Cyrene_v2/internal/runtime/multimodal/vision/image.go

package vision

import (
	"context"
	"encoding/base64"
	"fmt"
	"io"
	"net/http"
	"os"
	"strings"
)

// 读取图像字节与响应体的体积上限，使恶意或
// 畸形来源无法耗尽内存。50MB 上限与
// DevDocs/02-development-constraints.md §11 的文件上传限制一致。
const (
	maxImageBytes  = 50 << 20
	maxFetchedBody = 50 << 20
)

// DataURI 由 mime 与原始字节构造一个 base64 data: 图像 URI。
func DataURI(mime string, data []byte) string {
	return "data:" + mime + ";base64," + base64.StdEncoding.EncodeToString(data)
}

// Resolve 加载一个 ImageRef 对应的字节。路径与 base64 载荷在
// 本地读取；URL 用 http.Client 抓取（传入的 client，为 nil 时
// 用默认 client）。对 base64 输入，Data 字段原样作为
// 载荷返回，MIME 类型从 data: 前缀推断。
func Resolve(ctx context.Context, ref ImageRef, client *http.Client) (mime string, data []byte, err error) {
	switch ref.Source {
	case SourcePath:
		return readFile(ref.Data)
	case SourceBase64:
		return decodeBase64(ref.Data)
	case SourceURL:
		return fetchURL(ctx, ref.Data, client)
	default:
		return "", nil, fmt.Errorf("vision: %w: %q", ErrUnsupportedSource, ref.Source)
	}
}

// readFile 读取一个图像文件，并从扩展名推断其 MIME 类型。
func readFile(path string) (string, []byte, error) {
	f, err := os.Open(path)
	if err != nil {
		return "", nil, fmt.Errorf("vision: %w: %v", ErrBadImage, err)
	}
	defer func() { _ = f.Close() }()
	data, err := io.ReadAll(io.LimitReader(f, maxImageBytes))
	if err != nil {
		return "", nil, fmt.Errorf("vision: %w: %v", ErrBadImage, err)
	}
	return mimeByExt(path), data, nil
}

// decodeBase64 返回 base64/data: 载荷解码后的字节及其 MIME。
func decodeBase64(src string) (string, []byte, error) {
	mime, encoded, isDataURI := splitDataURI(src)
	if isDataURI {
		decoded, err := base64.StdEncoding.DecodeString(encoded)
		if err != nil {
			return "", nil, fmt.Errorf("vision: %w: %v", ErrBadImage, err)
		}
		return mime, decoded, nil
	}
	decoded, err := base64.StdEncoding.DecodeString(src)
	if err != nil {
		return "", nil, fmt.Errorf("vision: %w: %v", ErrBadImage, err)
	}
	return "image/png", decoded, nil
}

// fetchURL 通过 http(s) 下载一张图像；只允许 http/https 协议。
func fetchURL(ctx context.Context, rawURL string, client *http.Client) (string, []byte, error) {
	if !strings.HasPrefix(rawURL, "http://") && !strings.HasPrefix(rawURL, "https://") {
		return "", nil, fmt.Errorf("vision: %w: %q", ErrUnsupportedSource, rawURL)
	}
	if client == nil {
		client = defaultHTTPClient()
	}
	req, err := http.NewRequestWithContext(ctx, http.MethodGet, rawURL, nil)
	if err != nil {
		return "", nil, fmt.Errorf("vision: %w: %v", ErrFetch, err)
	}
	resp, err := client.Do(req)
	if err != nil {
		return "", nil, fmt.Errorf("vision: %w: %v", ErrFetch, err)
	}
	defer func() { _ = resp.Body.Close() }()
	if resp.StatusCode != http.StatusOK {
		return "", nil, fmt.Errorf("vision: %w: status %d", ErrFetch, resp.StatusCode)
	}
	data, err := io.ReadAll(io.LimitReader(resp.Body, maxFetchedBody))
	if err != nil {
		return "", nil, fmt.Errorf("vision: %w: %v", ErrFetch, err)
	}
	return mimeByContentType(resp.Header.Get("Content-Type")), data, nil
}

// splitDataURI 解析 "data:<mime>;base64,<payload>" 前缀。src 不是
// data: URI 时第二个返回值是 false。
func splitDataURI(src string) (mime, payload string, ok bool) {
	if !strings.HasPrefix(src, "data:") {
		return "", "", false
	}
	rest := strings.TrimPrefix(src, "data:")
	comma := strings.IndexByte(rest, ',')
	if comma < 0 {
		return "", "", false
	}
	meta := rest[:comma]
	payload = rest[comma+1:]
	mime = "image/png"
	if strings.HasPrefix(meta, ";base64") {
		return mime, payload, true
	}
	if idx := strings.Index(meta, ";base64"); idx >= 0 {
		if t := strings.TrimSuffix(meta[:idx], ";base64"); t != "" {
			mime = t
		}
		return mime, payload, true
	}
	return mime, payload, true
}

// mimeByExt 从文件路径扩展名推断 MIME 类型。
func mimeByExt(path string) string {
	switch strings.ToLower(pathExtension(path)) {
	case ".png":
		return "image/png"
	case ".jpg", ".jpeg":
		return "image/jpeg"
	case ".gif":
		return "image/gif"
	case ".webp":
		return "image/webp"
	default:
		return "image/png"
	}
}

// mimeByContentType 从 HTTP Content-Type 头返回 MIME 类型，
// 缺失或非图像时默认为 image/png。
func mimeByContentType(ct string) string {
	if ct == "" || !strings.HasPrefix(ct, "image/") {
		return "image/png"
	}
	return strings.TrimSpace(strings.Split(ct, ";")[0])
}

// pathExtension 返回带点号的文件扩展名，没有则返回空串。
func pathExtension(path string) string {
	base := path
	if idx := strings.LastIndexByte(path, '/'); idx >= 0 {
		base = path[idx+1:]
	} else if idx := strings.LastIndexByte(path, '\\'); idx >= 0 {
		base = path[idx+1:]
	}
	dot := strings.LastIndexByte(base, '.')
	if dot < 0 {
		return ""
	}
	return base[dot:]
}

// fileExists 报告 path 是否存在于磁盘且为常规文件。
func fileExists(path string) bool {
	info, err := os.Stat(path)
	return err == nil && info.Mode().IsRegular()
}

// defaultHTTPClient 构造 DevDocs/02-development-constraints.md §2.3
// 所述的 HTTP client。这里每次调用返回一个新 client，
// 而不是包级全局变量——后者为约束所禁止。
func defaultHTTPClient() *http.Client {
	return &http.Client{Transport: http.DefaultTransport}
}