#!/usr/bin/env python
# coding: utf-8

# This example makes use of the Google/OpenRefine client that can be found at https://github.com/PaulMakepeace/refine-client-py
# 
# The example is inspired by, and draws heavily from, Trevor Muñoz's tutorial posted at http://nbviewer.ipython.org/gist/trevormunoz/6265360
# 
# ---
# First we need to import some necessary libraries...

# In[4]:


get_ipython().system('~/anaconda/bin/pip install git+https://github.com/PaulMakepeace/refine-client-py.git')


# In[10]:


get_ipython().system('~/anaconda/bin/pip install refine-client')


# In[5]:


from google.refine import refine, facet


# Make sure you have OpenRefine running locally before going any further! (You can download and install it from [OpenRefine.org](http://openrefine.org).)
# 
# Once you've started OpenRefine in the normal way, get a link to the server...

# In[8]:


server = refine.RefineServer()
orefine = refine.Refine(server)


# Let's check everything's working by previewing three examples from the list of any projects that have previously been created...

# In[9]:


orefine.list_projects().items()[:3]


# If we want to work with one of these projects, we can open it directly:

# In[5]:


project_file = 'smalltest.csv'


# In[17]:


p=orefine.new_project(project_file=project_file)


# In[25]:


pr=p.get_rows(limit=10)


# In[14]:


dir(pr.rows[0])


# In[29]:


pr.rows[0].row


# In[32]:


f=facet.TextFacet('CompanyName')


# In[33]:


dir(f)


# In[42]:


fr=p.compute_facets(facet.TextFacet('CompanyName'))


# In[43]:


facets = fr.facets[0]
for k in sorted(facets.choices, key=lambda k: facets.choices[k].count, reverse=True):
    print facets.choices[k].count, k


# In[50]:


dir(pr.rows.rows_response)
pr.rows.rows_response


# In[53]:


dir(pr)


# In[54]:


dir(p)


# In[55]:


p.columns


# In[ ]: