Windows 7 64 java envirenment build
0 likes594 views
This document provides instructions for setting up the Berkeley parser on a Windows 7 64-bit system with a Java environment. It describes downloading the Java Development Kit (JDK), installing it, configuring environment variables, testing with a Hello World program, downloading the Berkeley parser files and an English grammar file, preparing a sample input file, and running the parser on the sample file to generate output. It also provides commands for training an English grammar on the WSJ corpus and testing the trained grammar.
![Run Berkeley parser on Windows 7 (64bit) & java environment building
recorded by Aaron
[http://www.linkedin.com/in/aaronhan]
1. Download JDK, JDK is a development environment for building
applications, applets, and components using the Java programming
language, from
“http://www.oracle.com/technetwork/java/javase/downloads/jdk7downloads-1880260.html”
2. Install “jdk-7u45-windows-x64.exe”
3. Configure the environment variables for windows 7 (64-bit).
[reference:
http://blog.163.com/wutianshui@126/blog/static/1869346220099455
115417/]
Go to your computer -> property - > higher system setting -> environment
variable](https://image.slidesharecdn.com/windows7-64javaenvirenmentbuild-131017083415-phpapp01/85/Windows-7-64-java-envirenment-build-1-320.jpg)
![public static void main(String[] args) {
System.out.println("Hello, World");
}
}
----------------------------------------------Open the cmd window, type the following code:
cd desktop
java HelloWorld
If it shows the following words, succeed!
Begin the new life of java programming?
5. Download Berkeley parser
(https://code.google.com/p/berkeleyparser/downloads/list).
Download parser file “BerkeleyParser-1.7.jar”
Download English grammar file “eng_sm6.gr”
6. Prepare segmented English corpus.
In the same directory of the downloaded tools, build an English
example file named as “sampleenglish.seg”, and put this sentence
“this is a small house .” into this file.](https://image.slidesharecdn.com/windows7-64javaenvirenmentbuild-131017083415-phpapp01/85/Windows-7-64-java-envirenment-build-6-320.jpg)
![Training began from 2013-11-10-16:30; finished at
Store the running details of training grammar by the command
“command…. > store.file.name” after training finished.
=====================================
Testing:
[[To the test the performance of a grammar you can use:
java -cp berkeleyParser.jar edu.berkeley.nlp.PCFGLA.GrammarTester path <WSJ location> -in <grammar-file>]]
the testing command of the trained grammar:](https://image.slidesharecdn.com/windows7-64javaenvirenmentbuild-131017083415-phpapp01/85/Windows-7-64-java-envirenment-build-10-320.jpg)
Windows 7 64 java envirenment build
- 1. Run Berkeley parser on Windows 7 (64bit) & java environment building recorded by Aaron [http://www.linkedin.com/in/aaronhan] 1. Download JDK, JDK is a development environment for building applications, applets, and components using the Java programming language, from “http://www.oracle.com/technetwork/java/javase/downloads/jdk7downloads-1880260.html” 2. Install “jdk-7u45-windows-x64.exe” 3. Configure the environment variables for windows 7 (64-bit). [reference: http://blog.163.com/wutianshui@126/blog/static/1869346220099455 115417/] Go to your computer -> property - > higher system setting -> environment variable
- 4. New -> Type “classpath” in the variable name, “C:Program FilesJavajdk1.7.0_45libdt.jar;C:Program FilesJava jdk1.7.0_45libtools.jar;.;” in the value of variable. Press the confirm button. New -> Type “java_home” in the variable name, “;C:Program FilesJavajdk1.7.0_45” in the end of previous value of variable. Press the confirm button.
- 5. Select the “path” variable from the existing list, press edit button, type “;%JAVA_HOME%bin” in the end of the previous value. 4. Build “Hello world” file to test your environment. Go to the desktop, build the file “HelloWorld.java”, put the following content into the file: --------------------------the following hello world java code---------------public class HelloWorld {
- 6. public static void main(String[] args) { System.out.println("Hello, World"); } } ----------------------------------------------Open the cmd window, type the following code: cd desktop java HelloWorld If it shows the following words, succeed! Begin the new life of java programming? 5. Download Berkeley parser (https://code.google.com/p/berkeleyparser/downloads/list). Download parser file “BerkeleyParser-1.7.jar” Download English grammar file “eng_sm6.gr” 6. Prepare segmented English corpus. In the same directory of the downloaded tools, build an English example file named as “sampleenglish.seg”, and put this sentence “this is a small house .” into this file.
- 7. 7. Run the Berkeley English parser for the first time. If you put the downloaded file in the directory “E:Berkeley Parser” Goto the “cmd” window, type the following commands. E: cd “Berkeley Parser” java -jar BerkeleyParser-1.7.jar -gr eng_sm6.gr -inputFile sampleenglish.seg -outputFile englishout.txt –render The above command “-render” is to generate the parsed picture. Wait for one moment, then it will generate the files “englishout.txt” and “thisisasmallhouse.png”. In the “englishout.txt” file there is the parsed sentence “( (S (NP (DT this)) (VP (VBZ is) (NP (DT a) (JJ small) (NN house))) (. .)) )”, and in the .png file, there is the parsed picture. =========================================== Train English grammar: Run Berkeley parser using “WSJ” English corpora. The “.mrg” documents contain the fully tagged English sentences while the “.prd” documents do not contain the POS tags of the words. Here we first use the fully tagged “.mrg” documents. Using the following command to training the English grammar: After around 8 hours, it finished the training as below:
- 8. ### above name is not suitable. It should be “TrainedGramEng.WSJ” instead of TrainedGramChi.CTB7.
- 9. Above running information shows the default training documents used by the Berkeley.parser. for WSJ corpora are from the document ID number 200 to the ID 2199. The following command can be run, which means that the Berkeley.parser for English WSJ corpora is defaulted for the “.mrg” documents instead of the “.prd” documents. the begging of the running information is as below:
- 10. Training began from 2013-11-10-16:30; finished at Store the running details of training grammar by the command “command…. > store.file.name” after training finished. ===================================== Testing: [[To the test the performance of a grammar you can use: java -cp berkeleyParser.jar edu.berkeley.nlp.PCFGLA.GrammarTester path <WSJ location> -in <grammar-file>]] the testing command of the trained grammar:
- 11. The testing score shows that: To record the running details of the testing stage, use the command: ============================================= Another training and testing of English grammar: The command of the training is below: After around xx hours, the training of the grammar finished as below:
- 12. Export the training record using below command, this will run the training once more, so you can use this command when you begin the training: The testing command: The result shows the scores as below, which are the same values with last testing run: