Content Based Document Retrieval System
Given a query, it retrieves the set of documents relevant to the submitted query using TF, TFIDF and BM25 term weighting measures. Appropriate snippets from the documents are also displayed.
pip install -r requirements.txt
Set the following values in index.py:
schema Indexing with or without stop-word removal, with or without term stemming
INDEX_NAME Name your index
PATH_TO_DATASET Path to dataset
Run python index.py to build the index.
Set index path in server.py and run python server.py to start the server.
/query/:query_string Takes page number, page length and weighting (tf, tfidf or bm25) as parameters and returns a JSON containing various fields including query time, number of results and an array of results
/log Logs the query
/precision Calculates mean average precision
/ndcg Calculates normalized discounted cumulative gain