def test_show_graph(self):
f = SomeFactor()
p = Pipeline(columns={"f": SomeFactor()})
# The real display_graph call shells out to GraphViz, which isn't a
# requirement, so patch it out for testing.
def mock_display_graph(g, format="svg", include_asset_exists=False):
return (g, format, include_asset_exists)
self.assertEqual(
getargspec(display_graph),
getargspec(mock_display_graph),
msg="Mock signature doesn't match signature for display_graph.",
)
patch_display_graph = patch(
"zipline.pipeline.graph.display_graph",
mock_display_graph,
)
with patch_display_graph:
graph, format, include_asset_exists = p.show_graph()
self.assertIs(graph.outputs["f"], f)
# '' is a sentinel used for screen if it's not supplied.
self.assertEqual(
sorted(graph.outputs.keys()),
["f", graph.screen_name],
)
self.assertEqual(format, "svg")
self.assertEqual(include_asset_exists, False)
with patch_display_graph:
graph, format, include_asset_exists = p.show_graph(format="png")
self.assertIs(graph.outputs["f"], f)
# '' is a sentinel used for screen if it's not supplied.
self.assertEqual(sorted(graph.outputs.keys()),
["f", graph.screen_name])
self.assertEqual(format, "png")
self.assertEqual(include_asset_exists, False)
with patch_display_graph:
graph, format, include_asset_exists = p.show_graph(format="jpeg")
self.assertIs(graph.outputs["f"], f)
self.assertEqual(sorted(graph.outputs.keys()),
["f", graph.screen_name])
self.assertEqual(format, "jpeg")
self.assertEqual(include_asset_exists, False)
expected = (r".*\.show_graph\(\) expected a value in "
r"\('svg', 'png', 'jpeg'\) for argument 'format', "
r"but got 'fizzbuzz' instead.")
with self.assertRaisesRegex(ValueError, expected):
p.show_graph(format="fizzbuzz")
def test_show_graph(self):
f = SomeFactor()
p = Pipeline(columns={'f': SomeFactor()})
# The real display_graph call shells out to GraphViz, which isn't a
# requirement, so patch it out for testing.
def mock_display_graph(g, format='svg', include_asset_exists=False):
return (g, format, include_asset_exists)
self.assertEqual(
inspect.getargspec(display_graph),
inspect.getargspec(mock_display_graph),
msg="Mock signature doesn't match signature for display_graph."
)
patch_display_graph = patch(
'zipline.pipeline.graph.display_graph',
mock_display_graph,
)
with patch_display_graph:
graph, format, include_asset_exists = p.show_graph()
self.assertIs(graph.outputs['f'], f)
# '' is a sentinel used for screen if it's not supplied.
self.assertEqual(sorted(graph.outputs.keys()), ['', 'f'])
self.assertEqual(format, 'svg')
self.assertEqual(include_asset_exists, False)
with patch_display_graph:
graph, format, include_asset_exists = p.show_graph(format='png')
self.assertIs(graph.outputs['f'], f)
# '' is a sentinel used for screen if it's not supplied.
self.assertEqual(sorted(graph.outputs.keys()), ['', 'f'])
self.assertEqual(format, 'png')
self.assertEqual(include_asset_exists, False)
with patch_display_graph:
graph, format, include_asset_exists = p.show_graph(format='jpeg')
self.assertIs(graph.outputs['f'], f)
# '' is a sentinel used for screen if it's not supplied.
self.assertEqual(sorted(graph.outputs.keys()), ['', 'f'])
self.assertEqual(format, 'jpeg')
self.assertEqual(include_asset_exists, False)
expected = (
r".*\.show_graph\(\) expected a value in "
r"\('svg', 'png', 'jpeg'\) for argument 'format', "
r"but got 'fizzbuzz' instead."
