"""Utilities."""
import asyncio
import logging
import math
import signal
import sys
import time
from dataclasses import fields, is_dataclass
from datetime import date, datetime, time as time_, timedelta, timezone
from typing import AsyncIterator, Awaitable, Callable, Iterator, List, Union
import eventkit as ev
globalErrorEvent = ev.Event()
"""
Event to emit global exceptions.
"""
EPOCH = datetime(1970, 1, 1, tzinfo=timezone.utc)
UNSET_INTEGER = 2 ** 31 - 1
UNSET_DOUBLE = sys.float_info.max
[docs]def df(objs, labels: List[str] = None):
"""
Create pandas DataFrame from the sequence of same-type objects.
Args:
labels: If supplied, retain only the given labels and drop the rest.
"""
import pandas as pd
from .objects import DynamicObject
if objs:
objs = list(objs)
obj = objs[0]
if is_dataclass(obj):
df = pd.DataFrame.from_records(dataclassAsTuple(o) for o in objs)
df.columns = [field.name for field in fields(obj)]
elif isinstance(obj, DynamicObject):
df = pd.DataFrame.from_records(o.__dict__ for o in objs)
else:
df = pd.DataFrame.from_records(objs)
if isinstance(obj, tuple):
_fields = getattr(obj, '_fields', None)
if _fields:
# assume it's a namedtuple
df.columns = _fields
else:
df = None
if labels:
exclude = [label for label in df if label not in labels]
df = df.drop(exclude, axis=1)
return df
[docs]def dataclassAsDict(obj) -> dict:
"""
Return dataclass values as ``dict``.
This is a non-recursive variant of ``dataclasses.asdict``.
"""
if not is_dataclass(obj):
raise TypeError(f'Object {obj} is not a dataclass')
return {field.name: getattr(obj, field.name) for field in fields(obj)}
[docs]def dataclassAsTuple(obj) -> tuple:
"""
Return dataclass values as ``tuple``.
This is a non-recursive variant of ``dataclasses.astuple``.
"""
if not is_dataclass(obj):
raise TypeError(f'Object {obj} is not a dataclass')
return tuple(getattr(obj, field.name) for field in fields(obj))
[docs]def dataclassNonDefaults(obj) -> dict:
"""
For a ``dataclass`` instance get the fields that are different from the
default values and return as ``dict``.
"""
if not is_dataclass(obj):
raise TypeError(f'Object {obj} is not a dataclass')
values = [getattr(obj, field.name) for field in fields(obj)]
return {
field.name: value for field, value in zip(fields(obj), values)
if value != field.default
and value == value
and not (isinstance(value, list) and value == [])}
[docs]def dataclassUpdate(obj, *srcObjs, **kwargs) -> object:
"""
Update fields of the given ``dataclass`` object from zero or more
``dataclass`` source objects and/or from keyword arguments.
"""
if not is_dataclass(obj):
raise TypeError(f'Object {obj} is not a dataclass')
for srcObj in srcObjs:
obj.__dict__.update(dataclassAsDict(srcObj))
obj.__dict__.update(**kwargs)
return obj
[docs]def dataclassRepr(obj) -> str:
"""
Provide a culled representation of the given ``dataclass`` instance,
showing only the fields with a non-default value.
"""
attrs = dataclassNonDefaults(obj)
clsName = obj.__class__.__qualname__
kwargs = ', '.join(f'{k}={v!r}' for k, v in attrs.items())
return f'{clsName}({kwargs})'
[docs]def isnamedtupleinstance(x):
"""From https://stackoverflow.com/a/2166841/6067848"""
t = type(x)
b = t.__bases__
if len(b) != 1 or b[0] != tuple:
return False
f = getattr(t, '_fields', None)
if not isinstance(f, tuple):
return False
return all(type(n) == str for n in f)
[docs]def tree(obj):
"""
Convert object to a tree of lists, dicts and simple values.
The result can be serialized to JSON.
"""
if isinstance(obj, (bool, int, float, str, bytes)):
return obj
elif isinstance(obj, (date, time_)):
return obj.isoformat()
elif isinstance(obj, dict):
return {k: tree(v) for k, v in obj.items()}
elif isnamedtupleinstance(obj):
return {f: tree(getattr(obj, f)) for f in obj._fields}
elif isinstance(obj, (list, tuple, set)):
return [tree(i) for i in obj]
elif is_dataclass(obj):
return {obj.__class__.__qualname__: tree(dataclassNonDefaults(obj))}
else:
return str(obj)
[docs]def barplot(bars, title='', upColor='blue', downColor='red'):
"""
Create candlestick plot for the given bars. The bars can be given as
a DataFrame or as a list of bar objects.
"""
import pandas as pd
import matplotlib.pyplot as plt
from matplotlib.lines import Line2D
from matplotlib.patches import Rectangle
if isinstance(bars, pd.DataFrame):
ohlcTups = [
tuple(v) for v in bars[['open', 'high', 'low', 'close']].values]
elif bars and hasattr(bars[0], 'open_'):
ohlcTups = [(b.open_, b.high, b.low, b.close) for b in bars]
else:
ohlcTups = [(b.open, b.high, b.low, b.close) for b in bars]
fig, ax = plt.subplots()
ax.set_title(title)
ax.grid(True)
fig.set_size_inches(10, 6)
for n, (open_, high, low, close) in enumerate(ohlcTups):
if close >= open_:
color = upColor
bodyHi, bodyLo = close, open_
else:
color = downColor
bodyHi, bodyLo = open_, close
line = Line2D(
xdata=(n, n),
ydata=(low, bodyLo),
color=color,
linewidth=1)
ax.add_line(line)
line = Line2D(
xdata=(n, n),
ydata=(high, bodyHi),
color=color,
linewidth=1)
ax.add_line(line)
rect = Rectangle(
xy=(n - 0.3, bodyLo),
width=0.6,
height=bodyHi - bodyLo,
edgecolor=color,
facecolor=color,
alpha=0.4,
antialiased=True
)
ax.add_patch(rect)
ax.autoscale_view()
return fig
[docs]def allowCtrlC():
"""Allow Control-C to end program."""
signal.signal(signal.SIGINT, signal.SIG_DFL)
[docs]def logToFile(path, level=logging.INFO):
"""Create a log handler that logs to the given file."""
logger = logging.getLogger()
logger.setLevel(level)
formatter = logging.Formatter(
'%(asctime)s %(name)s %(levelname)s %(message)s')
handler = logging.FileHandler(path)
handler.setFormatter(formatter)
logger.addHandler(handler)
[docs]def logToConsole(level=logging.INFO):
"""Create a log handler that logs to the console."""
logger = logging.getLogger()
logger.setLevel(level)
formatter = logging.Formatter(
'%(asctime)s %(name)s %(levelname)s %(message)s')
handler = logging.StreamHandler()
handler.setFormatter(formatter)
logger.handlers = [
h for h in logger.handlers
if type(h) is not logging.StreamHandler]
logger.addHandler(handler)
[docs]def isNan(x: float) -> bool:
"""Not a number test."""
return x != x
[docs]class timeit:
"""Context manager for timing."""
def __init__(self, title='Run'):
self.title = title
def __enter__(self):
self.t0 = time.time()
def __exit__(self, *_args):
print(self.title + ' took ' + formatSI(time.time() - self.t0) + 's')
[docs]def run(*awaitables: Awaitable, timeout: float = None):
"""
By default run the event loop forever.
When awaitables (like Tasks, Futures or coroutines) are given then
run the event loop until each has completed and return their results.
An optional timeout (in seconds) can be given that will raise
asyncio.TimeoutError if the awaitables are not ready within the
timeout period.
"""
loop = asyncio.get_event_loop()
if not awaitables:
if loop.is_running():
return
loop.run_forever()
f = asyncio.gather(*asyncio.Task.all_tasks())
f.cancel()
result = None
try:
loop.run_until_complete(f)
except asyncio.CancelledError:
pass
else:
if len(awaitables) == 1:
future = awaitables[0]
else:
future = asyncio.gather(*awaitables)
if timeout:
future = asyncio.wait_for(future, timeout)
task = asyncio.ensure_future(future)
def onError(_):
task.cancel()
globalErrorEvent.connect(onError)
try:
result = loop.run_until_complete(task)
except asyncio.CancelledError as e:
raise globalErrorEvent.value() or e
finally:
globalErrorEvent.disconnect(onError)
return result
def _fillDate(time: Union[time_, datetime]) -> datetime:
# use today if date is absent
if isinstance(time, time_):
dt = datetime.combine(date.today(), time)
else:
dt = time
return dt
[docs]def schedule(
time: Union[time_, datetime], callback: Callable, *args):
"""
Schedule the callback to be run at the given time with
the given arguments.
This will return the Event Handle.
Args:
time: Time to run callback. If given as :py:class:`datetime.time`
then use today as date.
callback: Callable scheduled to run.
args: Arguments for to call callback with.
"""
dt = _fillDate(time)
now = datetime.now(dt.tzinfo)
delay = (dt - now).total_seconds()
loop = asyncio.get_event_loop()
return loop.call_later(delay, callback, *args)
[docs]def sleep(secs: float = 0.02) -> bool:
"""
Wait for the given amount of seconds while everything still keeps
processing in the background. Never use time.sleep().
Args:
secs (float): Time in seconds to wait.
