hatebase

Python Version of Andrew Welter's Hatebase Wrapper, based on DanielJDufour's implementation


Keywords
conflict, hatespeech, language, nlp, python, tagging, api
License
MIT
Install
pip install hatebase==1.0.3

Documentation

hatebase

Python Version of Andrew Welter's Hatebase Wrapper Using the current 4.2 Version of the Hatebase API

Install

pip install hatebase

Requirements

pip install requests

HatebaseAPI functions

Initialize HatebaseAPI class

from hatebase import HatebaseAPI
key = 'yourHatebaseAPIKeyString'
hatebase = HatebaseAPI({"key": key})
# for more details, set debug to True
hatebase = HatebaseAPI({"key": key, "debug": True})

HatebaseAPI getVocabulary

# set filters for vocabulary query
filters = {"language": "eng"}
format = "json"

response = hatebase.getVocabulary(filters=filters, format=format)

# get some details from response
vocablist = response["result"]
results = response["number_of_results"]
pages = response["number_of_pages"]

HatebaseAPI getVocabularyDetails

format = "json"
details_filters = {'vocabulary_id': vocab_id}

response = hatebase.getVocabularyDetails(filters=details_filters, format=format)

HatebaseAPI getSightings

filters = {'is_about_nationality': '1', 'language': 'eng', 'country_id': 'US'}
format = "json"

response = hatebase.getSightings(filters=filters, format=format)

HatebaseAPI analyze

# TBD

HatebaseAPI getAnalysis

# TBD

Examples

Get All the Hate Speech in English About Nationality in the US

import json
import requests
from hatebase import HatebaseAPI

hatebase = HatebaseAPI({"key": key})
filters = {'is_about_nationality': '1', 'language': 'eng', 'country_id': 'US'}
format = "json"
json_response = hatebase.getSightings(filters=filters, format=format)

Get All Arabic Vocabulary

from json 
import requests
import pandas as pd
from hatebase import HatebaseAPI

hatebase = HatebaseAPI({"key": key})
filters = {"language": "ara"}
format = "json"
# initialize list for all vocabulary entry dictionaries
ara_vocab = []
response = hatebase.getVocabulary(filters=filters, format=format)
pages = response["number_of_pages"]
# fill the vocabulary list with all entries of all pages
# this might take some time...
for page in range(1, pages+1):
    filters["page"] = str(page) 
    response = hatebase.getVocabulary(filters=filters, format=format)
    ara_vocab.append(response["result"])

# create empty pandas df for all vocabulary entries
df_ara_vocab = pd.DataFrame()
# fill df
for elem in ara_vocab:
    df_ara_vocab = df_ara_vocab.append(elem)
# reset the df index
df_ara_vocab.reset_index(drop=True, inplace=True)    

For more documentation on the API check out https://github.com/hatebase/Hatebase-API-Docs

Testing

To test the package run

python -m unittest hatebase.tests.test