elhawlir / devsearch

A web search engine built with Python which uses TF-IDF and PageRank to sort search results.

devsearch

A web search engine built with Python which uses TF-IDF and PageRank to sort search results.

Stack

• Flask (Python 3)
• Scrapy
• LXML
• MongoEngine (MongoDB)
• Bootstrap 4

Requirements

• Docker
• Docker Compose

Setup

• Install Docker and Docker Compose.
• Clone or download this repo.
• Create a .env file from .env.example.
• Run docker-compose up.

Crawling

• Update the SPIDER_ALLOWED_DOMAINS variable in .env with domains you want the spider to crawl.
• Add at least one url to the crawl_list collection (in MongoDB) for the spider to start with.
• Run docker-compose run web flask crawl to crawl new web pages.
• You can add the --recrawl option to update pages already crawled: docker-compose run web flask crawl --recrawl True.

Indexing

• To index crawled pages, run docker-compose run web flask index.
• To compute TFIDF, run the following one after the other:
  • docker-compose run web flask idf
  • docker-compose run web flask tfidf
• To compute PageRank, run docker-compose run web flask rank.
• To compute page-word score, run docker-compose run web flask score.

Deploy

• Create a .env.secret file from .env.secret.example.
• Run docker-compose -f docker-compose.prod.yml up --build -d.