def test_source():
now = datetime.now()
# TODO : generate sample data with hypothesis
df = pandas.DataFrame(
data=[
[random.randint(-10, 10),
random.randint(-10, 10)],
[random.randint(-10, 10),
random.randint(-10, 10)],
],
columns=["random1", "random2"],
index=[now, now + timedelta(milliseconds=1)],
)
dm = DataModel(name="TestDataModel", data=df)
ds = dm.source
assert isinstance(ds, DataSource)
assert ds.name == "TestDataModel"
assert ds.column_names == df.reset_index().columns.to_list()
for c in df.columns:
assert (ds.data[c] == df[c]).all()
assert ds in dm._rendered_datasources
def test_render():
now = datetime.now()
# TODO : generate sample data with hypothesis
df = pandas.DataFrame(
data=[
[random.randint(-10, 10), random.randint(-10, 10)],
[random.randint(-10, 10), random.randint(-10, 10)],
],
columns=["random1", "random2"],
index=[now, now + timedelta(milliseconds=1)],
)
dm = DataModel(name="TestDataModel", data=df)
dv = DataView(model=dm)
assert dv.model == dm
rendered = dv.plot
assert isinstance(rendered, Plot)
assert dv.model == dm
rendered2 = dv.plot
assert isinstance(rendered, Plot)
assert rendered2 != rendered
def __init__(self, sample_size=255):
self.sample_size = sample_size
self.pregen_frame = pandas.DataFrame(
data={"RandomWalk": RandomWalk().sample(sample_size)})
self.current = 0
self.model = DataModel(data=self.pregen_frame[:self.current],
name="Random Walk",
debug=False)
def __init__(self, sample_size=255):
self.sample_size = sample_size
self.pregen_frame = pandas.DataFrame(
data={"MarkovChain": MarkovChain().sample(sample_size)}
)
self.current = 0
self.model = DataModel(
data=self.pregen_frame[: self.current], name="Markov Chain", debug=False
)
def test_data_patch():
now = datetime.now()
# TODO : generate sample data with hypothesis
df = pandas.DataFrame(
data=[
[random.randint(-10, 10),
random.randint(-10, 10)],
[random.randint(-10, 10),
random.randint(-10, 10)],
],
columns=["random1", "random2"],
index=[now, now + timedelta(milliseconds=1)],
)
dm = DataModel(name="TestDataModel", data=df)
no_patches = dm._patch(df)
assert len(no_patches) == 0
# same index, different values
df2 = pandas.DataFrame(
data=[
[random.randint(-20, -10),
random.randint(-20, -10)],
[random.randint(-20, -10),
random.randint(-20, -10)],
],
columns=["random1", "random2"],
index=[now, now + timedelta(milliseconds=1)],
)
patches = dm._patch(df2)
# Note : we need to drop timestamp index and retrieve the integer index
assert patches == {
col: [(i, v) for i, v in s.items()]
for col, s in df2.reset_index(drop=True).to_dict("series").items()
}
def test_data_stream():
now = datetime.now()
# TODO : generate sample data with hypothesis
df = pandas.DataFrame(
data=[
[random.randint(-10, 10),
random.randint(-10, 10)],
[random.randint(-10, 10),
random.randint(-10, 10)],
],
columns=["random1", "random2"],
index=[now, now + timedelta(milliseconds=1)],
)
dm = DataModel(name="TestDataModel", data=df)
no_streamable = dm._stream(df)
assert no_streamable.empty
# different indexes
df2 = pandas.DataFrame(
data=[
[random.randint(-10, 10),
random.randint(-10, 10)],
[random.randint(-10, 10),
random.randint(-10, 10)],
],
columns=["random1", "random2"],
index=[
now + timedelta(milliseconds=2), now + timedelta(milliseconds=3)
],
)
streamable = dm._stream(df2)
assert (streamable == df2).all().all()
def test_intindexed_data():
# TODO : generate sample data with hypothesis
df = pandas.DataFrame(
data=[
[random.randint(-10, 10),
random.randint(-10, 10)],
[random.randint(-10, 10),
random.randint(-10, 10)],
],
columns=["random1", "random2"],
index=[0, 1],
)
dm = DataModel(name="TestDataModel", data=df)
dd = dm.data
assert isinstance(dd, pandas.DataFrame)
assert (dd == df).all().all()
def test_datetimeindexed_data():
now = datetime.now()
# TODO : generate sample data with hypothesis
df = pandas.DataFrame(
data=[
[random.randint(-10, 10),
random.randint(-10, 10)],
[random.randint(-10, 10),
random.randint(-10, 10)],
],
columns=["random1", "random2"],
index=[now, now + timedelta(milliseconds=1)],
)
dm = DataModel(name="TestDataModel", data=df)
dd = dm.data
assert isinstance(dd, pandas.DataFrame)
assert (dd == df).all().all()
async def _internal_example():
# Minimal server test
import random
import asyncio
from datetime import datetime, timedelta
# Note : This is "created" before a document output is known
# and before a request is sent to the server
start = datetime.now()
random_data_model = DataModel(
name="random_data_model",
data=pandas.DataFrame(
data=[
[random.randint(-10, 10),
random.randint(-10, 10)],
[random.randint(-10, 10),
random.randint(-10, 10)],
],
columns=["random1", "random2"],
index=[start, start + timedelta(milliseconds=1)],
),
)
# Note : This is "created" before a document output is known and before a request is sent to the server
view = DataView(model=random_data_model)
view.plot_args(
title="Random Test",
plot_height=480,
tools="xpan, xwheel_zoom, reset",
toolbar_location="left",
y_axis_location="right",
x_axis_type="datetime",
sizing_mode="scale_width",
)
# filtering view on the left
fview = DataView(
model=random_data_model,
filter=
# only show when random2 values are positive
lambda dt: dt.random2 > 0 # TODO : convert to pandas syntax... somehow
# here (dimension -1) we are talking about attributes, not columns...
)
# Producer as a background task
async def compute_random(m, M):
tick = (random_data_model.data.index.to_list()
) # to help with full data generation
while True:
now = datetime.now()
tick.append(now)
print(now) # print in console for explicitness
new_data = {
"random1": [
random.randint(m, M) # change everything to trigger patch
for t in range(len(tick)) # + 1 extra element to stream
],
"random2":
random_data_model.data["random2"].to_list() +
[random.randint(m, M)], # only add one element to stream
}
# push FULL data updates !
# Note some derivative computation may require more than you think
random_data_model(
pandas.DataFrame(columns=["random1", "random2"],
data=new_data,
index=tick))
await asyncio.sleep(1)
# bg async task...
asyncio.create_task(compute_random(-10, 10))
return random_data_model, view, fview