)
with self.assertRaisesRegexp(ValueError, expected):
p.show_graph(format='fizzbuzz')
def test_show_graph(self):
f = SomeFactor()
p = Pipeline(columns={"f": SomeFactor()})
# The real display_graph call shells out to GraphViz, which isn't a
# requirement, so patch it out for testing.
def mock_display_graph(g, format="svg", include_asset_exists=False):
return (g, format, include_asset_exists)
assert getargspec(display_graph) == getargspec(
mock_display_graph
), "Mock signature doesn't match signature for display_graph."
patch_display_graph = patch(
"zipline.pipeline.graph.display_graph",
mock_display_graph,
)
with patch_display_graph:
graph, format, include_asset_exists = p.show_graph()
assert graph.outputs["f"] is f
# '' is a sentinel used for screen if it's not supplied.
assert sorted(graph.outputs.keys()) == ["f", graph.screen_name]
assert format == "svg"
assert include_asset_exists is False
with patch_display_graph:
graph, format, include_asset_exists = p.show_graph(format="png")
assert graph.outputs["f"] is f
# '' is a sentinel used for screen if it's not supplied.
assert sorted(graph.outputs.keys()) == ["f", graph.screen_name]
assert format == "png"
assert include_asset_exists is False
with patch_display_graph:
graph, format, include_asset_exists = p.show_graph(format="jpeg")
assert graph.outputs["f"] is f
assert sorted(graph.outputs.keys()) == ["f", graph.screen_name]
assert format == "jpeg"
assert include_asset_exists is False
expected = (
r".*\.show_graph\(\) expected a value in "
r"\('svg', 'png', 'jpeg'\) for argument 'format', "
r"but got 'fizzbuzz' instead."
)
with pytest.raises(ValueError, match=expected):
p.show_graph(format="fizzbuzz")
# # ### Filters # # In general, a **Filter** is a function from an asset at a particular moment in time to a boolean value (True of False). An example of a filter is a function indicating whether a security's price is below \$5. Given a security and a specific moment in time, this evaluates to either **True** or **False**. Filters are most commonly used for selecting sets of securities to include or exclude from your stock universe. Filters are usually applied using comparison operators, such as <, <=, !=, ==, >, >=. # # Viewing the Pipeline as a Diagram # # Zipline's Pipeline class comes with the attribute `.show_graph()` that allows you to render the Pipeline as a Directed Acyclic Graph (DAG). This graph is specified using the DOT language and consequently we need a DOT graph layout program to view the rendered image. In the code below, we will use the Graphviz pakage to render the graph produced by the `.show_graph()` attribute. Graphviz is an open-source package for drawing graphs specified in DOT language scripts. # In[ ]: import graphviz # Render the pipeline as a DAG pipeline.show_graph() # Right now, our pipeline is empty and it only contains a screen. Therefore, when we rendered our `pipeline`, we only see the diagram of our `screen`: # # ```python # AverageDollarVolume(window_length = 60).top(10) # ``` # # By default, the `.AverageDollarVolume()` class uses the `USEquityPricing` dataset, containing daily trading prices and volumes, to compute the average dollar volume: # # ```python # average_dollar_volume = np.nansum(close_price * volume, axis=0) / len(close_price) # ``` # The top of the diagram reflects the fact that the `.AverageDollarVolume()` class gets its inputs (closing price and volume) from the `USEquityPricing` dataset. The bottom of the diagram shows that the output is determined by the expression `x_0 <= 10`. This expression reflects the fact that we used `.top(10)` as a filter in our `screen`. We refer to each box in the diagram as a Term.
# Also rank and zscore (don't need to de-mean by sector, s)
factor_smoothed = (SimpleMovingAverage(inputs=[factor],
window_length=5).rank().zscore())
# add the unsmoothed factor to the pipeline
p.add(factor, 'Momentum_Factor')
# add the smoothed factor to the pipeline too
p.add(factor_smoothed, 'Smoothed_Momentum_Factor')
# ## visualize the pipeline
#
# Note that if the image is difficult to read in the notebook, right-click and view the image in a separate tab.
# In[ ]:
p.show_graph(format='png')
# ## run pipeline and view the factor data
# In[ ]:
df = engine.run_pipeline(p, factor_start_date, universe_end_date)
# In[ ]:
df.head()
# ## Evaluate Factors
#
# We'll go over some tools that we can use to evaluate alpha factors. To do so, we'll use the [alphalens library](https://github.com/quantopian/alphalens)
#
