Working on a salary dataset. Everything works fine except when I use the python-graphviz module in Spyder 3.3.2 to show the decision tree graph it simply shows an image icon in console window. Well the same thing works on other systems. What I'm missing here?
The output image is here.Console Output
from sklearn.tree import DecisionTreeClassifier
dtf = DecisionTreeClassifier()
dtf.fit(X_train, y_train)
from sklearn.tree import export_graphviz
export_graphviz(dtf, out_file="tree.dot", class_names=["Less than 50k",
"More than 50k"])
import graphviz
with open("tree.dot") as f:
dot_graph = f.read()
graphviz.Source(dot_graph)
(spyder maintainer here) This seems a limitation of QtConsole, which is the package that powers our IPython consoles.
Please open an issue on the repo referenced above about this so we don't forget to fix it in the future.
Related
I am trying to set the theme for bokeh plots on jupyter lab.
To do so I am following the official docs, copying the exact code from https://docs.bokeh.org/en/latest/docs/reference/themes.html.
However, I get an error cannot import name 'built_in_themes'.
I assume the default themes location changed, but can't seem to find anything else besides the said page documenting bokeh.themes.
I assume the default themes location changed
No, it has only ever been in the same location since it was added in version 1.0. The explanation is almost certainly that the version you are using is older than that.
As bigreddot said, same location.
Check your Bokeh version:
$ python3.7
>>> import sys
>>> print(sys.version)
>>> import bokeh
>>> print(bokeh.__version__)
Or via jupyter notebook:
import numpy as np
from bokeh.io import output_notebook, show
from bokeh.plotting import figure
Following the example here provided, the output should look like this:
Maybe you are just working in an environment with an older version.
I'm trying to do the Udacity mini project and I've got the latest version of the SKLearn library installed (20.2).
When I run:
from sklearn.decomposition import RandomizedPCA
I get the error:
ImportError: cannot import name 'RandomizedPCA' from 'sklearn.decomposition' (/Users/kintesh/Documents/udacity_ml/python3/venv/lib/python3.7/site-packages/sklearn/decomposition/__init__.py)
I actually even upgraded the version using:
pip3 install -U scikit-learn
Which upgraded from 0.20.0 to 0.20.2, which also uninstalled and reinstalled... so I'm not sure why it can't initialise sklearn.decomposition.
Are there any solutions here that might not result in completely uninstalling python3 from my machine?! Would ideally like to avoid that.
Any help would be thoroughly appreciated!
Edit:
I'm doing some digging and trying to fix this, and it appears as though the __init__.py file in the decomposition library on the SKLearn GitHub doesn't reference RandomizedPCA... has it been removed or something?
Link to the GitHub page
As it turns out, RandomizePCA() was depreciated in an older version of SKLearn and is simply a parameter in PCA().
You can fix this by changing the import statement to:
from sklearn.decomposition import PCA as RandomizedPCA
... and then your classifier looks like this:
pca = RandomizedPCA(n_components=n_components, svd_solver='randomized', whiten=True).fit(X_train)
However, if you're here because you're doing the Udacity Machine Learning course on Eigenfaces.py, you'll notice that the PIL library is also deprecated.
Unfortunately I don't have a solution for that one, but here's the GitHub issue page, and here's a kind hearted soul that used a Jupyter Notebook to solve their mini-project back when these repositories worked.
I hope this helps, and gives enough information for the next person to get into Machine Learning. If I get some time I might take a crack at recoding eigenfaces.py for SKLearn 0.20.2, but for now I'm just going to crack on with the rest of this course.
In addition to what #Aaraeus said, the PIL library has been forked to Pillow.
You can fix the PIL import error using
pip3 install pillow
I am trying to make a span in bokeh using jupyter widgets.
from ipywidgets import interact
import numpy as np
from scipy.stats import norm
from bokeh.sampledata.daylight import daylight_warsaw_2013
from bokeh.io import push_notebook, show, output_notebook
from bokeh.plotting import figure
from bokeh.models import Span
output_notebook()
p = figure()
x_axis = np.arange(-10, 10, 0.001)
# Mean = 0, SD = 2.
y_axis = norm.pdf(x_axis,0,2)
p.line(x_axis, y_axis, line_dash='solid', line_width=2)
cutoff = Span(location=1,
dimension='height', line_color='green',
line_dash='dashed', line_width=2)
p.add_layout(cutoff)
show(p, notebook_handle=True)
def update(new_cutoff_location):
cutoff.location = new_cutoff_location
push_notebook()
interact(update, new_cutoff_location = 1.0)
When I run this code I get ValueError: PATCH-DOC message requires at least one event at push_notebook(). I suspect this indicates that the update to cutoff.location isn't getting detected, so it it looks as if there are no changes to send. Passing the handle doesn't seem to make a difference. Looking at the sample code in this github issue, it looks like there used to be an set method on span elements, but there doesn't appear to be one on my span element cutoff. Maybe there is a different function I am supposed to call to register the change?
