Skip to content

Update 'arkouda_viz' package and av.explore() #179

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
vasslitvinov wants to merge 1 commit into
base: main
Choose a base branch
from
Draft

Update 'arkouda_viz' package and av.explore() #179

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 7 additions & 3 deletions arkouda_viz/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,14 +1,18 @@
<p align="center">
<img src="../pictures/logo.png"/>
<img src="pictures/logo.png"/>
</p>

This is a client only implementation of vizualizations using Arkouda. Thus, all code is python and uses only server elements currently included in the main arkouda repository.

## Functionality Implemented

- 'datashade()' - Takes an Arkouda DataFrame along with optional parameters and creates an interactive plot using datashader. The method then updates the plot based on the user's selections of a variety of widgets.
- 'hist()' - Plots a histogram for numeric data.

- 'crossfilter()' - Takes an Arkouda Dataframe and creates a scatterplot with the widgets of size and color that manipulates the points.
- 'area()' - Plots a histogram for numeric data as an area plot.

- 'boxWhisker()' - Plots a box plot for numeric data.

- 'explore()' - Creates an interactive 2-D histogram for numeric data.

## Usage

Expand Down
157 changes: 86 additions & 71 deletions arkouda_viz/client/arkouda_viz/viz.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,14 +2,25 @@
import holoviews as hv
import panel as pn
import param
from typing import Tuple, Union
from typing import Tuple, Union, Optional

_numeric_types = ["float64", "int64", "uint64"]

"""
The plotting engine to be used by default
unless specified explicitly.
"""


default_engine = "matplotlib"


"""
Helper method for setting up the plot rendering environment.
Parameters
----------
engine : str
The plotting engine.
engine : string
The plotting engine; 'default_engine' by default.
width : int
Width of the plot.
height : int
Expand All @@ -21,31 +32,38 @@
"""


def render_env(engine: str, width: int, height: int):
from bokeh.io import output_notebook
def render_env(engine: Optional[str], width: int, height: int):
global default_engine
if engine == None: engine = default_engine

def ensure(name):
if not hv.extension._loaded or hv.Store.current_backend != name:
hv.extension(name)

output_notebook()
if engine in ("bokeh", "b"):
hv.extension("bokeh")
ensure("bokeh")
return dict(width=width, height=height)
elif engine in ("plotly", "p"):
hv.extension("plotly")
ensure("plotly")
return dict(width=width, height=height)
elif engine in ("matplotlib", "m"):
hv.extension("matplotlib")
return dict(fig_inches=(5, 5))
ensure("matplotlib")
return dict(fig_inches=(int(width/100), int(height/100)))
else:
raise ValueError("Please provide a supported plotting engine.")


"""
Plots an area plot for numeric data.
Plots a histogram for numeric data as an area plot.
The X axis is always range(0, bins).
Parameters
----------
data : ak.DataFrame or ak.pdarray
The data to be plotted.
bins : int
Number of bins to divide the data into.
engine : string
The plotting engine.
The plotting engine; 'default_engine' by default.
width : int
Width of the plot.
height : int
Expand All @@ -59,8 +77,8 @@ def render_env(engine: str, width: int, height: int):

def area(
data: Union[ak.DataFrame, ak.pdarray] = None,
bins=10,
engine: str = "matplotlib",
bins: int = 10,
engine: Optional[str] = None,
width: int = 500,
height: int = 500,
):
Expand All @@ -70,7 +88,7 @@ def area(
numeric_columns = [
col
for col, dtype in data.dtypes.items()
if dtype in ["float64", "int64"]
if dtype in _numeric_types
]
if len(numeric_columns) == 0:
raise ValueError(
Expand All @@ -79,14 +97,13 @@ def area(
data = data[numeric_columns]
h, b = ak.histogram(data[data.columns[0]], bins=bins)
var = pn.widgets.Select(
name="variable", value=data.columns[0], options=data.columns
name="variable", value=data.columns[0], options=data.columns.to_list()
)
all = pn.widgets.Checkbox(name="all")

@pn.depends(var.param.value, all.param.value)
def create_figure(var, all):
if all:
var.disabled = True
overlay = hv.Overlay()
for column in data.columns:
h, b = ak.histogram(data[column], bins=bins)
Expand All @@ -98,7 +115,7 @@ def create_figure(var, all):

