函数式 API

所有核心函数均可单独 import：

import {
  crop, resize, rotate, rotate90, rotate90Once, flip, filter,
  watermark, applyWatermarkImage, applyTextWatermark, resolvePosition,
  compress, convert, metadata, parseExif, normalizeMime,
  createZip, crc32,
  imgkitBatch,
} from 'imgpilot';

浏览器工具函数：

import {
  loadImage, getImageData, putImageData, createCanvas,
  toBlob, toDataURL, renderText,
  browserEncoder, browserTextRenderer,
} from 'imgpilot';

crop

function crop(image: ImageDataLike, options: CropOptions): ImageDataLike

CropOptions：

• x, y, width, height：区域裁剪
• aspectRatio：宽高比裁剪（与上互斥）
• align：宽高比模式下的对齐方式（Position）

resize

function resize(image: ImageDataLike, options: ResizeOptions): ImageDataLike

ResizeOptions：

• width, height：目标尺寸（至少一个）
• fit：'exact' | 'contain' | 'cover' | 'fill'
• algorithm：'nearest' | 'bilinear'

rotate / flip

function rotate(image: ImageDataLike, degrees: number): ImageDataLike
function rotate90(image: ImageDataLike, deg: number): ImageDataLike
function rotate90Once(image: ImageDataLike): ImageDataLike
function flip(image: ImageDataLike, axis: 'horizontal' | 'vertical'): ImageDataLike

rotate 支持任意角度，画布扩展以容纳完整图像，空白区域为透明。rotate90 是 90° 整数倍的高效实现，rotate90Once 是单次 90° 顺时针旋转。

filter

function filter(image: ImageDataLike, options: FilterOptions): ImageDataLike

FilterOptions 所有字段可选：

• grayscale 0-1
• sepia 0-1
• brightness -1~1
• contrast -1~1
• blur >=0（高斯半径，可分离卷积）
• invert 0-1
• opacity 0-1
• hueRotate 角度
• saturate 0-1（1 为不变）

watermark

function watermark(image, options, textRenderer?): ImageDataLike
function applyWatermarkImage(image, wmImage, options): ImageDataLike
function applyTextWatermark(image, text, options, textRenderer?): ImageDataLike
function resolvePosition(...)

compress / convert

async function compress(image, options): Promise<CompressResult>
async function convert(image, mimeType, quality?): Promise<Blob>

compress 在指定 maxSize 时启用二分搜索：在 [0.1, quality] 范围内迭代 12 次，找最大满足体积约束的质量值。

metadata

function metadata(image: ImageDataLike): ImageMetadata

返回 { width, height, size, pixels, channels, averageBrightness, hasAlpha }。

parseExif

function parseExif(buf: ArrayBuffer): ExifInfo

从 JPEG 原始字节中解析 EXIF 元信息。返回 { orientation, make, model, dateTime, gps }，非 JPEG 或无 EXIF 字段时返回空对象。

imgkitBatch

function imgkitBatch(sources: Array<File | Blob | string | ArrayBuffer>): Promise<Pipeline[]>

批量加载多张图片，返回各自独立的 Pipeline 实例数组。

const pipelines = await imgkitBatch([fileA, fileB, fileC]);
pipelines[0].resize({ width: 800 });
pipelines[1].filter({ grayscale: 1 });

createZip / crc32

function createZip(entries: ZipEntry[]): Blob
function crc32(data: Uint8Array): number

createZip 将多个文件打包为 ZIP Blob，支持浏览器端下载。crc32 为 CRC-32 校验值计算。

const zipBlob = createZip([
  { name: 'result-1.jpg', data: uint8Array1 },
  { name: 'result-2.png', data: uint8Array2 },
]);

浏览器工具函数

function loadImage(source: File | Blob | string | ArrayBuffer): Promise<HTMLImageElement>
function getImageData(source: HTMLImageElement | HTMLCanvasElement | OffscreenCanvas): ImageDataLike
function putImageData(canvas: HTMLCanvasElement | OffscreenCanvas, data: ImageDataLike): void
function createCanvas(w: number, h: number): HTMLCanvasElement | OffscreenCanvas
function toBlob(canvas: HTMLCanvasElement | OffscreenCanvas, mime: ImageMimeType, quality?: number): Promise<Blob>
function toDataURL(canvas: HTMLCanvasElement | OffscreenCanvas, mime: ImageMimeType, quality?: number): string
function renderText(text: string, options: { font: string; color: string; rotate?: number }): ImageDataLike

工具函数：

function createImageData(w: number, h: number): ImageDataLike
function cloneImageData(src: ImageDataLike): ImageDataLike
function getPixel(data: Uint8ClampedArray, offset: number): [number, number, number, number]
function setPixel(data: Uint8ClampedArray, offset: number, r: number, g: number, b: number, a: number): void
function packRGBA(r: number, g: number, b: number, a: number): number
function clamp(v: number, lo: number, hi: number): number
function lerp(a: number, b: number, t: number): number
function degToRad(deg: number): number
function gaussian(x: number, sigma: number): number
function gaussianKernel(sigma: number, radius: number): number[]
function binarySearchQuality(...): number