Getting Started

Installation

the easiest way to install is with pip

# install from pipi
pip install bokeh_wordcloud2

# install from github
pip install git+https://github.com/joranbeasley/bokeh_wordcloud2.git#bokeh_wordcloud

# clone and build
git clone pip git+https://github.com/joranbeasley/bokeh_wordcloud2.git
cd bokeh_wordcloud2
pip install .  # alternatively `python setup.py install`

Examples

Interactive Jupyter Notebook

Feel free to play with the jupyter notebook available at https://gke.mybinder.org/v2/gh/joranbeasley/bokeh-wordcloud-notebook/master?filepath=bokeh_wordcloud2.ipynb

First Word Cloud

here is a super simple word cloud to get you started

you can find this example at Simple Example Source

you can see its html here Simple Example HMTL

from bokeh.io import show
from bokeh.models import ColumnDataSource
from bokeh_wordcloud2 import WordCloud2

data = [
    ['susan',6], ['tom',3], ['frankie',8],
    ['roger',7], ['amy',9], ['nicole',10],
    ['joran',5], ['mark',4], ['brianne',7],
    ['michael',8], ['greg',4], ['adrian',6],
    ['drew',9]
]
names,weights = zip(*data)
test1 = ColumnDataSource({'names':names,'weights':weights})
# we will specify just "blue" for the color
wordcloud = WordCloud2(source=test1,wordCol="names",sizeCol="weights",colors="blue")
show(wordcloud)

when you run the application you should see a webpage open with your newly created Wordcloud

data = [
    ['susan',6],
    ['tom',3],
    ['frankie',8],
    ['roger',7],
    ['amy',9],
    ['nicole',10],
    ['joran',5],
    ['mark',4],
    ['brianne',7],
    ['michael',8],
    ['greg',4],
    ['adrian',6],
    ['drew',9]
]
names,weights = zip(*data)
test1 = ColumnDataSource({'names':names,'weights':weights})
sdp = WordCloud2(source=test1,wordCol="names",sizeCol="weights",colors=['pink','blue','green'])
show(sdp)
    

Using a corpus instead

if you dont specify a sizeCol, then it will extract all the words from the wordCol and use the counts as a size

you can find this example at Example Source

Color Options

for the color options we can pass in a single color that will be assigned to all words

# all the words are pink
WordCloud2(source=test2,wordCol="titles",colors='pink')

or we can pass in a list, and randomly select from it:

# pick a random color
random_colors=['pink','blue']
WordCloud2(source=test2, wordCol="titles", colors=random_colors)

or if our datasource has a column for colur we can pass in the name of that (only works if you specify sizeCol)

colors = [['red','green','blue','purple'][i%4]
            for i in range(len(test1.data['words']))]
test1.data['colorsCol']  = colors # assign a new column
WordCloud2(source=test1,wordCol="names",sizeCol="weights",
           colors="colorsCol")  # use our column name instead

or we can supply a javascript callback that returns a string, but we name it slightly differently

colorFun = CustomJS(code="""
console.log("PICKING A COLOR FOR:",cb_obj)
console.log("Got Word:",cb_data['word'],cb_data['weight'],cb_data['fontSize'])
return "red" # dont forget to RETURN a value
""")

WordCloud2(source=test1,wordCol="names",sizeCol="weights",colorsFun=colorFun)

you can find this example at Color Options

you can see its html here Color Options HMTL

Clicks And Hovers

you can subscribe to either clicks or hovers with a javascript object

you can find this example at JS Callbacks

you can see its html here JS Callbacks HMTL

Python Click Callback

Note

this only applies when running bokeh server bokeh run my_app.py

you can also subscribe to the click handler in python if you are running with bokeh server

wordcloud = WordCloud2(source=test1, wordCol="names", sizeCol="weights", colors=['pink', 'blue', 'green'])

def clicked_word(evt):
    print("GOT:",evt)
    data=test1.data
    if(7 < evt.weight < 20 ):
        new_weight = evt.weight + random.choice([-1,1,1,2])
    elif evt.weight < 7:
        new_weight = evt.weight + random.choice([1,2])
    elif evt.weight < 20:
        new_weight = evt.weight - random.choice([1,2,3])
    weights = list(data['weights'])
    weights[data['names'].index(evt.word)] = new_weight
    # make sure to reassign back to data a new dict, or the difference might not be noticed
    test1.data = {'names':data['names'][:],'weights':weights}

# subscribe to the click event
wordcloud.on_click(clicked_word)
curdoc().add_root(column(wordcloud))

you can find this example at Bokeh Server Example

since you must be running with bokeh serve app.py I cannot show you the html