widgets = pn.WidgetBox(var, all, width=200)
return pn.Row(widgets, create_figure).servable("Area")
if isinstance(data, ak.pdarray) and data.dtype in ["int64", "float64"]:
if isinstance(data, ak.pdarray) and data.dtype in _numeric_types:
h, b = ak.histogram(data, bins=bins)
return hv.Area(h.to_ndarray()).opts(**opts)
else:
Expand All @@ -118,7 +135,7 @@ def create_figure(var, all):
bins : int
Number of bins to divide the data into.
engine : string
The plotting engine.
The plotting engine; 'default_engine' by default.
width : int
Width of the plot.
height : int
Expand All @@ -133,7 +150,7 @@ def create_figure(var, all):
def hist(
data: Union[ak.DataFrame, ak.pdarray] = None,
bins=10,
engine: str = "matplotlib",
engine: Optional[str] = None,
width: int = 500,
height: int = 500,
):
Expand All @@ -143,7 +160,7 @@ def hist(
numeric_columns = [
col
for col, dtype in data.dtypes.items()
if dtype in ["float64", "int64"]
if dtype in _numeric_types
]
if len(numeric_columns) == 0:
raise ValueError(
Expand All @@ -152,7 +169,7 @@ def hist(
data = data[numeric_columns]
h, b = ak.histogram(data[data.columns[0]], bins=bins)
var = pn.widgets.Select(
name="variable", value=data.columns[0], options=data.columns
name="variable", value=data.columns[0], options=data.columns.to_list()
)

@pn.depends(var.param.value)
Expand All @@ -161,7 +178,7 @@ def create_figure(var):

widgets = pn.WidgetBox(var, width=200)
return pn.Row(widgets, create_figure).servable("Histogram")
if isinstance(data, ak.pdarray) and data.dtype in ["int64", "float64"]:
if isinstance(data, ak.pdarray) and data.dtype in _numeric_types:
h, b = ak.histogram(data, bins=bins)
return hv.Histogram((h.to_ndarray(), b.to_ndarray())).opts(**opts)
else:
Expand All @@ -179,7 +196,7 @@ def create_figure(var):
data : ak.DataFrame or ak.pdarray
The data to be plotted.
engine : string
The plotting engine.
The plotting engine; 'default_engine' by default.
width : int
Width of the plot.
height : int
Expand All @@ -193,7 +210,7 @@ def create_figure(var):

def boxWhisker(
data: Union[ak.DataFrame, ak.pdarray] = None,
engine: str = "matplotlib",
engine: Optional[str] = None,
width: int = 5,
height: int = 5,
):
Expand All @@ -213,7 +230,7 @@ def boxWhisker(
data = data[numeric_columns]

var = pn.widgets.Select(
name="variable", value=data.columns[0], options=data.columns
name="variable", value=data.columns[0], options=data.columns.to_list()
)

@pn.depends(var.param.value)
Expand All @@ -238,14 +255,14 @@ def create_figure(var):

return boxwhisker.opts(
hv.opts.Bounds(alpha=0.5, color="blue"),
hv.opts.HLine(color="red", linewidth=2, xlim=(0, 1)),
hv.opts.HLine(color="red", line_width=2, xlim=(0, 1)),
hv.opts.Segments(color="black"),
hv.opts.Points(color="green"),
)

widgets = pn.WidgetBox(var, width=200)
return pn.Row(widgets, create_figure).servable("Box and Whisker")
if isinstance(data, ak.pdarray) and data.dtype in ["int64", "float64"]:
if isinstance(data, ak.pdarray) and data.dtype in _numeric_types:
sorted_data = ak.sort(data)

values = {
Expand All @@ -266,7 +283,7 @@ def create_figure(var):

return boxwhisker.opts(
hv.opts.Bounds(alpha=0.5, color="blue"),
hv.opts.HLine(color="red", linewidth=2, xlim=(0, 1)),
hv.opts.HLine(color="red", line_width=2, xlim=(0, 1)),
hv.opts.Segments(color="black"),
hv.opts.Points(color="green"),
)
Expand All @@ -289,7 +306,7 @@ def create_figure(var):
yBin : int
Number of bins to divide the y data into.
engine : string
The plotting engine.
The plotting engine; 'bokeh' by default.
width : int
Width of the plot.
height : int
Expand All @@ -300,18 +317,16 @@ def create_figure(var):
An image with or without a variable dropdown menu based in single or multiple columns.
"""


def explore(
data: Union[ak.DataFrame, Tuple[ak.pdarray, ak.pdarray]] = None,
xBin: int = 10,
yBin: int = 10,
engine: str = "bokeh",
xBin: int = 100,
yBin: int = 100,
engine: Optional[str] = "bokeh",
width: int = 500,
height: int = 500,
):
render_env(engine, width=width, height=height)
opts = render_env(engine, width=width, height=height)
pn.extension()
full_data = None
if data is not None:
if isinstance(data, ak.DataFrame):
numeric_columns = [
Expand All @@ -323,16 +338,16 @@ def explore(
raise ValueError(
"The provided DataFrame does not have at least two numeric columns."
)
full_data = data[numeric_columns]
elif (
isinstance(data, tuple)
and len(data) == 2
and all(isinstance(item, ak.pdarray) for item in data)
):
full_data = ak.DataFrame(data[0], data[1])
data = ak.DataFrame({"0":data[0], "1":data[1]})
numeric_columns = ["0", "1"]
else:
raise ValueError(
"Invalid data. Please provide an ak.Dataframe or [ak.pdarray, ak.pdarray]."
"Invalid data. Please provide an ak.Dataframe or (ak.pdarray, ak.pdarray)."
)
else:
raise ValueError("Please provide data.")
Expand All @@ -342,10 +357,10 @@ class Explore(param.Parameterized):
label="color map", default="turbo", objects=hv.plotting.list_cmaps()
)
x_var = param.Selector(
label="x-variable", default=data.columns[0], objects=data.columns
label="x-variable", default=numeric_columns[0], objects=numeric_columns
)
y_var = param.Selector(
label="y-variable", default=data.columns[1], objects=data.columns
label="y-variable", default=numeric_columns[1], objects=numeric_columns
)
enable_slider_checkbox = pn.widgets.Checkbox(
name="remove outliers", value=False
Expand All @@ -355,40 +370,38 @@ class Explore(param.Parameterized):
)