"""
run(asyncio.sleep(secs))
return True
[docs]def timeRange(
start: Union[time_, datetime],
end: Union[time_, datetime],
step: float) -> Iterator[datetime]:
"""
Iterator that waits periodically until certain time points are
reached while yielding those time points.
Args:
start: Start time, can be specified as datetime.datetime,
or as datetime.time in which case today is used as the date
end: End time, can be specified as datetime.datetime,
or as datetime.time in which case today is used as the date
step (float): The number of seconds of each period
"""
assert step > 0
delta = timedelta(seconds=step)
t = _fillDate(start)
tz = timezone.utc if t.tzinfo else None
now = datetime.now(tz)
while t < now:
t += delta
while t <= _fillDate(end):
waitUntil(t)
yield t
t += delta
[docs]def waitUntil(t: Union[time_, datetime]) -> bool:
"""
Wait until the given time t is reached.
Args:
t: The time t can be specified as datetime.datetime,
or as datetime.time in which case today is used as the date.
"""
now = datetime.now(t.tzinfo)
secs = (_fillDate(t) - now).total_seconds()
run(asyncio.sleep(secs))
return True
[docs]async def timeRangeAsync(
start: Union[time_, datetime],
end: Union[time_, datetime],
step: float) -> AsyncIterator[datetime]:
"""Async version of :meth:`timeRange`."""
assert step > 0
delta = timedelta(seconds=step)
t = _fillDate(start)
tz = timezone.utc if t.tzinfo else None
now = datetime.now(tz)
while t < now:
t += delta
while t <= _fillDate(end):
await waitUntilAsync(t)
yield t
t += delta
[docs]async def waitUntilAsync(t: Union[time_, datetime]) -> bool:
"""Async version of :meth:`waitUntil`."""
now = datetime.now(t.tzinfo)
secs = (_fillDate(t) - now).total_seconds()
await asyncio.sleep(secs)
return True
[docs]def patchAsyncio():
"""Patch asyncio to allow nested event loops."""
import nest_asyncio
nest_asyncio.apply()
[docs]def startLoop():
"""
Use nested asyncio event loop for Jupyter notebooks.
This is not needed anymore in Jupyter versions 5 or higher.
"""
def _ipython_loop_asyncio(kernel):
"""Use asyncio event loop for the given IPython kernel."""
loop = asyncio.get_event_loop()
def kernel_handler():
kernel.do_one_iteration()
loop.call_later(kernel._poll_interval, kernel_handler)
loop.call_soon(kernel_handler)
try:
if not loop.is_running():
loop.run_forever()
finally:
if not loop.is_running():
loop.run_until_complete(loop.shutdown_asyncgens())
loop.close()
patchAsyncio()
loop = asyncio.get_event_loop()
if not loop.is_running():
from ipykernel.eventloops import register_integration, enable_gui
register_integration('asyncio')(_ipython_loop_asyncio)
enable_gui('asyncio')
[docs]def useQt(qtLib: str = 'PyQt5', period: float = 0.01):
"""
Run combined Qt5/asyncio event loop.
Args:
qtLib: Name of Qt library to use, can be 'PyQt5' or 'PySide2'.
period: Period in seconds to poll Qt.
"""
def qt_step():
loop.call_later(period, qt_step)
if not stack:
qloop = QEventLoop()
timer = QTimer()
timer.timeout.connect(qloop.quit)
stack.append((qloop, timer))
qloop, timer = stack.pop()
timer.start(0)
qloop.exec_()
timer.stop()
stack.append((qloop, timer))
if qtLib not in ('PyQt5', 'PySide2'):
raise RuntimeError(f'Unknown Qt library: {qtLib}')
if qtLib == 'PyQt5':
from PyQt5.Qt import QApplication, QTimer, QEventLoop
else:
from PySide2.QtWidgets import QApplication
from PySide2.QtCore import QTimer, QEventLoop
global qApp
qApp = QApplication.instance() or QApplication(sys.argv) # type: ignore
loop = asyncio.get_event_loop()
stack: list = []
qt_step()
[docs]def parseIBDatetime(s: str) -> Union[date, datetime]:
"""Parse string in IB date or datetime format to datetime."""
if len(s) == 8:
# YYYYmmdd
y = int(s[0:4])
m = int(s[4:6])
d = int(s[6:8])
dt = date(y, m, d)
elif s.isdigit():
dt = datetime.fromtimestamp(int(s), timezone.utc)
else:
# YYYYmmdd HH:MM:SS
# or
# YYYY-mm-dd HH:MM:SS.0
ss = s.replace(' ', '').replace('-', '')[:16]
dt = datetime.strptime(ss, '%Y%m%d%H:%M:%S')
return dt