Yhteystiedot:

Osoite:
Yleisen kielitieteen laitos
PL 9
Siltavuorenpenger 20A
00014 Helsingin yliopisto

Puhelin: +358 (09) 1911 (vaihde)
Faksi: +358 (09) 191 29307

Kurssin materiaalit

1. Installing open source software.

1.1. Installing Xerces.

• Lecture notes: Apache Software Foundation
  • Old IBM tutorials: Installing Xerces, Installing Xalan
• Optional:
  • Install Xerces in your own directory and set your CLASSPATH.
  • Use Xerces SAX sample Counter.java to parse the file message.xml.

1.2. Installing jEdit.

• Lecture notes: jEdit.
  • Articles: jEdit (Wikipedia), Installing jEdit (About.com).
• Practical work:
  • Install jEdit in your own directory.
  • Learn about jEdit by studying the jEdit help menu.

2. Reusing published examples.

2.1. Basic XSLT processing with Java.

• Lecture notes: Java and XSLT by Eric Burke (O'Reilly)
• Optional:
  • Use SimpleJaxp.java to transform message.xml with message.xslt.

2.2. Transforming CSV spreadsheet files to HTML.

• Lecture notes: Custom SAX parser described in Java and XSLT
  • AbstractXMLReader.java, CSVXMLReader.java, SimpleCSVProcessor.java.
• Practical work:
  • Use SimpleCSVProcessor.java to transform presidents.csv with csvToHTMLTable.xslt.
• Assignment 1: Transforming CSV spreadsheet files to XML.

3. Building software with Ant.

3.1. Installing Ant

• Lecture notes: Apache Ant project
  • Installing Ant
• Practical work:
  • Install Ant in your own directory.
  • Set ANT_HOME to the location where you install it.
    export PATH=$PATH:$ANT_HOME/bin

3.2. Using Ant to compile and execute.

• Lecture notes: Ant manual
  • IBM tutorial: Apache Ant 101
• Practical work:
  • Write a simple buildfile build.xml with a target "compile" that compiles SimpleJaxp.java. Run the target by "ant compile".
  • Add a target "simplejaxp" that executes SimpleJaxp with two files (arguments) message.xml and message.xslt.
• Assignment 2: Building software with Ant.

4. Open source speech technology.

4.1. Java Speech API.

• Lecture notes: Java Speech API Programmer's Guide (Sun)
  • Java Speech Markup Language (JSML)

4.2. Installing FreeTTS.

• Lecture notes: FreeTTS
• Practical work (on your own workstation using earphones):
  • Install FreeTTS in your own directory.
  • Set up the Java Speech API.
  • Run the JSAPI Demo Player. Experiment with pitch and speed.

4.3. Optional: Non-Java open source speech tools.

• Festival speech synthesizer (Edinburgh)
• Flite synthesizer and FestVox voices (CMU)
• Sphinx speech recognition (CMU)
  Update: The new version Sphinx-4 is written in Java.

5. Open source language technology.

5.1. Open source NLP tools.

• Lecture notes: Natural Language Software Registry (DFKI)
  • Open NLP (Edinburgh), NLP Farm (Linköping)
• Practical work. Can you find:
  • a free morphology analyser for English? ... for Finnish?
  • a free chart parser? a free sentence generator? ... in Java?

5.2. Installing OpenCCG.

• Lecture notes: OpenCCG
• Practical work:
  • Download OpenCCG version 0.8.5 (not 0.8.6) to your own directory.
  • Follow the instructions in README about setting environment variables, building OpenCCG using Ant, and finally "Trying it out".
  • Try: tccg> the teacher bought the policeman a book

6. Testing with JUnit.

6.1. Installing JUnit.

• Lecture notes: JUnit.org
• Practical work:
  • Install JUnit in your own directory and set JUNIT_HOME.
    Add junit.jar to your CLASSPATH, for example:
    export CLASSPATH=.:$JUNIT_HOME/junit.jar
  • Installation test: cd $JUNIT_HOME
    All on one line: java junit.swingui.TestRunner junit.samples.AllTests

6.2. Testing with JUnit.

• Lecture notes: JUnit Primer by Mike Clark
  • Test Infected: Programmers love writing tests by Erich Gamma & Kent Beck
  • Incremental Development with Ant and JUnit (IBM)
• Assignment 3: Testing with Ant and JUnit.

