ReadMe

This document will guide you to setup and run the system on your PC.
The folder in which you find this Readme you will be able to see the following folder structure

1> AllResults -> all the generated results along with Precision and Recall along with Metrics such as MAP and MRR for each indexing system
2> WebEngine -> this is the folder structure in which the project runs and all the source code
3> Report -> the Analysis report of the project
4> Lucene -> the lucene retrival model with source
5> ReadMe



The project uses libraries such as 
1> BeautifulSoup
2> NLTK - natural language text processing libraries


The WebEngine project is the one which is setup for the proejct to run,
The output of one stage is given as the input to the next stage.

step 1: The Parser is used to tokenize the raw corpus which is in the folder "cacm" and produce tokens in folder "tokens"
step 2: the "tokens" folder is taken as input for next part where the BM25, TFIDF and Smoothed Query Likelyhood model systems read tokenize and 
		output to respective folders the results for each queries "bm25Results", "tfidfResults", "SmoothQResults" and also this system produces 
		results using pseudo relevance feedback on BM25 system and produces results in "PseudoRelBM25Ranking" folder
step 3: The same tokens are taken as input for the Task-2 also which produces stopping and stemmed results for all three systems which are "BM25", "TFIDF" and "smoothedQuerylikelyhood model"
		All the results are produced in "Task2" folder 
step 4: The Snippet generator is in phase II folder which takes the input form the "bm25forSnippet" folder and produces snippets results for each query.
step 5: the Evaluaton code is "metrics.py", the folder "metrics/input" takes in results for all queries for each system and then produces precision and 
		recall in for each query and also MAP and MRR are saved in Stats.txt file.
step 6: Phase II folder contains code for snippet generator

Extra Credit:

step 6: The noiseGenerator.py is used to generate the noise queries and are saved in folder "noiceQueries"
step 7: The softQueryEngine.py reads the noise queries and does soft matching and produces output for BM25 indexing model in folder "noiceQueries/Results"