filter

Documentation

Overview

Package filter implements the PDF stream decode filters of ISO 32000-1 §7.4 that text extraction needs: FlateDecode, LZWDecode, ASCIIHexDecode, ASCII85Decode and RunLengthDecode, with the PNG and TIFF predictors. The image codecs (DCTDecode, JPXDecode, CCITTFaxDecode, JBIG2Decode) are not stream filters here: Apply rejects them, and ImageCodec identifies them so a caller can stop a filter chain at the codec and hand the still-encoded image data to a decoder.

The package knows nothing about the PDF object model: callers extract /DecodeParms fields into Params.

Index

• func Apply(r io.Reader, filterName string, p Params) (io.Reader, error)
• func ImageCodec(name string) (string, bool)
• type Params
• type Releaser

Functions

func Apply

func Apply(r io.Reader, filterName string, p Params) (io.Reader, error)

Apply wraps r with the named decode filter. The returned reader fails on malformed data rather than panicking; callers bound total output.

func ImageCodec

func ImageCodec(name string) (string, bool)

ImageCodec reports whether name (full or abbreviated) is one of the image codecs, returning its full name: DCTDecode, JPXDecode, CCITTFaxDecode or JBIG2Decode.

Types

type Params

type Params struct {
	Predictor        int
	Colors           int
	BitsPerComponent int
	Columns          int
	NoEarlyChange    bool // LZW /EarlyChange 0 (the non-default variant)
}

Params carries the /DecodeParms fields a filter consumes. The zero value means the spec defaults (Predictor 1, Colors 1, BitsPerComponent 8, Columns 1, EarlyChange 1).

type Releaser

type Releaser interface {
	Release()
}

Releaser is implemented by readers that hold pooled resources. A caller that has finished with a stream calls Release to return them; a reader that is never released is simply garbage collected. Reads after Release fail.