This PR significantly improves the efficiency of PDF processing with JPEG images, achieving 10x performance by avoiding unnecessary JPEG decode when insert JPEG into PDF.
The image buffer is now directly used as binary data when inserting JPEG images into PDF, skipping the decode step altogether.
Reason for Change:
image-rs supports retrieving the dimensions and color type without decode. This eliminates the need to decode JPEGs just to obtain these properties.
In addition, JPEG can be embedded directly into PDF as binary data. This is possible by the DCTDecode filter. Therefore, decoding JPEGs is redundant.
Implementation Details:
Modified the image_from function to accept a binary buffer and a file path. The image_from now checks the format of the buffer, if it is JPEG, directly uses the buffer for PDF embedding without decode.
For non-JPEG formats, the image is still decoded to determine properties like dimensions and color type, but JPEG images bypass the decode step entirely.
These optimizations lead to a significant performance gain, especially in scenarios involving large numbers of JPEG images.
Description:
This PR significantly improves the efficiency of PDF processing with JPEG images, achieving 10x performance by avoiding unnecessary JPEG decode when insert JPEG into PDF.
The image buffer is now directly used as binary data when inserting JPEG images into PDF, skipping the decode step altogether.
Reason for Change:
image-rs supports retrieving the
dimensionsandcolor typewithout decode. This eliminates the need to decode JPEGs just to obtain these properties.In addition, JPEG can be embedded directly into PDF as binary data. This is possible by the
DCTDecodefilter. Therefore, decoding JPEGs is redundant.Implementation Details:
Modified the
image_fromfunction to accept a binarybufferand a filepath. Theimage_fromnow checks the format of the buffer, if it is JPEG, directly uses the buffer for PDF embedding without decode.For non-JPEG formats, the image is still decoded to determine properties like
dimensionsandcolor type, but JPEG images bypass the decode step entirely.These optimizations lead to a significant performance gain, especially in scenarios involving large numbers of JPEG images.