6.3. Optional: Extreme Programming.

• What is Extreme Programming?
• Java Extreme Programming Cookbook by Eric Burke and Brian Coyner
  • Chapter 4: JUnit

7. Version control: RCS and CVS.

7.1. RCS and CVS.

• Lecture notes: CVS: Concurrent Versions System
  • Basic Source Control using RCS by Bolinger & Bronson (O'Reilly)
• Practical work:
  • In your Assignment 2 directory, make a new subdirectory "RCS".
    Checkin your Ant buildfile, with description "Assignment 2 buildfile".
  • Checkout the buildfile. Rename targets "prepare", "clean" to "init", "tidy".
    View the change by: rcsdiff build.xml
  • Checkin the buildfile, with log message "Renamed targets".
    View the file's history by: rlog build.xml

7.2. Downloading open source software using CVS.

• Lecture notes: Midiki: Mitre Dialogue Kit
• Practical work: Download the latest version of Midiki using CVS.
  • Execute this CVS command (all on one line): cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/midiki checkout -P midiki
  • Use Ant to build Midiki: cd midiki; ant build-midiki
  • Run Midiki (all on one line): java -cp target/midiki-0.1.4.jar org.mitre.dm.qud.domain.diagnosis.diagnosis_dm

8. Developing with IDEs.

8.1. Integrated Development Environments (IDEs).

• Lecture notes: JBuilder development environment (click on >> button)
  • JBuilder proprietary IDE with a free version (Borland)
  • NetBeans open source IDE (Sun)
  • Eclipse open source IDE (IBM)

8.2. Avoiding problems when using IDEs.

• Avoiding "IDE dependency": IDE Developer's Guide (Apache Jakarta)
• Avoiding "IDE lock-in": Top 15 Ant Best practices by Eric Burke (O'Reilly)
• Optional: 10 Reasons to dump your Java IDE by Brad O'Hearne -- and a reply.

8.3. Make your own IDE by extending jEdit.

• Read about installing and using plugins and window docking in jEdit.
• Optional: Install the AntFarm plugin and use it to redo Assignment 2.
• Optional: Install the JUnit plugin and use it to redo Assignment 3.

9. Open source databases.

9.1. Relational database systems.

• Lecture notes: Introducing Oracle (O'Reilly)
  • MySQL "The world's most popular open source database"
  • PostgreSQL "The world's most advanced open source database"

9.2. Structured Query Language (SQL).

• Lecture notes:
  • Interactive introductory tutorial: SQLcourse.com
  • SQL queries from the command line: MySQL tutorial
  • Optional: Chapter 3: Query Primer from Learning SQL (O'Reilly)
• Practical work:
  • Use MySQL on venus: venus$ mysql ...
    Use the test database: mysql> use test; ...
  • Create a table for the LiTeR staff information.
    Call it username_liter (with your own username).
  • Insert rows into the table for the four LiTeR staff members.
    Practise selecting different data.

10. Java and SQL.

10.1. Java Database Connectivity (JDBC).

• Lecture notes:
  • Chapter 2: JDBC from Java Enterprise in a Nutshell (O'Reilly)
  • Open source JDBC driver for MySQL: MySQL Connector/J
• Practical work:
  • Install Connector/J in your own directory and set your CLASSPATH.

10.2. Java, SQL and XML.

• Lecture notes:
  • IBM tutorial: Using JDBC to insert data from XML into a database
  • IBM tutorial: Using JDBC to extract data into XML
• Assignment 4: Transforming SQL database tables to XML.

