In [ ]:

    

from pymongo import MongoClient


    

In [ ]:

    

client = MongoClient('localhost', 27017)


    

In [ ]:

    

db = client['planet_four']


    

In [ ]:

    

from odo import odo


    

In [ ]:

    

odo('mongodb://localhost:27017/planet_four::planet_four_subjects',
    'hdfstore://Users/klay6683/data/planet4/2015-03-31.hdf5::/df')


    

In [ ]:

    

import blaze as bz


    

In [ ]:

    

bz.Data('mongodb://localhost/planet_four::planet_four_subjects')


    

In [ ]:

    

blaze.__version__


    

In [ ]:

    

subjects = db['planet_four_subjects']
classifications = db['planet_four_classifications']
users = db['planet_four_users']


    

In [ ]:

    

df = pd.DataFrame(list(classifications.find({'user_name':'michaelaye'})))


    

In [ ]:

    

df.annotations[1]


    

In [ ]:

    

subjects.find_one().keys()


    

In [ ]:

    

from toolz import take


    

In [ ]:

    

from bson.objectid import ObjectId


    

In [ ]:

    

from datashape import discover


    

In [ ]:

    

items = list(take(2, classifications.find()))
oid_cols = [k for k,v in items[0].items() if isinstance(v, ObjectId)]
print("OID_cols:\n",oid_cols)
for item in items:
    for col in oid_cols:
        del item[col]
print('items[0].keys():\n',items[0].keys())
discover(items)


    

In [ ]:

    

with open('classification_fields.txt','w') as f:
    for key in items[0].keys():
        f.write(key)
        f.write('\n')


    

In [ ]:

    

!cat classification_fields.txt


    

In [ ]:

    

pwd


    

In [ ]: