org.apache.tika.detect
Class TextDetector

java.lang.Object
  org.apache.tika.detect.TextDetector

All Implemented Interfaces:

Serializable, Detector

public class TextDetector
extends Object
implements Detector

Content type detection of plain text documents. This detector looks at the beginning of the document input stream and considers the document to be a text document if no ASCII (ISO-Latin-1, UTF-8, etc.) control bytes are found. As a special case some control bytes (up to 2% of all characters) are also allowed in a text document if it also contains no or just a few (less than 10%) characters above the 7-bit ASCII range.

Note that text documents with a character encoding like UTF-16 are better detected with MagicDetector and an appropriate magic byte pattern.

Since:

Apache Tika 0.3

See Also:

Serialized Form

Constructor Summary

TextDetector()
Constructs a TextDetector which will look at the default number of bytes from the beginning of the document.

TextDetector(int bytesToTest)
Constructs a TextDetector which will look at a given number of bytes from the beginning of the document.

Method Summary

MediaType

detect(InputStream input, Metadata metadata) Looks at the beginning of the document input stream to determine whether the document is text or not.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

TextDetector

public TextDetector()

Constructs a TextDetector which will look at the default number of bytes from the beginning of the document.

TextDetector

public TextDetector(int bytesToTest)

Constructs a TextDetector which will look at a given number of bytes from the beginning of the document.

Method Detail

detect

public MediaType detect(InputStream input,
                        Metadata metadata)
                 throws IOException

Looks at the beginning of the document input stream to determine whether the document is text or not.

Specified by:

detect in interface Detector

Parameters:

input - document input stream, or null

metadata - ignored

Returns:

"text/plain" if the input stream suggest a text document, "application/octet-stream" otherwise

Throws:

IOException - if the document input stream could not be read