In [2]:

    

import pandas as pd
import numpy as np
%matplotlib inline


    

In [3]:

    

from pattern.en import parsetree


    

In [4]:

    

vcodes_df = pd.read_csv("vcodes.csv")


    

In [5]:

    

vcodes_df.head()


    

    
Out[5]:






  

    

      

      
code
      
description
    
  
  

    

      
0
      
V010
      
Pedestrian injured in collision with pedal cyc...
    
    

      
1
      
V011
      
Pedestrian injured in collision with pedal cyc...
    
    

      
2
      
V019
      
Pedestrian injured in collision with pedal cyc...
    
    

      
3
      
V020
      
Pedestrian injured in collision with two- or t...
    
    

      
4
      
V021
      
Pedestrian injured in collision with two- or t...
    
  

In [6]:

    

vcodes_df['parsetree'] = vcodes_df['description'].map(parsetree)


    

In [7]:

    

vcodes_df.head()


    

    
Out[7]:






  

    

      

      
code
      
description
      
parsetree
    
  
  

    

      
0
      
V010
      
Pedestrian injured in collision with pedal cyc...
      
[(Pedestrian, injured, in, collision, with, pe...
    
    

      
1
      
V011
      
Pedestrian injured in collision with pedal cyc...
      
[(Pedestrian, injured, in, collision, with, pe...
    
    

      
2
      
V019
      
Pedestrian injured in collision with pedal cyc...
      
[(Pedestrian, injured, in, collision, with, pe...
    
    

      
3
      
V020
      
Pedestrian injured in collision with two- or t...
      
[(Pedestrian, injured, in, collision, with, tw...
    
    

      
4
      
V021
      
Pedestrian injured in collision with two- or t...
      
[(Pedestrian, injured, in, collision, with, tw...
    
  

In [8]:

    

ptree = vcodes_df.ix[118, 2]


    

In [9]:

    

vcodes_df.ix[118, 1]


    

    
Out[9]:





'Motorcycle rider injured in collision with car, pick-up truck or van: Unspecified motorcycle rider injured in nontraffic accident'

In [10]:

    

for sentence in ptree:
    for chunk in sentence.chunks:
        print chunk


    

    




Chunk('Motorcycle rider/NP')
Chunk('injured/VP')
Chunk('in/PP')
Chunk('collision/NP')
Chunk('with/PP')
Chunk('car/NP')
Chunk('pick-up truck/NP')
Chunk('van/NP')
Chunk('Unspecified motorcycle rider/NP')
Chunk('injured/VP')
Chunk('in/PP')
Chunk('nontraffic accident/NP')

In [12]:

    

ptree = vcodes_df.ix[377, 2]


    

In [14]:

    

for sentence in ptree:
    for chunk in sentence.chunks:
        print chunk


    

    




Chunk('Occupant/NP')
Chunk('of/PP')
Chunk('pick-up truck/NP')
Chunk('van injured/VP')
Chunk('in/PP')
Chunk('collision/NP')
Chunk('with/PP')
Chunk('car/NP')
Chunk('pick-up truck/NP')
Chunk('van/NP')
Chunk('Unspecified/VP')
Chunk('occupant/NP')
Chunk('of/PP')
Chunk('pick-up truck/NP')
Chunk('van injured/VP')
Chunk('in/PP')
Chunk('nontraffic accident/NP')

In [16]:

    

vcodes_df.ix[377, 1]


    

    
Out[16]:





'Occupant of pick-up truck or van injured in collision with car, pick-up truck or van: Unspecified occupant of pick-up truck or van injured in nontraffic accident'

In [23]:

    

for sentence in vcodes_df.ix[378, 2]:
    for chunk in sentence.chunks:
        print chunk


    

    




Chunk('Occupant/NP')
Chunk('of/PP')
Chunk('pick-up truck/NP')
Chunk('van injured/VP')
Chunk('in/PP')
Chunk('collision/NP')
Chunk('with/PP')
Chunk('car/NP')
Chunk('pick-up truck/NP')
Chunk('van/NP')
Chunk('Person/NP')
Chunk('injured/VP')
Chunk('while/PP')
Chunk('boarding/VP')
Chunk('alighting/ADJP')

In [24]:

    

vcodes_df.ix[378, 1]


    

    
Out[24]:





'Occupant of pick-up truck or van injured in collision with car, pick-up truck or van: Person injured while boarding or alighting'

In [26]:

    

sent = ptree[0]


    

In [30]:

    

vcodes_df['parsetree'].map(lambda x: len(x[0].subjects)).unique()


    

    
Out[30]:





array([0])

In [47]:

    

pu_truck = sent.chunks[2]


    

In [48]:

    

pu_truck.string


    

    
Out[48]:





u'pick-up truck'