Zach Kagan Zach Kagan -3 years ago 100
JSON Question

How can I JSON serialize an object from google's natural language API? (No __dict__ attribute)

I'm using the Google Natural Language API for a project tagging text with sentiment analysis. I want to store my NL results as JSON. If a direct HTTP request is made to Google then a JSON response is returned.

However when using the provided Python libraries an object is returned instead, and that object is not directly JSON serializable.

Here is a sample of my code:

import os
import sys
import oauth2client.client
from import enums, language_service_client
from import language_service_pb2

class LanguageReader:
# class that parses, stores and reports language data from text

def __init__(self, content=None):

# attempts to autheticate credentials from env variable
except oauth2client.client.ApplicationDefaultCredentialsError:
print("=== ERROR: Google credentials could not be authenticated! ===")
print("Current enviroment variable for this process is: {}".format(os.environ['GOOGLE_APPLICATION_CREDENTIALS']))
print("to set the authentication credentials manually")

self.language_client = language_service_client.LanguageServiceClient()
self.document = language_service_pb2.Document()
self.document.type = enums.Document.Type.PLAIN_TEXT
self.encoding = enums.EncodingType.UTF32

self.results = None

if content is not None:

def read_content(self, content):
self.document.content = content
self.language_client.analyze_sentiment(self.document, self.encoding)
self.results = self.language_client.analyze_sentiment(self.document, self.encoding)

Now if you were to run:

sample_text="I love R&B music. Marvin Gaye is the best. 'What's Going On' is one of my favorite songs. It was so sad when Marvin Gaye died."
resp = LanguageReader(sample_text).results
print resp

You would get:

document_sentiment {
magnitude: 2.40000009537
score: 0.40000000596
language: "en"
sentences {
text {
content: "I love R&B music."
sentiment {
magnitude: 0.800000011921
score: 0.800000011921
sentences {
text {
content: "Marvin Gaye is the best."
begin_offset: 18
sentiment {
magnitude: 0.800000011921
score: 0.800000011921
sentences {
text {
content: "\'What\'s Going On\' is one of my favorite songs."
begin_offset: 43
sentiment {
magnitude: 0.40000000596
score: 0.40000000596
sentences {
text {
content: "It was so sad when Marvin Gaye died."
begin_offset: 90
sentiment {
magnitude: 0.20000000298
score: -0.20000000298

Which is not JSON. It's an instance of the object. And it has no __dict__ attribute attribute so it is not serializable by using json.dumps().

How can I either specify that the response should be in JSON or serialize the object to JSON?

Answer Source

Edit: @Zach noted Google's protobuf Data Interchange Format. It seems the preferred option would be to use these protobuf.json_format methods:

from google.protobuf.json_format import MessageToDict, MessageToJson 

self.dict = MessageToDict(self.results)
self.json = MessageToJson(self.results)

From the docstring:

MessageToJson(message, including_default_value_fields=False, preserving_proto_field_name=False)
    Converts protobuf message to JSON format.

      message: The protocol buffers message instance to serialize.
      including_default_value_fields: If True, singular primitive fields,
          repeated fields, and map fields will always be serialized.  If
          False, only serialize non-empty fields.  Singular message fields
          and oneof fields are not affected by this option.
      preserving_proto_field_name: If True, use the original proto field
          names as defined in the .proto file. If False, convert the field
          names to lowerCamelCase.

      A string containing the JSON formatted protocol buffer message.
Recommended from our users: Dynamic Network Monitoring from WhatsUp Gold from IPSwitch. Free Download