In Pandas missing data seems to be represented as numpy.NaN. We need to make sure that no element is bound to an NaN by accident!

Hey! We can remove any code formatting (like all those ugly tabs) from the code generation and just run some jslint on the generated code to make it human pretty. Has the added benefit of testing the outgoing code for errors...

A 2D contour plot would be nice!

The sample code in the readme works as it is, but if I change line 17 to:

fig += d3py.geoms.Bar(x="time", y="temp",fill="red")

It fails to render in firefox or chrome. However, all of the requests (js, json, html) return 200 except for the favicon.ico.

Hello,

There is a method stub called _clanup() in the Figure class which simply raises NotImplementedError. What is the purpose of it?

-jk

python d3py_bar.py
Cleanup after exception: <type 'exceptions.AttributeError'>: 'module' object has no attribute 'xAxis'
Cleaning temp files
Traceback (most recent call last):
File "d3py_bar.py", line 12, in
p += d3py.xAxis(x = "apple_type")
AttributeError: 'module' object has no attribute 'xAxis'
Cleaning temp files
Exception AttributeError: "'Bar' object has no attribute 'cleanup'" in <bound method Bar.del of <d3py.geoms.Bar object at 0x272f650>> ignored

great library. putting in tooltips for scatterplots would be awesome

As the title says... this should help with ipython compatibility and would vastly simplify cleanup. This could be done with simple modifications to the figure object and a new HTTPServer object (HTTPFileStreamServer?).

I tried plotting something with x-axis data of type long and it gave me the following error on line 164 of

TypeError: 0 is not JSON serializable

The line that threw the error is: https://github.com/mikedewar/D3py/blob/master/d3py/d3py.py#L164

It'd be nice to let this work on non-localhosts (like ec2 instances). Right now d3py/d3py_template.html has localhost hard coded, ie:

<script type="text/javascript" charset="utf-8" src="localhost:{{ port }}/static/d3.js"></script>',</code>

It would be nice if this was templated to

<script type="text/javascript" charset="utf-8" src="{{ host }}:{{ port }}/static/d3.js"></script>',

...And anywhere else that change would have to be made.

@bartdegoede's time series scales need to be converted to the new axis generators

right now you need to import figure. This should be import d3py

The Figure class uses cStringIO.StringIO(). This limits the html and js output to ASCII. Using StringIO.StringIO() would allow use of Unicode. Should Figure be converted to allow for unicode.

-jk

For easy of reference, here (http://docs.python.org/2/library/stringio.html) is the link to python docs.

We should have multiple functions available to add to from the geometries in the figure object. Two that are required are an init function and a draw function. New functions should handled as well.

In my opinion, this should be done with a dictionary of javascript functions which are merged under addition before rendering. Merging javascript function objects appends them to another so we should be fine.

Also, is there a way to define a scope within a javascript block (similar to wrapping code in {}'s in C code)? This would help with variable munging when we concatenate functions.

please add LICENSE file.

... otherwise port 8000 remains in use and no new figs can be launched.

Subplots would be nice!

The main figure should be able to handle an arbitrary number of data files. This will be very important when implementing maps since we also need to provide the map paths to the client.

(Problem exposed while dealing with pull #25)

del is not what we want to cleanup files: the reference count could be at zero even if we have stale files lying around!

One solution would be to use enter and exit and require the user to create a figure object with the with statement:

with Figure() as fig:
...do stuff with fig...

ideas?

We need to install the static files somewhere where we can later retrieve them (/usr/share/python-d3py?)

d3 has great support for javascript's Date objects. We should be able to pass in datetime objects and have them converted...

It would be cool to draw a ven diagram from python sets

autocall = True doesn't seem to do whatever it's supposed to do!

suggested idiom:

df = pandas.DataFrame(
    {
        "count" : [1,4,7,3,2,9],
    },
    index = ["a", "b", "c", "d", "e", "f"]
)

with d3py.Figure(df) as p:
    p += d3py.Bar(y = "count")
    p += d3py.xAxis()
    p.show()

this would draw the bar chart with the index as the x-bins.

At the moment, you need to manually run python server.py before doing any plotting. There's no real reason that the user should have to do this - we should start a server when the user first plots something and keep it going until some appropriate time. I don't know what that time should be...

Well.... the title explains it. I would update it, but you know... I have some nebulous "other stuff" to do.

