django search example

django-search-example

This project is an example for using Django's full text search based on PostgreSQL.

The version found in the main branch provides an application with a search form. The actual search uses Django's icontains which on an SQL level maps to like while ignoring the case of letters. This works only for simple search requests and is very inefficient.

The search functionality is gradually improved in additional feature branches:

Project set up

Before you can run the application in the main branch you need to install:

1. Python 3.8+
2. Docker (used to run a PostgreSQL database server)

Next start downloading some ebooks:

sh scripts/rsync_gutenberg.sh

Downloading the entire library of Project Gutenberg will take hours. But for testing this script only needs to run for a couple of minutes until it downloaded a few dozen or hundreds of ebooks.

You can leave this running in the background in case you want a larger amount of ebooks eventually.

Next, launch the docker container providing the PostgreSQL database server:

docker-compose up

While the server runs, you cannot enter new commands in the terminal, so open a new one for the commands to follow. (Alternatively you could add the option --detach to the above call.)

After that, set up the poetry environment and pre-commit hook:

sh scripts/setup_project.sh

Once this is finished, open a poetry shell:

poetry shell

Any further commands should be run in this shell.

Next, set up the database:

sh scripts/reset_local_database.sh

This creates or clears the database and loads a few ebooks into it.

Running the local development server

Finally, you can run the local development server. Optionally you can specify the port to avoid clashes with existing services:

python manage.py runserver 8078

To search, navigate to http://127.0.0.1:8078/ and enter a single word search term, for example "house". (Later branches allow more sophisticated searches).

To browse the documents available, navigate to http://127.0.0.1:8078/admin/gutensearch/document/. For the login, use admin as username and deMo.123 as password.

Learning text search

After that, open the slides stored in Full text search with Django and PostgreSQL.odp and work your way through them. For each new feature introduced there is an educational pull request showing the code needed to implement it (based on the previous feature branch).