I'm on bokeh 0.12.11 with jupyter 1.0.0, jupyter-client 5.1.0, jupyter-console 5.2.0, jupyter-core 4.4.0
This appears to be a regression in Bokeh 0.12.11. Your code works with version 0.12.10 so the immediate workaround is to downgrade. I've made GitHub issue here that you can follow. We will issue a new point release with a fix ASAP.
UPDATE: The issue is now fixed in recent versions of Bokeh
I am using Plotly to make graphs in my IPython notebook. I am able to view graphs on my IPython notebook when I upload them on GitHub they are displayed as blank spaces.
I read on the web that Plotly currently does not support iframes and hence the issue, but is there a workaround?
Here's the link to my GitHub Ipython notebook:
https://github.com/dhavalbhinde/bhinde_dhaval_spring2017/blob/master/Finals/Analysis%203.ipynb
Please, can someone advice how should I handle them?
I found a way to show Plotly plots on Github. They aren’t interactive anymore but it’s better than nothing.
First
import plotly.io as pio
pio.renderers
you can see the list of available renders.
*if you get an error on this step you can simply just install orca:
conda install -c plotly plotly-orca
and then there are 2 possible ways.
you can pass "svg" to .show() like this:
fig = px.scatter_3d(iris, x=transformed_iris['component1'], y=transformed_iris['component2'], z=transformed_iris['component3'],color='species')
fig.show(renderer="svg")
or you can set the pio.renderers.default to svg:
pio.renderers.default = "svg"
Import these and this and it will work :
from plotly.offline import plot, iplot, init_notebook_mode
import plotly.graph_objs as go
init_notebook_mode(connected=True)
After exporting a .dot file using scikit-learn's handy export_graphviz function.
I am trying to render the dot file using Graphviz into a cell in my Jupyter Notebook:
import graphviz
from IPython.display import display
with open("tree_1.dot") as f:
dot_graph = f.read()
display(graphviz.Source(dot_graph))
However the out[ ] is just an empty cell.
I am using graphviz 0.5 (pip then conda installed), iPython 5.1, and Python 3.5
The dot file looks correct here are the first characters:
digraph Tree {\nnode [shape=box, style="filled", color=
iPython display seems to work for other objects including Matplotlib plots and Pandas dataframes.
I should note the example on Graphviz' site also doesn't work.
It's possible that since you posted this, changes were made so you might want to update your libraries if that's possible.
The versions of relevance here I used are:
Python 2.7.10
IPython 5.1.0
graphviz 0.7.1
If you have a well formed .dot file, you can display it to the jupyter out[.] cell by the following:
import graphviz
with open("tree_1.dot") as f:
dot_graph = f.read()
# remove the display(...)
graphviz.Source(dot_graph)
this solution allows you to insert DOT text directly (without saving it to file first)
# convert a DOT source into graph directly
import graphviz
from IPython.display import display
source= '''\
digraph sample {
A[label="AL"]
B[label="BL"]
C[label="CL"]
A->B
B->C
B->D
D->C
C->A
}
'''
print (source)
gvz=graphviz.Source(source)
# produce PDF
#gvz.view()
print (gvz.source)
display(gvz)
Try to use pydotplus.
import pydotplus
by (1.1) Importing the .dot from outside
pydot_graph = pydotplus.graph_from_dot_file("clf.dot")
or (1.2) Directly using the .export_graphviz output
dt = tree.DecisionTreeClassifier()
dt = clf.fit(x,y)
dt_graphviz = tree.export_graphviz(dt, out_file = None)
pydot_graph = pydotplus.graph_from_dot_data(dt_graphviz)
(2.) and than display the pyplot graph using
from IPython.display import Image
Image(pydot_graph.create_png())
try to reinstall graphviz
conda remove graphviz
conda install python-graphviz
graphviz.Source(dot_graph).view()
graphviz.Source(dot_graph).view()