The new method should look like

data = DataFrame({'y' = [1,2,3], 'x' = [1,2,3]})
p = Figure(data)
p += Line(x='x', y='y')
p += DataFrame({'z' = [3,4,5], 'x' = [2,3,4]})
p += Line(x='x', y='z')

and the internal data frame should have merged the xs and added missing data for the ys and zs that don't have data points.

Hello Mike,

I was looking at the d3py python module. At the moment, it is tightly coupled with python’s SimpleHTTPServer. From a desktop, this works well interactively.

Have you considered a Model/View split? This would provide a good way to deploy d3. Figure and its subclasses would make fine models, but all code which is specific to a web server would be refactored into a class which knows about the Simple HTTP Server. Another class would know how to write to the file system. It might look something like:

// encapsulates all code related to the SimpleHTTPServer implementation.
vHttp = createHTTP(“www.example.com”, 8080)
// creates a class which knows one to write the d3 code to the filesystem for deployment.
vFs = createFS(“~/workspace/opportunity/static”)

// This call is just a placeholder. It will take fewer args as web specific args
// would be passed to creatHTTP().
fig = PandasFigure(...)

// The user can show it. The functionality will be the as it is now.
vHttp.show(fig)
// and/or they can save it to a file system which could be handed off to deploy on the website.
vFs.save(fig)

The added benefit of this pattern is it can extended. This may be particularly useful for HTTP as there are a plethora of web servers and web technologies. New classes could be created for integration and/or deployment.

What do you think?

Regards,
John

When I ctrl-c out of a show when using 'with' python pukes a big long error, even though we direct the user to ctrl-c out of the blocking loop. Let's at least make the error not happen (and let's think about re-writing this at some point eh?).

There should be the facility to plot multiple line graphs and stacked bar charts

Allow the user to specify x and y limits manually.

The user should be able to specify that the x, y or both axes should be log-scale

Hello,
thanks for this great module!
I would like to know the proper technique to plot multiple line on a PandasFigure?
What id the proper Data Frame and how to call it?
for example:
i would like to plot two lines on the same figure defined by x,y.
How can i tel d3py to use a dataframe like that:
x y
0 [1, 2] [10, 11]
1 [3, 4] [13, 14]
thanks

Found a small issue in trying to run the example session in the README file. Adding the following line at the top corrects the example:

import numpy as np

A box plot would be nice!

right now

JavaScript() + draw

casts draw to a Javascript object, which isn't terribly clear.

python test.py
you can find your chart at http://localhost:8000/basic_example/basic_example.html
Shutting down httpd
Cleaning temp files

The browser window opens but the server is already shut down.

Source code:

import d3py
import pandas
import numpy as np
# some test data
T = 100
# this is a data frame with three columns (we only use 2)
df = pandas.DataFrame({
    "time" : range(T),
    "pressure": np.random.rand(T),
    "temp" : np.random.rand(T)
})
## build up a figure, ggplot2 style
# instantiate the figure object
fig = d3py.Figure(df, name="basic_example", width=300, height=300) 
# add some red points
fig += d3py.geoms.Point(x="pressure", y="temp", fill="red")
# writes 3 files, starts up a server, then draws some beautiful points in Chrome
fig.show()

I tried to run the sample program from README:

import d3py
import pandas
import numpy as np

# some test data
T = 100
# this is a data frame with three columns (we only use 2)
df = pandas.DataFrame({
    "time" : range(T),
    "pressure": np.random.rand(T),
    "temp" : np.random.rand(T)
})
## build up a figure, ggplot2 style
# instantiate the figure object
fig = d3py.Figure(df, name="basic_example", width=300, height=300) 
# add some red points
fig += d3py.geoms.Point(x="pressure", y="temp", fill="red")
# writes 3 files, starts up a server, then draws some beautiful points in Chrome
fig.show() 

Error:

Traceback (most recent call last):
  File "sample.py", line 15, in <module>
    fig = d3py.Figure(df, name="basic_example", width=300, height=300) 
  File "/Users/amjith/.virtualenvs/d3py/lib/python2.7/site-packages/d3py/d3py.py", line 97, in __init__
    self.filemap = {"static/d3.js":{"fd":open("static/d3.js","r"), "timestamp":time.time()},}
IOError: [Errno 2] No such file or directory: 'static/d3.js'
Error in clean-up: 'Figure' object has no attribute 'httpd'

Here's my pip freeze

d3py==0.1.0
numpy==1.6.2
pandas==0.10.0
python-dateutil==2.1
pytz==2012h
six==1.2.0
wsgiref==0.1.2

Can we set the default width and height using CSS? It would be nice if this could be styled

At the moment, d3py needs some modifications to work on Python 3. I think all the supporting modules are available on py 3. How about d3py?

take axis labels from the data frame column names unless explicity overruled

I cloned d3py and tried to run the d3py_graph.py example, but ran into a problem.

