Remove dependency on pandas (#105)

* Remove dependency on pandas

* Remove debug print

Author: shifucun

datacommons/BUILD.bazel

@@ -1,10 +1,6 @@
 package(default_visibility = ["//visibility:public"])
-load("@requirements//:requirements.bzl", "requirement")
 
 py_library(
     name = "datacommons",
     srcs = glob(["*.py"]),
-    deps = [
-        requirement("pandas"),
-    ]
 )

datacommons/__init__.py

@@ -21,4 +21,4 @@
 from datacommons.populations import get_populations, get_observations, get_pop_obs, get_place_obs
 
 # Other utilities
-from .utils import set_api_key, clean_frame, flatten_frame
+from .utils import set_api_key

datacommons/core.py

@@ -28,8 +28,6 @@
 
 from collections import defaultdict
 
-import pandas as pd
-
 import datacommons.utils as utils
 import requests
 
@@ -40,7 +38,7 @@ def get_property_labels(dcids, out=True):
   """ Returns the labels of properties defined for the given :code:`dcids`.
 
   Args:
-    dcids (:obj:`list` of :obj:`str`): A list of nodes identified by their
+    dcids (:obj:`iterable` of :obj:`str`): A list of nodes identified by their
       dcids.
     out (:obj:`bool`, optional): Whether or not the property points away from
       the given list of nodes.
@@ -99,6 +97,7 @@ def get_property_labels(dcids, out=True):
     }
   """
   # Generate the GetProperty query and send the request
+  dcids = list(dcids)
   url = utils._API_ROOT + utils._API_ENDPOINTS['get_property_labels']
   payload = utils._send_request(url, req_json={'dcids': dcids})
 
@@ -120,8 +119,7 @@ def get_property_values(dcids,
   """ Returns property values of given :code:`dcids` along the given property.
 
   Args:
-    dcids (Union[:obj:`list` of :obj:`str`, :obj:`pandas.Series`]): dcids to get
-      property values for.
+    dcids (:obj:`iterable` of :obj:`str`): dcids to get property values for.
     prop (:obj:`str`): The property to get property values for.
     out (:obj:`bool`, optional): A flag that indicates the property is directed
       away from the given nodes when set to true.
@@ -131,15 +129,8 @@ def get_property_values(dcids,
       aggregated over all given nodes.
 
   Returns:
-    When :code:`dcids` is an instance of :obj:`list`, the returned property
-    values are formatted as a :obj:`dict` from a given dcid to a list of its
-    property values.
-
-    When :code:`dcids` is an instance of :obj:`pandas.Series`, the returned
-    property values are formatted as a :obj:`pandas.Series` where the `i`-th
-    entry corresponds to property values associated with the `i`-th given dcid.
-    The cells of the returned series will always contain a :obj:`list` of
-    property values.
+    Returned property values are formatted as a :obj:`dict` from a given dcid
+    to a list of its property values.
 
   Raises:
     ValueError: If the payload returned by the Data Commons REST API is
@@ -160,21 +151,11 @@ def get_property_values(dcids,
       "geoId/21": ["Kentucky"],
       "geoId/24": ["Maryland"],
     }
-
-    Next, we specify :code:`dcids` as a :obj:`pandas.Series`
-
-    >>> import pandas as pd
-    >>> dcids = pd.Series(["geoId/06", "geoId/21", "geoId/24"])
-    >>> get_property_values(dcids, "name")
-    0    [California]
-    1      [Kentucky]
-    2      [Maryland]
-    dtype: object
   """
   # Convert the dcids field and format the request to GetPropertyValue
-  dcids, req_dcids = utils._convert_dcids_type(dcids)
+  dcids = list(dcids)
   req_json = {
-    'dcids': req_dcids,
+    'dcids': dcids,
     'property': prop,
     'limit': limit
   }
@@ -205,9 +186,6 @@ def get_property_values(dcids,
   # Make sure each dcid is in the results dict, and convert all sets to lists.
   results = {dcid: sorted(list(unique_results[dcid])) for dcid in dcids}
 
-  # Format the results as a Series if a Pandas Series is provided.
-  if isinstance(dcids, pd.Series):
-    return pd.Series([results[dcid] for dcid in dcids], index=dcids.index)
   return results
 
 
@@ -221,7 +199,7 @@ def get_triples(dcids, limit=utils._MAX_LIMIT):
   *predicate*).
 
   Args:
-    dcids (:obj:`list` of :obj:`str`): A list of dcids to get triples for.
+    dcids (:obj:`iterable` of :obj:`str`): A list of dcids to get triples for.
     limit (:obj:`int`, optional): The maximum total number of triples to get.
 
   Returns:
@@ -249,6 +227,7 @@ def get_triples(dcids, limit=utils._MAX_LIMIT):
     }
   """
   # Generate the GetTriple query and send the request.
+  dcids = list(dcids)
   url = utils._API_ROOT + utils._API_ENDPOINTS['get_triples']
   payload = utils._send_request(url, req_json={'dcids': dcids, 'limit': limit})
 

datacommons/examples/core.py

@@ -77,7 +77,7 @@ def main():
   # To expand on a column with get_property_values, the data frame has to be
   # flattened first. Clients can use flatten_frame to do this.
   utils._print_header('Flatten the Frame')
-  pd_frame = dc.flatten_frame(pd_frame)
+  pd_frame = pd_frame.explode('county')
   print(pd_frame)
 
   # Get the names for each city.
@@ -87,7 +87,7 @@ def main():
 
   # Format the final frame.
   utils._print_header('The Final Frame')
-  pd_frame = dc.flatten_frame(pd_frame)
+  pd_frame = pd_frame.explode('city')
   print(pd_frame)
 
 

datacommons/examples/places.py

@@ -53,7 +53,7 @@ def main():
   # Get all CensusTracts in these two counties.
   utils._print_header('Get Census Tracts')
   pd_frame['tracts'] = dc.get_places_in(pd_frame['county'], 'CensusTract')
-  pd_frame = dc.flatten_frame(pd_frame)
+  pd_frame = pd_frame.explode('tracts')
   print(pd_frame)
 
 

datacommons/examples/populations.py

@@ -57,16 +57,16 @@ def main():
   # DataFrame with Santa Clara and Montgomery County.
   utils._print_header('Initialize the DataFrame')
   pd_frame = pd.DataFrame({'state': ['geoId/06', 'geoId/21', 'geoId/24']})
-  pd_frame['state_name'] = dc.get_property_values(pd_frame['state'], 'name')
-  pd_frame = dc.flatten_frame(pd_frame)
-  print(pd_frame)
+  pd_frame['state_name'] = pd_frame['state'].map(
+    dc.get_property_values(pd_frame['state'], 'name'))
+  pd_frame = pd_frame.explode('state_name').reset_index(drop=True)
 
   # Get populations for employed individuals
   utils._print_header('Add Population and Observation to DataFrame')
-  pd_frame['employed_pop'] = dc.get_populations(
+  pd_frame['employed_pop'] = pd_frame['state'].map(dc.get_populations(
     pd_frame['state'],
     'Person',
-    constraining_properties={'employment': 'BLS_Employed'})
+    constraining_properties={'employment': 'BLS_Employed'}))
 
   # Add the observation for employed individuals
   pd_frame['employed_count'] = dc.get_observations(
@@ -81,7 +81,7 @@ def main():
   # Final dataframe. Use the convenience function "clean_frame" to convert
   # columns to numerical types.
   utils._print_header('Final Data Frame')
-  pd_frame = dc.clean_frame(pd_frame)
+  pd_frame = pd_frame.dropna().reset_index(drop=True)
   print(pd_frame)
 
 

datacommons/examples/query.py

@@ -21,7 +21,6 @@
 from __future__ import print_function
 
 import datacommons as dc
-import pandas as pd
 
 
 def main():
@@ -37,12 +36,9 @@ def main():
 ''')
   print('> Issuing query.\n{}'.format(query))
 
-  # Initialize the Query instance.
-  dc_query = dc.Query(sparql=query)
-
   # Iterate through all the rows in the results.
   print('> Printing results.\n')
-  for row in dc_query.rows():
+  for row in dc.query(query_string=query):
     print('  {}'.format(row))
 
 

datacommons/places.py

@@ -23,7 +23,6 @@
 from __future__ import print_function
 
 import datacommons.utils as utils
-import pandas as pd
 
 import requests
 
@@ -33,21 +32,13 @@ def get_places_in(dcids, place_type):
       :code:`place_type`.
 
   Args:
-    dcids (Union[:obj:`list` of :obj:`str`, :obj:`pandas.Series`]): Dcids to get
-      contained in places.
+    dcids (:obj:`iterable` of :obj:`str`): Dcids to get contained in places.
     place_type (:obj:`str`): The type of places contained in the given dcids to
       filter by.
 
   Returns:
-    When :code:`dcids` is an instance of :obj:`list`, the returned
-    :obj:`Place`'s are formatted as a :obj:`dict` from a given dcid to a list of
-    places identified by dcids of the given `place_type`.
-
-    When :code:`dcids` is an instance of :obj:`pandas.Series`, the returned
-    :obj:`Place`'s are formatted as a :obj:`pandas.Series` where the `i`-th
-    entry corresponds to places contained in the place identified by the dcid
-    in `i`-th cell if :code:`dcids`. The cells of the returned series will always
-    contain a :obj:`list` of place dcids of the given `place_type`.
+    The returned :obj:`Place`'s are formatted as a :obj:`dict` from a given
+    dcid to a list of places identified by dcids of the given `place_type`.
 
   Raises:
     ValueError: If the payload returned by the Data Commons REST API is
@@ -70,26 +61,14 @@ def get_places_in(dcids, place_type):
         # and 53 more
       ]
     }
-
-    We can also specify the :code:`dcids` as a :obj:`pandas.Series` like so.
-
-    >>> import pandas as pd
-    >>> dcids = pd.Series(["geoId/06"])
-    >>> get_places_in(dcids, "County")
-    0    [geoId/06041, geoId/06089, geoId/06015, geoId/...
-    dtype: object
-
   """
-  # Convert the dcids field and format the request to GetPlacesIn
-  dcids, req_dcids = utils._convert_dcids_type(dcids)
+  dcids = list(dcids)
   url = utils._API_ROOT + utils._API_ENDPOINTS['get_places_in']
   payload = utils._send_request(url, req_json={
-    'dcids': req_dcids,
+    'dcids': dcids,
     'place_type': place_type,
   })
 
   # Create the results and format it appropriately
   result = utils._format_expand_payload(payload, 'place', must_exist=dcids)
-  if isinstance(dcids, pd.Series):
-    return pd.Series([result[dcid] for dcid in dcids], index=dcids.index)
   return result