Recently I had a job to create some statistics on interactions between our visitors and our landing page. When a visitor
downloads our product, requires a license or buys the product, a new line is inserted in the log file of the web server.
All I have is a big log file with a lot of information.
For example DownloadServlet for each download request writes in the container's log a line in the format
[webapp 2008/10/06 16:12:16] - 192.168.12.124, /download/next-reports-setup-1.7-jre.exe
What I want is to see some statistics about downloads.
My idea is to convert the log file in a collection of business objects (Download in my case) and to make some query
using
josql
or
cqengine a
nd extract the required information. I can also insert all downloads in a
database table and use sql to extract the required information.
With
Log2j I can convert a log's lines in java objects with a single line:
new Log2j()
.map(Download.class)
.addEntityHandler(new DownloadHandler())
.parse(new InputStreamReader(input));
What is Log2j?
Log2j is an open source (Apache license) tiny (around 15KB) log file to POJOs converter, with zero dependencies and a quick learning curve.
No XML, only Java.
My first step is to create a simple POJO class with the name Download having some properties.
@RegexEntity(pattern = "PATTERN")
public class Download {
// [webapp 2008/10/06 16:12:16] - 192.168.12.124, /download/next-report-setup-1.7-jre.exe, f13dfc7fe609480297a0b15d611676b4
public static final String PATTERN = "\\[webapp\\s"
+ "(20[0-1][0-9]/\\d{2}/\\d{2}\\s\\d{2}:\\d{2}:\\d{2})" // date
+ "\\]\\s-\\s<\\$>\\s"
+ "(\\d{1,3}\\.\\d{1,3}\\.\\d{1,3}\\.\\d{1,3})" // ip
+ ",\\s/download/"
+ "([^,]*)" // file
+ ".*";
@RegexField(group = 1, converter = MyDateConverter.class)
private Date date;
@RegexField(group = 3)
private String ip;
@RegexField(group = 4)
private String file;
// getters and setters
}
I added annotation @RegexEntity(pattern = "PATTERN") to class Download. The @RegexEntity annotation may have a
parameter named pattern which represents the constant name that returns the regex string.
The default value is PATTERN (in Download class I added pattern value for transparency).
This annotation is informing Log2j that all log's lines that respect the pattern will be transformed in
Download objects.
The second step is to specify how Log2j will transform text fragments from log's lines in Download's properties.
For this purpose I used
@RegexField annotation. This annotation takes a mandatory parameter named group that represents
the group index from the java's regex Matcher object and an optional parameter named converter that will be used by
Log2j to convert the text fragment in property's value.
The third step is to create a
DownloadHandler for handling Download objects.
public class DownloadHandler implements EntityHandler<Download> {
private int count;
@Override
public void beforeFirstEntity() {
count = 0;
}
@Override
public void handleEntity(Download entity) {
count++;
// only display the entity
System.out.println(entity);
}
@Override
public void afterLastEntity() {
System.out.println("Handled " + count + " Download entities");
}
}
The methods beforeFirstEntity() and afterLastEntity() are callback methods.
Log2j will call these methods one time at the start/end of log parsing.
For example you can start a database transaction, clear an entity table in beforeFirstEntity()
and commit the database transaction in afterLastEntity().
In the example above my DownloadHandler init count variable to zero in beforeFirstEntity(), prints all
download objects to System.out in handleEntity() and prints a count with download entities in afterLastEntity().
Another important concept in Log2j is Converter.
A
Converter is used by Log2j to transform a text fragment into a POJO property's value.
Log2j comes with builtin converters for all primitive values (Boolean, Byte, Short, ...).
Log2j also allows you to register new general converters (for each field's type) or you can register a converter only for a particular field.
public class MyDateConverter implements Converter<Date> {
@Override
public Date convert(String text) {
try {
return new SimpleDateFormat("yyyy/MM/dd").parse(text);
} catch (ParseException e) {
e.printStackTrace();
}
return null;
}
}
For example MyDateConverter is used in Download class by Log2j to transform the text fragment in Date object. If you want to use MyDateConverter for all POJO fields with type Date you can do it with:
new Log2j()
...
registerConverter(new MyDateConverter();
In my next article about Log2j I will talk about validations and how you can interrogate the collection of POJOs.
{{ parent.title || parent.header.title}}
{{ parent.tldr }}
{{ parent.linkDescription }}
{{ parent.urlSource.name }}