browscap-java

A blazingly fast and memory efficient (thread-safe) Java client on top of the BrowsCap CSV source files. The BrowsCap version currently shipped is: 6000040.

Description

This library can be used to parse useragent headers in order to extract information about the used browser, browser version, platform, platform version and device type. Very useful to determine if the client is a desktop, tablet or mobile device or to determine if the client is on Windows or Mac OS (just to name a few examples).

Algorithm

We got some questions on the how and why of our algorithm and why it is "blazingly fast and efficient". In short, this is how our algorithm works:

1. All CSV lines are read and parsed in our own data structures (e.g. "Rule" objects). -- This doesn't involve regular expressions (which are memory consuming and costly) but uses a smart way of doing substrings on the CSV expression. -- The results of the substring operations (startsWith, endsWith, findIndices in SearchableString) are cached, so subsequent calls are very fast.
2. When all rules are generated, they're sorted by size and alphabet, so the first match can be returned immediately.
3. When looking up a useragent, all rules are filtered based on the "parts" of an expression. Most rules can be easily discarded because they don't contain a specific substring.
4. The filtering mechanism is based on bitset operations, which are very fast for large data sets.

Notes

• Although this library is very fast, implementing a cache is advisable. Since cache strategies differ per usecase, this library doesn't ship with one out of the box.
• All BrowsCap fields are available by configuration, but the following fields are loaded by default:
  • browser (e.g. Chrome)
  • browserType (e.g. Browser or Application)
  • browserMajorVersion (e.g. 80 in case of Chrome)
  • deviceType (e.g. Mobile Phone, Desktop, Tablet, Console, TV Device)
  • platform (e.g. Android, iOS, Win7, Win8, Win10)
  • platformVersion (e.g. 4.2, 10 depending on what the platform is)
• The fields are configurable by specifying a list of BrowsCapFields in the constructor of the UserAgentParser.
• The CSV file is read in a streaming way, so it's processed line by line. This makes it more memory efficient than loading the whole into memory first.
• 1000+ user agents are tested in the unit tests.

Future

Possible new features we're thinking of (and are not yet present):

• Auto-update the BrowsCap CSV

Maven

Add this to the dependencies in your pom.xml.

<dependency>
  <groupId>com.blueconic</groupId>
  <artifactId>browscap-java</artifactId>
  <version>1.3.1</version>
</dependency>

Usage

import com.blueconic.browscap.BrowsCapField;
import com.blueconic.browscap.Capabilities;
import com.blueconic.browscap.ParseException;
import com.blueconic.browscap.UserAgentParser;
import com.blueconic.browscap.UserAgentService;

// ... class definition

// create a parser with the default fields
final UserAgentParser parser = new UserAgentService().loadParser(); // handle IOException and ParseException

// or create a parser with a custom defined field list
// the list of available fields can be seen inthe BrowsCapField enum
final UserAgentParser parser =
        new UserAgentService().loadParser(Arrays.asList(BrowsCapField.BROWSER, BrowsCapField.BROWSER_TYPE,
                BrowsCapField.BROWSER_MAJOR_VERSION,
                BrowsCapField.DEVICE_TYPE, BrowsCapField.PLATFORM, BrowsCapField.PLATFORM_VERSION,
                BrowsCapField.RENDERING_ENGINE_VERSION, BrowsCapField.RENDERING_ENGINE_NAME,
                BrowsCapField.PLATFORM_MAKER, BrowsCapField.RENDERING_ENGINE_MAKER));

// It's also possible to supply your own ZIP file by supplying a correct path to a ZIP file in the constructor.
// This can be used when a new BrowsCap version is released which is not yet bundled in this package.
// final UserAgentParser parser = new UserAgentService("E:\\anil\\browscap.zip").loadParser();

// parser can be re-used for multiple lookup calls
final String userAgent = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.81 Safari/537.36";
final Capabilities capabilities = parser.parse(userAgent);

// the default fields have getters
final String browser = capabilities.getBrowser();
final String browserType = capabilities.getBrowserType();
final String browserMajorVersion = capabilities.getBrowserMajorVersion();
final String deviceType = capabilities.getDeviceType();
final String platform = capabilities.getPlatform();
final String platformVersion = capabilities.getPlatformVersion();

// the custom defined fields are available
final String renderingEngineMaker = capabilities.getValue(BrowsCapField.RENDERING_ENGINE_MAKER);

// do something with the values