Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time
section x-masysma-name title date lang author keywords x-masysma-version x-masysma-repository x-masysma-website x-masysma-owned x-masysma-copyright
UNI DKE U11/5: Information Retrieval
2016/09/11 22:09:55
Linux-Fan, (
1 Quick Information Retrieval Version Coypright (c) 2016 For further info send an e-mail to


This “Quick Information Retrieval”-System (short: QuickIR) is a program to query a set of text documents using the common TF-IDF and cosine similarity. This allows a keyword-based search which takes into account that certain words which appear in a lot of documents are less relevant.

Not to raise any wrong expectations: It is called Quick because it was written quickly, not because it runs quickly! The code has been written to be very changeable, that is: there is large freedom of choosing an algorithm for simple ranking, e. g. cosine similarity, dot product similarity etc. and on the other hand, there are a lot of function calls which could be optimized away¹ (or “precomputed”) if this feature was not needed.

This was designed as a solution to an exercise presented at the TU Darmstadt's DKE course in summer 2016.


Compile this as follows (using any Java 7 compiler):

javac -encoding UTF-8 $(find ma -type f -name '*.java')

Alternatively, it is possible to compile by using $ ant or $ ant jar.


Invoke the program as follows

java ma.quickir.Main <arguments>

where arguments is a specfic commandline.

Currently, only <arguments> of the format TEXT <query> are supported.

Also, in order to tell the application which files to consult for “search indexing” (will be wholly in-memory and re-generated for every request) you have to give a list of file names via STDIN.


find /usr/share/common-licenses -type f | \
	java -cp minitools:. ma.quickir.Main TEXT "Apache"

will output (on my Debian Jessie system):

/usr/share/common-licenses/Apache-2.0 (0.10426607)

Thus, this query yielded only one result. Let's try to confuse the search engine with adding a few trivial words:

find /usr/share/common-licenses -type f | \
	java -cp minitools:. ma.quickir.Main TEXT "Apache and to that"

will output (formatted in a tabular manner for enhanced readability in the documentation):

Apache-2.0 (0.13838476) LGPL-3 (0.05926972) GPL-1 (0.057670876) Artistic (0.053396318) GPL-2 (0.052671414) LGPL-2 (0.047885068) LGPL-2.1 (0.04646213) GFDL-1.2 (0.041788474) GFDL-1.3 (0.039808597) BSD (0.03747116) GPL-3 (0.03579121)

Now we have more results but sill the Apache license is found first, although it is not that it contains many of the words “and”, “to” and “that” compared with the other licenses which we can find out by removing “Apache” from the list of words to be searched for:

find /usr/share/common-licenses -type f | \
	java -cp minitools:. ma.quickir.Main TEXT "and to that"

will output (also formatted for enhanced readability in the documentation):

LGPL-3 (0.13050869) GPL-1 (0.12698813) Artistic (0.11757579) GPL-2 (0.11597958) LGPL-2 (0.10544031) LGPL-2.1 (0.10230709) Apache-2.0 (0.100169055) GFDL-1.2 (0.09201595) GFDL-1.3 (0.08765637) BSD (0.08250946) GPL-3 (0.078810304)

Of course, you are invited to try the program on other interesting texts :)


As this was created for learning purposes, it's features are only rudimentary. Especially, this search is

  • fully case-sensitive
  • unaware of any punctuation
  • insufficient at determining word boundaries (a “new word” is whenever a character not in [A-Za-z_] occurs)
  • slow (see “Introduction” for details)
  • single-threaded (This problem really calls for parallelization but for now, this is not implemented).

¹) Using better design, it may be possible to combine both aspects but that is not currently implemented.


This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see

This program is licensed under GPLv3. The full text can be found under minitools/ma/tools2/license.txt.


An example of how to write a very simple Information Retrieval system for text files.






No releases published


No packages published