params = Explore()
cols = full_data.columns
initial_xrange = (ak.min(full_data[cols[0]]), ak.max(full_data[cols[0]]))
initial_yrange = (ak.min(full_data[cols[1]]), ak.max(full_data[cols[1]]))

def make_data(x_range, y_range, cmap, x_var, y_var):
if x_range is None or y_range is None or not x_range or not y_range:
binned_data = ak.histogram2d(
full_data[x_var], full_data[y_var], bins=(xBin, yBin)
)
return hv.Image(binned_data[0].to_ndarray(), bounds=(0, 0, 1, 1)).opts(
cmap=cmap, width=width, height=height, color_bar=True
)
server_widget = pn.widgets.StaticText(name="", value="", styles = {'color': 'red'})

# use caching to preserve the current range when changing the color map and to
# avoid unnecessary re-histogramming, which sometimes happens with hv.streams
class Cache: pass
h_cache = Cache()
h_cache.x_var, h_cache.y_var, h_cache.histo, h_cache.range2 = None, None, None, None

def make_data(x_range, y_range, cmap, x_var, y_var, h_cache=h_cache):
if x_var == h_cache.x_var and y_var == h_cache.y_var \
and (x_range == None or (x_range, y_range) == h_cache.range2):
histo, range2 = h_cache.histo, h_cache.range2
else:
subset_data = data[
(full_data[x_var] >= x_range[0])
& (full_data[x_var] <= x_range[1])
& (full_data[y_var] >= y_range[0])
& (full_data[y_var] <= y_range[1])
]
x_span = x_range[1] - x_range[0]
y_span = y_range[1] - y_range[0]
binned_data = ak.histogram2d(
subset_data[x_var], subset_data[y_var], bins=(1000, 1000)
)
if x_range == None: range = None
else: range = (x_range, y_range)
server_widget.value = "server processing"
histo,xbins,ybins = ak.histogram2d(data[x_var], data[y_var], bins=(xBin, yBin), range=range)
server_widget.value = ""
histo = histo.to_ndarray().T[::-1,:]
range2 = ((xbins[0], xbins[-1]), (ybins[0], ybins[-1])) # ((xmin,xmax),(ymin,ymax))
h_cache.x_var, h_cache.y_var, h_cache.histo, h_cache.range2 = x_var, y_var, histo, range2

return hv.Image(
binned_data[0].to_ndarray(),
bounds=(x_range[0], y_range[0], x_range[0] + x_span, y_range[0] + y_span),
).opts(cmap=cmap, width=width, height=height, colorbar=True)
histo,
bounds=(range2[0][0], range2[1][0], range2[0][1], range2[1][1]) # xmin, ymin, xmax, ymax
).opts(cmap=cmap, colorbar=True, **opts)

@pn.depends(
cmap=params.param.cmap, x_var=params.param.x_var, y_var=params.param.y_var
)
def update(cmap, x_var, y_var):
stream = hv.streams.RangeXY(x_range=initial_xrange, y_range=initial_yrange)
stream = hv.streams.RangeXY()
dmap = hv.DynamicMap(
lambda x_range, y_range: make_data(x_range, y_range, cmap, x_var, y_var),
streams=[stream],
Expand All @@ -400,8 +413,10 @@ def update(cmap, x_var, y_var):
params.param.cmap,
params.param.x_var,
params.param.y_var,
params.enable_slider_checkbox,
params.z_score_threshold_slider,
# these are currently unused:
#params.enable_slider_checkbox,
#params.z_score_threshold_slider,
server_widget,
width=200,
)
return pn.Row(widget_column, update)
return pn.Row(widget_column, update)
2 changes: 1 addition & 1 deletion arkouda_viz/client/setup.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@

setup(
name="arkouda_viz",
version="0.0.0",
version="0.0.1",
description="Vizualizations for Arkouda.",
long_description=long_description,
long_description_content_type="text/markdown",
Expand Down