Document Retrieval

Python implementation of Document Retrieval System

Created by Gerardo Roa Dabike

Oct 2015

USE: python (options)

OPTIONS: Help : -h : print this help message and exit

PARAMETERS/BEHAVIOURS:

-c FILE : FILE as a COLLECTION_FILE. 
-q FILE : FILE as a QUERIE_FILE
-p INT : INT is the id of one query inside the QUERIE_FILE. This is a filter of "-q" option, and works together.
-t STRING : STRING is a string of query words separated each by full stop ".". This option cannot be together with "-q" option. 
-i FILE : FILE as a name of INDEX_FILE. If it is not specify one, the system will assume "index.txt" by default. This option only work with "-c" option declared.
-I : the file INDEX_FILE must be created before use it.
-r FILE : FILE as a RESULT_FILE. If not present but option "B" or "-F" or "-D" are declared, result is printed in screen. 
-a INT : INT as a length of output ranking. By default the value is 10.
-s FILE : FILE as a STOPLIST_FILE
-S : use PorterStemmer. 
-L : use LancasterStemmer.

If none options, "-S" or "-L", are declared, system will not use stemmed process.

If both options, "-S" and "-L", are declared, system use PorterStemmer.

-T : apply TreebankWordTokenizer. The system work with WordPunctTokenizer by default. 
-B : term weight in vector are BINARY.
-F : term weight in vector are TERM FREQUENCY.
-D : term weight in vector are TF.IDF.

Only one of the options "-B", "-F" and "-D" can be declared.

EXAMPLES:

1. Create default index of collection COLLECTIONFILE without stop list, using PorterStemmer and WordPunctTokenizer python -c COLLECTIONFILE -I -S

2. Create index INDEXFILE without stop list, using LancasterStemmer and WordPunctTokenizer python -c COLLECTIONFILE -i INDEXFILE -I -L

3. Retrieve Document using QUERIEFILE, a exist INDEXFILE, saving in OUTPUTFILE using STOPLISTFILE, PorterStemmer, TreebankWordTokenizer and Term Frequency Aproach python -q QUERIEFILE -i INDEXFILE -r OUTPUTFILE -s STOPLISTFILE -S -T -F