This demonstrates of the use of the RedisGraph to store and analyze named entities (people, places, objects, etc.). The data collection involves applying a Named Entity Recognition (NER) model from SpaCy to a set of blog posts.

An example application uses Flask and provides you to the ability to navigate the blog posts via the cooccurrences of the named entityies. It also provides full text search on the graph and displays the collect of named entities from the search results. The application has a simple two-tier architecture where the Flask-based web application uses Cypher to directly query RedisGraph.

The graph is constructed from harvesting and analyzing blog post articles through a process separate from RedisGraph. The metadata about the articles and the article text are stored in RedisGraph. Separately, an NER model is applied to the content of the articles and the named entities extracted are also stored in the graph in relation to the articles.

Setup

The implementation is written in Python 3. To run the code, all you will need to do is create a python environment and install the required packages:

pip install -r requirements.txt

For a quick start, see the project documentation and for more detailed information, see below.

What you can do next

Collecting Data

See how articles are harvest and the NER model is applied. More

Data Architecture

See how the graph is constructed. More

Ingesting Data

All about how to ingest articles and named entities into RedisGraph. More

Querying Data

Understand how to use Cypher to query the named entities and articles. More

Demo Application

Understand how to run the demo web application. More

Serverless

You can deploy the client application via serverless More