README.md

tl-hermes

News Aggregations and Sentiment Analysis app

Using the hermes python module

Add export PYTHONPATH=/path/to/tl-hermes/python/ to your favourite *sh's configuration file. Then hermes can be used e.g. with

import hermes.firebaseInterface as fbi
from hermes.nappy.tools import Content
from hermes.api import theguardian

Structure of the hermify script

1. Firebase + NaturalLanguage app/client setup
2. Create day tag, e.g. "2018-12-11"
3. Get list of childnodes of '/' in RTDB and create/reference the nodeToday
4. Make an API call + parse the response into dict/json-like object
5. Create Content instance & do sentiment analysis
6. Deduplication -> check if content of response is in RTDB and if not:
   1. Push content to RTDB (returns the randomly generated node key)
   2. Send a POST request to a cloud function, which extracts the newly generated content from RTDB and pushes it into BigQuery

How to use the hermify script

1. Create a folder api_configs in the same directory as hermes.py.
2. Place an API config like the example config in python/hermes/api/guardian_example_settings.json in api_configs and adapt the contents accordingly (side note: empty strings for one or both of the date sections will result in the script just requesting today's data.
3. call the script with python hermify.py (python 3 of course)

Dependencies

Python==3.7.1 For packages, please refer to the requirements.txt. Install them via pip install -r requirements.txt.