11. WordNet: a lexical database.

11.1. Princeton WordNet.

• Lecture notes: WordNet
  • Introduction to WordNet by George Miller et al. (Princeton)
• Practical work:
  • Start the WordNet browser on venus: venus$ wnb &
  • Nouns: Query hyponyms of "student": what kind of student are you?
    Compare hypernyms of "student" and "professor": what's the difference?
    Compare hyponyms of "professor" and "lecturer": is WordNet US English?
  • Adjectives: Compare "big", "large, "great". What are their antonyms?
    Which combinations of "big/large/great sister/uncle/toe" are collocations?

11.2. Other WordNets.

• List maintained by Global WordNet Association includes:
  • Estonian WordNet (University of Tartu) (see also EuroWordNet)
  • RussNet WordNet for Russian (St Petersburg State University)
• Commercial products (Oracle database, Java software) include:
  • Russian WordNet (Petersburg State Transport University)

12. WordNet, SQL and Java.

12.1. WordNet SQL databases.

• Lecture notes: WordNet: Related projects
  • Open source WordNet SQL builder for MySQL and PostgreSQL

12.2. Java API for WordNet.

• Practical work: Open source JWNL (Java WordNet Library)
  • Install JWNL in your own directory and set your CLASSPATH
    (add jwnl.jar, utilities.jar, commons-logging.jar).
  • Edit file_properties.xml to specify the location of WordNet:
    <param name="dictionary_path" value="/usr/share/wordnet"/>
  • Compile Examples.java and run it:
    java net.didion.jwnl.utilities.Examples file_properties.xml
• Assignment 5: Accessing WordNet from Java.

13. GATE: an IDE for NLP.

13.1. GATE.

• Lecture notes: GATE and the GATE graphical development environment
  • GATE 2.2 User Guide by Hamish Cunningham et al. (Sheffield)
• Practical work:
  • Copy the shell script gate-2.2 to your directory and make it executable.
    Start GATE 2.2 by: venus$ gate-2.2 &
  • Copy Shakespeare's sonnet130.xml and sonnet.dtd to your directory.
    Load sonnet130.xml into GATEas document Sonnet130 (see Creating documents) and view the document contents.

13.2. ANNIE.

• Lecture notes: ANNIE: A Nearly-New Information Extraction system
• Practical work:
  • Create a GATE corpus Sonnets (see Creating corpora) with Sonnet130 in it.
  • Load ANNIE (see ANNIE - a ready-made information extraction system for English) and run ANNIE on the Sonnets corpus.
  • View Sonnet130 (see Inspecting the processing results).
    Check original markups of the XML document and new markups by ANNIE.
    Surprises? <JobTitle>: mistress, <Unknown>: Shakespeare, sonnet, heaven.
  • Use WordNet to get the hypernyms and hyponyms of the "unknown" words Shakespeare, sonnet, heaven. Save the results in a plain text file.
• Assignment 6: Annotating a corpus with GATE

14. GATE, Java and WordNet.

14.1. GATE data stores.

• Lecture notes: Serial data stores in GATE
  • Optional: Database setup for GATE (Oracle or PostgreSQL)
• Practical work:
  • Run GATE and create a serial data store in your directory. Load Sonnet130 and save it as XML in the data store.
  • Run ANNIE and save Sonnet130 as XML again (with a new name).
  • Problem solving: When I ran Sentence Splitter on Sonnet130, it went on and on without finishing. How did I find out what was wrong? (Hint and Answer)

14.2. GATE, Java and WordNet.

• Lecture notes:
  • Using GATE components in Java: GATE example code
  • Using WordNet 1.6 inside GATE 2.2: WordNet in GATE
  • GATE has its own WordNet API based on JWNL. Example: WordNetApp.java
  • GATE was developed with JUnit testing. Example: TestWordNet.java
• Assignment 7: Using GATE and WordNet from Java
• Assignment 8: Using GATE and ANNIE from Java
• Assignment 9: Installing an open source spoken dialogue system