imagemeta

Install the imagemeta library using go get to start extracting EXIF and XMP metadata from images and camera RAW formats.

go get github.com/evanoberholster/imagemeta

To extract basic metadata like image type, camera make, model, and original date/time, use the imagemeta.Decode function. This function accepts an io.ReadSeeker (such as an *os.File).

package main

import (
	"fmt"
	"os"

	"github.com/evanoberholster/imagemeta"
)

func main() {
	f, err := os.Open("image.jpg")
	if err != nil {
		panic(err)
	}
	defer f.Close()

	ex, err := imagemeta.Decode(f)
	if err != nil {
		panic(err)
	}

	fmt.Println("type:", ex.ImageType)
	fmt.Println("make:", ex.Make)
	fmt.Println("model:", ex.Model)
	fmt.Println("datetime:", ex.DateTimeOriginal)
}

The github.com/evanoberholster/imagemeta/meta/xmp package allows you to parse XMP data from two sources:

1. Sidecar files: Use xmp.ParseXmp(file) to parse a standalone .xmp file.
2. Embedded blocks: Use xmp.Parse(file) to extract XMP data embedded directly within an image file.

// Parse a sidecar .xmp file
x, err := xmp.ParseXmp(file)

// Parse embedded XMP from an image
x, err := xmp.Parse(file)

If you need to identify a file format before performing a full decode, use the imagetype package. The Scan function provides fast format detection.

import "github.com/evanoberholster/imagemeta/imagetype"

t, err := imagetype.Scan(r)

The imagemeta package provides several specialized decoding entry points for specific file formats. All methods accept an io.ReadSeeker:

• Decode(io.ReadSeeker): General purpose decoder
• DecodeJPEG(io.ReadSeeker)
• DecodeTiff(io.ReadSeeker)
• DecodeCR3(io.ReadSeeker)
• DecodeCRW(io.ReadSeeker)
• DecodeHeif(io.ReadSeeker)
• DecodePng(io.ReadSeeker)
• PreviewCR3(io.ReadSeeker): Extracts embedded CR3 previews

Use the following methods to get human-readable or structural information about an Entry:

• Name() string: Returns the string name of the tag based on its ID and IfdType (e.g., "ImageWidth").
• Size() uint32: Returns the total size in bytes of the tag's value (Type.Size() * UnitCount).

fmt.Printf("Tag: %s, Size: %d bytes\n", entry.Name(), entry.Size())

You can pass PreviewOption functions to TIFFPreviews, ExtractPreview, or PreviewTIFF to control extraction behavior.

Currently, the primary option is:

• WithMaxPreviewLength(n int64): Sets a maximum byte length for previews. If a preview's declared length is greater than n, it will be ignored by TIFFPreviews or cause ExtractPreview to return ErrNoPreview. This is useful for preventing large, potentially malicious allocations based on attacker-controlled metadata.

// Limit previews to 5MB to prevent large memory allocations
opts := []imagemeta.PreviewOption{imagemeta.WithMaxPreviewLength(5 * 1024 * 1024)}

previews, err := imagemeta.TIFFPreviews(file, opts...)

Use PreviewTIFF(r io.ReadSeeker, opts ...PreviewOption) ([]byte, error) as a convenience method to quickly get the largest renderable JPEG preview from a TIFF-based file.

This function internally calls TIFFPreviews and attempts to ExtractPreview each one in order until it succeeds. If no valid preview can be extracted, it returns ErrNoPreview.

// Quickly get the best available preview
previewBytes, err := imagemeta.PreviewTIFF(file)
if err != nil {
    if errors.Is(err, imagemeta.ErrNoPreview) {
        fmt.Println("No valid preview found")
    } else {
        log.Fatal(err)
    }
}

While Decode() handles automatic detection, you can use format-specific functions if you already know the image type. These functions return an exif.Exif object or an error.

Supported specific decoders:

• DecodeCR3(r io.ReadSeeker): Decodes Canon CR3 files.
• DecodeTiff(r io.ReadSeeker): Decodes TIFF or DNG files.
• DecodeCR2(r io.ReadSeeker): Decodes Canon CR2 files (uses DecodeTiff).
• DecodeARW(r io.ReadSeeker): Decodes Sony ARW files (uses DecodeTiff).
• DecodeCRW(r io.ReadSeeker): Decodes Canon CRW files.
• DecodeHeif(r io.ReadSeeker): Decodes HEIF files.
• DecodeJPEG(r io.ReadSeeker): Decodes JPEG files.
• DecodePng(r io.ReadSeeker): Decodes PNG files.

Use ExtractPreview(r io.ReadSeeker, p PreviewImage, opts ...PreviewOption) ([]byte, error) to read the raw bytes of a specific preview identified by a PreviewImage struct.

Important details:

• The function validates the JPEG SOI marker (0xff 0xd8) before returning the data.
• It performs bounds checking against the actual file size to prevent out-of-bounds reads.
• If the preview exceeds the configured maxLength in PreviewOption, it returns ErrNoPreview.
• Returns a newly allocated byte slice containing the JPEG data.

// Extract the first (largest) preview found
if len(previews) > 0 {
    data, err := imagemeta.ExtractPreview(file, previews[0])
    if err != nil {
        if errors.Is(err, imagemeta.ErrNoPreview) {
            // Handle no preview
        }
        log.Fatal(err)
    }
    // 'data' now contains the JPEG bytes
}

Use NewEntry for standard TIFF entries where the value is stored at a file offset. Use NewInlineEntry for BigTIFF entries where the value is packed inline within the 8-byte value field. NewInlineEntry automatically sets the embedded flag to true, signaling to parsers that ValueOffset should be treated as data rather than a file pointer.

// Standard entry
entry := tag.NewEntry(id, typ, unitCount, valueOffset, directoryType, ifdIndex, byteOrder)

// Inline entry (BigTIFF)
entry := tag.NewInlineEntry(id, typ, unitCount, lo, hi, directoryType, ifdIndex, byteOrder)