search

SIMPATICOProject / SimpaticoTAEServer

This is the official repository of the Text Adaptation Engine server of Simpatico.
GNU Lesser General Public License v3.0
7 stars 3 forks source link

readme

Content:

architecture: Folder containing a visual description of the architecture of the first version of the TAE server.
lexical_simplification_server: Folder containing the code for the local Lexical Simplification server of SIMPATICO.
syntactic_simplification_server: Folder containing the code for the local Syntactic Simplification server of SIMPATICO.
main_TAE_server: Folder containing the code for the TAE web server of SIMPATICO.
resources.txt: File containing a map of resources to be used by the simplification servers of SIMPATICO.
configurations.txt: File containing port configurations for the servers.

Installation instructions:

Local Variables

export PYTHONIOENCODING=utf8

Without docker

1) In order to run the servers, you will need a Python 2.6+ installation (preferably Anaconda), and a Java 1.8+ installation.

2) Install the following Python 2 libraries:

    kenlm
    gensim
    nltk
    sklearn
    keras
    numpy
    h5py

3) Download or clone the code from https://github.com/SIMPATICOProject/SimpaticoTAEServer onto a folder of choice (ex: /home/user/SimpaticoTAEServer).

4) Download the data pack from http://www.quest.dcs.shef.ac.uk/simpatico/simplifier_data.tar.gz and unpack it into the root folder of the code (ex: resulting in /home/user/SimpaticoTAEServer/data). These are the files referenced in the "resources.txt" file.

5) Download the Stanford Tagger (full) from http://nlp.stanford.edu/software/stanford-postagger-full-2015-04-20.zip and unpack it into a folder of choice (ex: /home/user/stanford-postagger-full-2015-04-20).

6) The version 3.7.0 of the Stanford CoreNLP is already available in the data pack. 
    - However, if you want to use a different version, you can download from http://stanfordnlp.github.io/CoreNLP/. In this case you will need to change the path to your CoreNLP version into the resources.txt file (corenlp_dir parameter).

With docker

  1. Clone the project
  2. Download archive from http://www.quest.dcs.shef.ac.uk/simpatico/simplifier_data.tar.gz and save into docker-data folder, uncompress the archive
  3. Download archive from http://nlp.stanford.edu/software/stanford-postagger-full-2015-04-20.zip and save into docker-data folder, uncompress the archive
  4. Create docker image sudo docker build -t tae .
  5. Create container sudo docker run -v <PROJECT_PATH>/docker-data:/app/data --name tae-cont tae

Running instructions:

1) In order to run a fully functional version of the SIMPATICO TAE server, you will have to run the following components:

    - The Stanford Tagger servers for English, Spanish, Galician and Italian: Receives requests from the Lexical Simplification local server for tagging.
    - The Stanford Dependency Parser servers: Receives requests from the Syntactic Simplification local server for parsing.
    - The Lexical Simplification local server: Receives requests from the main TAE web server for lexical simplifications.
    - The Syntactic Simplification local server: Receives requests from the main TAE web server for syntactic simplifications.
    - The main TAE web server: Receives requests from the web for Text Adaptation.

2) How to run the Stanford Tagger servers:

    Go to the folder where you unpacked the Stanford Tagger and run the following command:

        java -mx2G -cp "*:lib/*:models/*" edu.stanford.nlp.tagger.maxent.MaxentTaggerServer -model ./data/english.tagger -port 2020 &
        java -mx2G -cp "*:lib/*:models/*" edu.stanford.nlp.tagger.maxent.MaxentTaggerServer -model ./data/spanish.tagger -port 3030 &
        java -mx2G -cp "*:lib/*:models/*" edu.stanford.nlp.tagger.maxent.MaxentTaggerServer -model ./data/galician.tagger -port 4040 &
        java -mx2G -cp "*:lib/*:models/*" edu.stanford.nlp.tagger.maxent.MaxentTaggerServer -model ./data/italian.tagger -port 5050 &

    Then you will have a tagging servers running at ports 2020, 3030, 4040, and 5050. The ports chosen MUST be the ones specified on the "configurations.txt" file.

