rsusik / magetdc

MAG on ETDC

MAG on ETDC

This is MAG variant that searches patterns in ETDC encoded text. This source code was written for research purpose and has a minimal error checking. The code may be not very readable and comments may not be adequate. There is no warranty, your use of this code is at your own risk.

Note: Please refer to following link for source codes of MAG algorithm: https://github.com/rsusik/mag

Requirements

• C++ compiler compatible
• Unix based 64-bit OS (compilation works also in Cygwin)
• Python 3 (for testing)
• Docker (optionally)

Compilation

To compile the code run below line commands:

git clone https://github.com/rsusik/mag_etdc mag_etdc
cd mag_etdc
make all

Running (MANUALLY)

Note: it is recomended to run tests using python script etdc_test.py which is described in next section (testing)

./filename [enc_patt] [m], [enc_set], [u], [k], [q], [sig], [denom], [set]

where:

• filename - one of compiled executable (e.g. etdc_mag_l2)
• enc_patt - /path/to encoded patterns
• enc_set - /path/to encoded set (e.g. /path/to/english.200MB)
• m - pattern length
• u - FAOSO param
• k - FAOSO param
• q - q-gram size
• sig - alphabet size
• denom - specifies the offset of index tag (n/demon)
• set - file set (used only for other file names generation, e.g. dictionary)

example:

./path/to/etdc_mag /path/to/english.200MB.processed.100.32.patt.bin 32 /path/to/english.200MB.processed.etdc.bin 4 2 3 4 100 /path/to/english.200MB

The enc_patt and enc_set can be generated by running etdc_tool executabe:

etdc_tool [path/to/text/set] [pnumber]

where:

• path/to/text/set - is a path to corpus (file previousely processed with etdc_parse_text.py script)
• pnumber - is a number of patterns to be generated

The program generates following files:

• [text_filename].processed.dict - dictionary
• [text_filename].processed.etdc.bin - encoded set
• [text_filename].processed.[p_number].[m].bin - encoded patterns
• [text_filename].processed.[p_number].[m].txt - patterns in plain text

The file [text_filename].processed should be generated with for example script etdc_parse_text.py:

python etdc\_parse\_text.py -c english.200MB

Quick Steps

1. Compile source codes (see Compilation section).
2. Parse the input file. 2.1. Replace set_loc value in file etdc_parse_text.py so it points to location where the sets are stored (e.g. english.200MB). 2.2. Execute etdc_parse_text.py (see Running section).
3. Encode the text file and generate patterns. 3.1. Execute etdc_tool (see Running section).
4. Execute tests. 4.1. Replace parameters in test script etdc_test.py. 4.1.1. set_loc - location of files generated by etdc_tool. 4.1.2. alg_loc - location of executables. 4.2. Execute the tests (see Testing section).

Testing

There is a python script that helps in testing the performance of given algorithms. The script downloads automaticaly Pizza&Chilli corpus and generates the patterns, etc.

To execute a test run following command:

usage: etdc_test.py [-h] [-r R] [-a A] [-c C] [-m M] [-u U] [-k K] [-q Q] [-s S] [-d D]

optional arguments:
  -h, --help           show this help message and exit
  -r R, --npatterns R  number of patterns
  -a A, --algorithm A  algorithm[s] to be tested
  -c C, --corpus C     corpus
  -m M, --length M     pattern length[s] (e.g. 8,16,32)
  -u U, --faosou U     FAOSO parameter U
  -k K, --faosok K     FAOSO parameter k
  -q Q, --q-gram Q     q-gram size
  -s S, --sigma S      dest. alph. size
  -d D, --denom D      denominator (e.g. 100,1000)

Example:

python3 etdc_test.py -a etdc_mag,etdc_mag_l2 -c english.100MB -r 100,1000 -m 8,32 -q 2,3,4,5,6,7,8,9,10 -s 5 -u 4 -k 2 -d 100,10000,1000000

The above script may execute algorithms with incorrect parameters (the correctness only depends on you), and thus generate some errors in the output which can be ignored.

Docker

The simplest way you can test the algorithm is by using docker. All you need to do is to:

• Clone the git repository:

git clone https://github.com/rsusik/magetdc magetdc
cd magetdc

• Build the image:

docker build -t magetdc .

• Or pull from repository:

docker pull rsusik/magetdc
docker tag docker.io/rsusik/magetdc magetdc

• Run container:

docker run --rm magetdc

• Additionally you may add parameters (as mentioned in Testing section):

docker run --rm magetdc -c english.100MB -m 16 -r 100

Authors

• Robert Susik
• Szymon Grabowski
• Kimmo Fredriksson