Personalized Search Engine

Implementation of a personalized search engine using Elastic Search on tweets.

A demo using Angular has also been built.

Group

• Christian Bernasconi 816423
• Marco Ripamonti 806785

Description

Tweets from five different 5 people of two different domains have been gathered using Tweepy. Those users are then used to build five custom user profiles so that searches between users are different from one another.

Elastic Search has been used to build the search engine.

Setup environment

To setup the python environment run the following commands:

python -m venv venv 
.\venv\scripts\activate (WINDOWS)
.\venv\bin\activate (LINUX)
pip install -r requirements.txt

Index creation and search

To test the implemented functionalities run the following command:

python -m project

Be sure to have Elasticsearch running on http://localhost:9200 . N.B.: if you want to create a new index you need to specify it in ./create_index_conf.py. An example is provided in ./create_index_conf_sample.py

Demo

First of all you have to copy ./elasticsearch.yml to ../elasticsearch-version/config/ to enable cross-origin requests. Then, you have to execute the following commands:

cd ./demo
npm install
npm run start

The demo should be available at http://localhost:4200 (be sure to have Elasticsearch running on http://localhost:9200)

Get tweets

Tweets are already provided in ./data. You can download new tweets executing the command:

python -m getTweets

Please notice: to use the twitter API you need to put your KEYS in ./secrets.py. An example is provided in ./secrets_sample.py