$ git clone git://github.com/mikedewar/d3py.git
[...]
$ cd d3py/
$ python setup.py install
[...]
$ cd examples/
$ python d3py_graph.py 
Traceback (most recent call last):
  File "d3py_graph.py", line 15, in <module>
    with d3py.NetworkXFigure(G, width=500, height=500) as p:
  File "[...]/local/lib/python2.7/site-packages/d3py/networkx_figure.py", line 39, in __init__
    port=port, **kwargs
TypeError: __init__() takes exactly 10 arguments (9 given)
Error in clean-up: 'NetworkXFigure' object has no attribute 'httpd'

I just discovered d3py a few minutes ago, so forgive me if I've missed something. I got the demo to run like this:

$ git diff
diff --git a/examples/d3py_graph.py b/examples/d3py_graph.py
index 99c73ba..b1d9e6a 100644
--- a/examples/d3py_graph.py
+++ b/examples/d3py_graph.py
@@ -12,6 +12,6 @@ G.add_edge(3,4)
 G.add_edge(4,2)

 # use 'with' if you are writing a script and want to serve this up forever
-with d3py.NetworkXFigure(G, width=500, height=500) as p:
+with d3py.NetworkXFigure(G, width=500, height=500, host='localhost') as p:
     p += d3py.ForceLayout()
     p.show()

Unlike PandasFigure(Figure), NetworkXFigure(Figure) does not have a default host argument.

Traceback (most recent call last):
File "testd3.py", line 1, in
import d3py
File "/usr/local/lib/python2.7/dist-packages/d3py/init.py", line 2, in
from geoms import *
File "/usr/local/lib/python2.7/dist-packages/d3py/geoms/init.py", line 5, in
from axis import Axis
ImportError: No module named axis

when running the example code

import d3py
import pandas
import numpy as np

some test data

T = 100

this is a data frame with three columns (we only use 2)

df = pandas.DataFrame({
"time" : range(T),
"pressure": np.random.rand(T),
"temp" : np.random.rand(T)
})

build up a figure, ggplot2 style

instantiate the figure object

fig = d3py.Figure(df, name="basic_example", width=300, height=300)

add some red points

fig += d3py.geoms.Point(x="pressure", y="temp", fill="red")

writes 3 files, starts up a server, then draws some beautiful points in Chrome

fig.show()

It would be nice to have a meaningful time axis for the line graph

System: Fedora 16
Installed from git repo

When i run any of the examples, i get:

$ python d3py_scatter.py
you can find your chart at http://localhost:8000/my_figure/my_figure.html

(gnome-open:12202): GConf-WARNING **: Client failed to connect to the D-BUS daemon:
Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.
GConf Error: No D-BUS daemon running

Shutting down httpd

DBUS is installed and running.

Thanks!

scenario:
In python web applications, one would want to insert d3 visualization with d3py by "printing" html snippet to an existing html. For example, googleVis package in R provides such functionality in its print function, which can be used with R markdown to produce html page easily.

Every time a geom is added, the axes should update themselves to reflect the new geom's ranges

I get the following error when running the example code in the README:

File "d3py.py", line 1, in <module>
    import d3py
  File "/Users/sloria1/projects/sandbox/d3py.py", line 15, in <module>
    fig = d3py.Figure(df, name="basic_example", width=300, height=300) 
AttributeError: 'module' object has no attribute 'Figure'

From the README:

import d3py
import pandas
import numpy as np

# some test data
T = 100
# this is a data frame with three columns (we only use 2)
df = pandas.DataFrame({
    "time" : range(T),
    "pressure": np.random.rand(T),
    "temp" : np.random.rand(T)
})
## build up a figure, ggplot2 style
# instantiate the figure object
fig = d3py.Figure(df, name="basic_example", width=300, height=300) 
# add some red points
fig += d3py.geoms.Point(x="pressure", y="temp", fill="red")
# writes 3 files, starts up a server, then draws some beautiful points in Chrome
fig.show()

The user should be able to map a variable to the color of each shape on the graph. In ggplot, this would run:

require(ggplot2)
df = data.frame(x=rnorm(100), y=rnorm(100, 10, 1), color=rep(1:4, 25))
ggplot(data = df)+geom_point(aes(x = x, y= y, color=color))

It seems like a simple enough fix for geom_point at least. Is there a reason why this hasn't been done?