In [1]:

    

from harness import Harness
import pandas
from harness.python.ext import HarnessExtension


    

In [2]:

    

class NetworkXMixin:
    graph = None

class NetworkXExtension(HarnessExtension):
    imports = 'networkx'
    mixin = NetworkXMixin
    
    def keywords(self, dataframe):
        return {
            'G': lambda: dataframe.graph,
            'df': lambda: dataframe,
        }
    
    def to_graph(self, source, target, edge_attr=True, df=None):
        df.graph = self.module_.from_pandas_dataframe(
            df, source, target, edge_attr
        )
        return df.graph
    
    def callback(self, dataframe, value):
        if isinstance(value, self.module_.Graph):
            return dataframe
        return value
    
    def edges(self, df):
        return df.graph.edges()
    
    def nodes(self, df):
        return df.graph.nodes()


    

In [3]:

    

extensions = [
    'harness.python.ext.base.JinjaExtension',
    'harness.python.ext.SciKit.SciKitExtension', 
    'harness.python.ext.Bokeh.BokehModelsExtension',     
    'harness.python.ext.Bokeh.BokehPlottingExtension',
    'harness.python.ext.Bokeh.BokehChartsExtension',
    '__main__.NetworkXExtension'
]


    

In [4]:

    

df = Harness(pandas.np.random.randn(10,2), extensions=extensions)


    

In [5]:

    

df.to_graph(source=0, target=1)


    

    
Out[5]:






  

    

      

      
0
      
1
    
  
  

    

      
0
      
-1.117878
      
1.262091
    
    

      
1
      
-1.359881
      
0.319997
    
    

      
2
      
1.434786
      
0.632406
    
    

      
3
      
-0.627768
      
-2.226961
    
    

      
4
      
-1.352356
      
-0.043742
    
    

      
5
      
0.577460
      
-0.980929
    
    

      
6
      
0.677719
      
-0.031278
    
    

      
7
      
0.874833
      
0.123264
    
    

      
8
      
-0.496750
      
0.627734
    
    

      
9
      
1.089115
      
-0.090981
    
  

In [6]:

    

df.to_graph(0, 1).edges()


    

    
Out[6]:





[(0.31999679444203105, -1.3598809977366302),
 (1.2620913813930383, -1.1178781224538343),
 (-0.49675022780319023, 0.62773419742106085),
 (1.4347864011340115, 0.63240646399100764),
 (-0.043741889938154949, -1.3523562694971041),
 (-0.031278057054313593, 0.67771868270860069),
 (0.12326369370788465, 0.87483306496355573),
 (0.57746008474171939, -0.98092933524192161),
 (-0.6277682035366926, -2.2269610526154016),
 (1.0891147149001867, -0.090980690694986377)]

In [7]:

    

df.edges()


    

    
Out[7]:





[(0.31999679444203105, -1.3598809977366302),
 (1.2620913813930383, -1.1178781224538343),
 (-0.49675022780319023, 0.62773419742106085),
 (1.4347864011340115, 0.63240646399100764),
 (-0.043741889938154949, -1.3523562694971041),
 (-0.031278057054313593, 0.67771868270860069),
 (0.12326369370788465, 0.87483306496355573),
 (0.57746008474171939, -0.98092933524192161),
 (-0.6277682035366926, -2.2269610526154016),
 (1.0891147149001867, -0.090980690694986377)]

In [8]:

    

Harness(df.spring_layout()).transpose().sample(2)


    

    
Out[8]:






  

    

      

      
0
      
1
    
  
  

    

      
0.319997
      
0.072384
      
0.232406
    
    

      
-1.352356
      
0.875902
      
0.171825
    
  

In [9]:

    

%matplotlib inline
df.draw_circular()


    

In [10]: