PhotoshopAPI

The PhotoshopTest/documents directory contains a collection of test files used to verify that PhotoshopAPI accurately and efficiently parses various Photoshop file formats and combinations. These documents are used by the testing suite to validate full parsing and the correctness of specific internal sections or algorithms.

Key Details:

• Generation Version: All test documents were generated using Photoshop Version 22.3.2 (Photoshop 2022).
• Resolution: To maintain repository size, individual test documents are typically low-resolution (rarely exceeding 64x64).

The PhotoshopAPI provides specialized 2D Geometry structures designed to facilitate image transformations, such as warping.

Key characteristics:

• Optimization: These structures are optimized for execution speed rather than maximum flexibility.
• Current Scope: Currently, these structures are primarily used for manipulating SmartObject layers.
• Future Support: Planned updates may extend these capabilities to allow more generic image transformations across different layer types.

To work with geometry, you will need to explore operations, meshes, Bezier curves, and SmartObject warping specifically.

The Python bindings provide a 'pythonic' interface to the underlying C++ API. While they share the same core logic, the function signatures and structures may differ from the C++ implementation to better suit Python usage patterns. The main classes you will interact with include:

• LayeredFile: For reading and writing Photoshop files.
• ImageLayer_16bit: For handling high-bit depth layer data.
• ImageLayer_8bit: For handling standard bit depth layer data.
• Enums: For constants like blend modes.
• Utility: For helper functions.

The SmartObject namespace provides several specialized structures for defining warp transformations on Smart Objects. Depending on the desired effect, you can use one of the following three warp types:

1. SmartObject::Warp: The standard warp structure for general mesh-based distortions.
2. SmartObject::PuppetWarp: A specialized structure designed for puppet warp transformations, typically used for more complex, localized deformations.
3. SmartObject::PerspectiveWarp: A structure specifically for applying perspective transformations to a Smart Object.

For 16-bit and 32-bit files, Photoshop uses prediction encoded zip compression for the main data, regardless of whether the compressed size is actually smaller than the uncompressed data.

However, the merged image data section (used for compatibility) is always saved uncompressed. This can result in significantly larger file sizes because the merged section misses out on the performance benefits of Zip Prediction.

PhotoshopAPI uses the c-blosc2 library to perform on-the-fly compression and decompression of image data in system memory. This approach solves two problems:

1. Memory Overhead: Storing uncompressed image data (especially for large PSB files with many layers) can lead to massive memory consumption. In-memory compression significantly reduces this footprint.
2. Performance: Because the compression/decompression is highly optimized, it can actually be faster than a standard memcpy() when passing data, leading to faster overall parsing speeds.

The core abstraction responsible for this behavior is the ImageChannel class, which is used internally by both LayeredFile and PhotoshopFile.

The ImageData structure is the primary container for image data within the PhotoshopAPI. It encapsulates the raw pixel information and metadata required to represent an image in memory.

Note: Detailed member definitions (such as width, height, channels, or pixel buffers) are provided by the Doxygen documentation generated from the source code.