3) How to run the Stanford parser servers:

    These servers are run when the syntactic simplifier server is started, therefore, there is no need to run them externally.

4) How to run the Lexical Simplification local server:

    Navigate to the "lexical_simplification_server" folder, then run the following command:

        nohup python -u Run_TCP_Lexical_Simplifier_Server.py &

    Then you will have a server receiving requests at the "ls_local_server_port" port specified in the "configurations.txt" file.

5) How to run the Syntactic Simplification local server:

    Navigate to the "syntactic_simplification_server" folder, then run the following command:

        nohup python -u Run_TCP_Syntactic_Simplifier_Server.py &

    Then you will have a server receiving requests at the "ss_local_server_port" port specified in the "configurations.txt" file.

6) How to run the main TAE web server:

    Navigate to the "main_TAE_server" folder, then run the following command:

        nohup python -u Run_TAE_Simplification_Server.py &

    Then you will have a server receiving requests at the "main_tae_server_port" port specified in the "configurations.txt" file.

Testing instructions:

1) How to test the Stanford Tagger server:

    Go to the folder where you unpacked the Stanford Tagger and run the following command:

        java -cp stanford-postagger.jar edu.stanford.nlp.tagger.maxent.MaxentTaggerServer -client -port 2020

    Then just type a complete English sentence and press enter. The port chosen MUST be the one specified on the "configurations.txt" file.

2) How to test the Lexical Simplification local server:

    Navigate to the "lexical_simplification_server/tests" folder, then run the following commands:

        python English_Test_Simplifier.py
        python Spanish_Test_Simplifier.py
        python Galician_Test_Simplifier.py
        python Italian_Test_Simplifier.py

    You should receive responses with simplifications after a few seconds.
    You can also open the aforementioned scripts to see how they work.

3) How to test the Syntactic Simplification local server:
    Navigate to the "syntactic_simplification_server/tests" folder, then run the following commands:

        python English_Test_Simplifier.py
        python Galician_Test_Simplifier.py
        python Spanish_Test_Simplifier.py

    You should receive responses with simplifications after a few seconds.
    You can also open the aforementioned scripts to see how they work.

4) How to test the main TAE web server:

    Navigate to the "main_TAE_server/tests" folder, then run the following commands:

        python Lexical_English_Test.py
        python Lexical_Spanish_Test.py
        python Lexical_Galician_Test.py
        python Lexical_Italian_Test.py
        python Syntactic_English_Test.py
        python Syntactic_Galician_Test.py
        python Syntactic_Spanish_Test.py

    Then you will have a server receiving requests at the "main_tae_server_port" port specified in the "configurations.txt" file.
    You can also test it from a web browser through URLs such as:

        http://<server_ip>:<main_tae_server_port>/?type=lexical&target=unorthodox&sentence=His%20fighting%20technique%20is%20very%20unorthodox%20.&index=5&lang=en
        http://<server_ip>:<main_tae_server_port>/?type=lexical&target=collegamento&sentence=gestione%20dati%20del%20nucleo%20familiare%20in%20caso%20di%20casa%20famiglia%20(%20nucleo%20familiare%20)%20-%20da%20collegamento%20all%20anagrafe%20.&index=16&lang=it
        http://<server_ip>:<main_tae_server_port>/?type=lexical&target=medran&sentence=Os%20líquens%20que%20medran%20no%20residuo%20asfáltico%20suxiren%20a%20sua%20.&index=3&lang=spanish
        http://<server_ip>:<main_tae_server_port>/?type=syntactic&sentence=His%20fighting%20technique%20is%20very%20unorthodox%20.
        http://<server_ip>:<main_tae_server_port>/?type=syntactic&sentence=Os%20líquens%20que%20medran%20no%20residuo%20asfáltico%20suxiren%20a%20sua%20.