||
win_names: list[str] = [
|
||||
'Interactive Brokers', # tws running in i3
|
||||
'IB Gateway', # gw running in i3
|
||||
# 'IB', # gw running in i3 (newer version?)
|
||||
]
|
||||
|
||||
combos: dict[str, str] = {
|
||||
# only required if we need a connection reset.
|
||||
'connection': ('ctrl+alt+r', 12),
|
||||
|
||||
# data feed reset.
|
||||
'data': ('ctrl+alt+f', 6)
|
||||
}
|
||||
|
||||
for name in win_names:
|
||||
results = t.find_titled(name)
|
||||
print(f'results for {name}: {results}')
|
||||
if results:
|
||||
con = results[0]
|
||||
print(f'Resetting data feed for {name}')
|
||||
win_id = str(con.window)
|
||||
w, h = con.rect.width, con.rect.height
|
||||
|
||||
# TODO: only run the reconnect (2nd) kc on a detected
|
||||
# disconnect?
|
||||
key_combo, timeout = combos[reset_type]
|
||||
# for key_combo, timeout in [
|
||||
# # only required if we need a connection reset.
|
||||
# # ('ctrl+alt+r', 12),
|
||||
# # data feed reset.
|
||||
# ('ctrl+alt+f', 6)
|
||||
# ]:
|
||||
await trio.run_process([
|
||||
'xdotool',
|
||||
'windowactivate', '--sync', win_id,
|
||||
|
||||
# move mouse to bottom left of window (where there should
|
||||
# be nothing to click).
|
||||
'mousemove_relative', '--sync', str(w-4), str(h-4),
|
||||
|
||||
# NOTE: we may need to stick a `--retry 3` in here..
|
||||
'click', '--window', win_id,
|
||||
'--repeat', '3', '1',
|
||||
|
||||
# hackzorzes
|
||||
'key', key_combo,
|
||||
# ],
|
||||
# timeout=timeout,
|
||||
])
|
||||
|
||||
# re-activate and focus original window
|
||||
await trio.run_process([
|
||||
'xdotool',
|
||||
'windowactivate', '--sync', str(orig_win_id),
|
||||
'click', '--window', str(orig_win_id), '1',
|
||||
])
|
||||
return True
|
||||
|
|
Loading…
Reference in New Issue