Mouse_Gene_Phenotypes error

[pnrobinson]

When I run this code in my own notebook, http://nbviewer.jupyter.org/github/biolink/ontobio/blob/master/notebooks/Mouse_Gene_Phenotypes.ipynb

z, xlabels, ylabels = aset.similarity_matrix(diseases, diseases)
import plotly.plotly as py
import plotly.graph_objs as go
import numpy as np
trace = go.Heatmap(z=-np.array(z),
                       x=xlabels,
                       y=ylabels)
py.iplot([trace], filename='labelled-heatmap')

leads to the following error message

Aw, snap! We didn't get a username with your request.
Don't have an account? https://plot.ly/api_signup
Questions? accounts@plot.ly
---------------------------------------------------------------------------
PlotlyError                               Traceback (most recent call last)
<ipython-input-17-ba5d83e243b8> in <module>()
      6                        x=xlabels,
      7                        y=ylabels)
----> 8 py.iplot([trace], filename='labelled-heatmap')

~/bin/anaconda3/lib/python3.6/site-packages/plotly/plotly/plotly.py in iplot(figure_or_data, **plot_options)
    171         embed_options['height'] = str(embed_options['height']) + 'px'
    172 
--> 173     return tools.embed(url, **embed_options)
    174 
    175 

~/bin/anaconda3/lib/python3.6/site-packages/plotly/tools.py in embed(file_owner_or_url, file_id, width, height)
    393         else:
    394             url = file_owner_or_url
--> 395         return PlotlyDisplay(url, width, height)
    396     else:
    397         if (get_config_defaults()['plotly_domain']

~/bin/anaconda3/lib/python3.6/site-packages/plotly/tools.py in __init__(self, url, width, height)
   1440         def __init__(self, url, width, height):
   1441             self.resource = url
-> 1442             self.embed_code = get_embed(url, width=width, height=height)
   1443             super(PlotlyDisplay, self).__init__(data=self.embed_code)
   1444 

~/bin/anaconda3/lib/python3.6/site-packages/plotly/tools.py in get_embed(file_owner_or_url, file_id, width, height)
    298                 "'{1}'."
    299                 "\nRun help on this function for more information."
--> 300                 "".format(url, plotly_rest_url))
    301         urlsplit = six.moves.urllib.parse.urlparse(url)
    302         file_owner = urlsplit.path.split('/')[1].split('~')[1]

PlotlyError: Because you didn't supply a 'file_id' in the call, we're assuming you're trying to snag a figure from a url. You supplied the url, '', we expected it to start with 'https://plot.ly'.
Run help on this function for more information.

The notebook should explain how to set this up correctly

[cmungall]

This doesn't help you much right now, but I'm thinking we should use seaborn rather than plotly. Seems to be the most popular choice for data science in python - and doesn't require use of a server. Thoughts @kltm @kshefchek?

[pnrobinson]

@cmungall @kltm @kshefchek Here is an alternative that works with Seaborn thanks to my colleague @adeslatt https://github.com/TheJacksonLaboratory/robinson-lab-notebooks/blob/master/Linked_Data_Biolink.ipynb

from ontobio.assoc_factory import AssociationSetFactory
afactory = AssociationSetFactory()
HUMAN='NCBITaxon:9606'
aset=afactory.create(ontology=ont, subject_category='disease',object_category='phenotype',taxon=HUMAN)
[t]=ont.search('Nephroblastoma')
diseases=aset.query([t])
dlabels = ["{} '{}'".format(d, aset.label(d)) for d in diseases]
z, xlabels, ylabels = aset.similarity_matrix(diseases, diseases)
import seaborn as sns
import numpy as np 
import pandas as pd

sns.set(color_codes=True)
zarray = np.array(z)
zarray_df = pd.DataFrame(zarray, columns=dlabels, index=dlabels)

cm = sns.clustermap(zarray_df, method="average",metric="Euclid")