free-ide-proxy/cmd/vision/main.go

package main

import (
	"bytes"
	"encoding/base64"
	"encoding/json"
	"fmt"
	"io"
	"net/http"
	"os"
	"path/filepath"
	"strings"
	"time"
)

const defaultAPIKey = "opx-57ffcb013651f7acae558cd9ab094dbd719fdec980a6ad53"

type visionRequest struct {
	Model    string         `json:"model"`
	Messages []visionMessage `json:"messages"`
	Stream   bool           `json:"stream"`
}

type visionMessage struct {
	Role    string      `json:"role"`
	Content interface{} `json:"content"`
}

type visionResponse struct {
	ID      string         `json:"id"`
	Object  string         `json:"object"`
	Created int64          `json:"created"`
	Model   string         `json:"model"`
	Choices []visionChoice `json:"choices"`
}

type visionChoice struct {
	Index        int           `json:"index"`
	Message      visionContent `json:"message"`
	FinishReason string        `json:"finish_reason"`
}

type visionContent struct {
	Role    string      `json:"role"`
	Content interface{} `json:"content"`
}

func main() {
	if len(os.Args) < 2 {
		fmt.Fprintf(os.Stderr, "Usage: vision <image-path> [prompt]\n")
		fmt.Fprintf(os.Stderr, "Example: vision C:\\path\\to\\photo.jpg \"Describe this image\"\n")
		os.Exit(1)
	}

	imagePath := os.Args[1]
	prompt := "Describe esta imagen en detalle."
	if len(os.Args) > 2 {
		prompt = strings.Join(os.Args[2:], " ")
	}

	ext := strings.ToLower(filepath.Ext(imagePath))
	var mediaType string
	switch ext {
	case ".jpg", ".jpeg":
		mediaType = "image/jpeg"
	case ".png":
		mediaType = "image/png"
	case ".webp":
		mediaType = "image/webp"
	default:
		fmt.Fprintf(os.Stderr, "Unsupported image format: %s (supported: jpg, png, webp)\n", ext)
		os.Exit(1)
	}

	data, err := os.ReadFile(imagePath)
	if err != nil {
		fmt.Fprintf(os.Stderr, "Error reading file: %v\n", err)
		os.Exit(1)
	}

	b64 := base64.StdEncoding.EncodeToString(data)
	dataURL := fmt.Sprintf("data:%s;base64,%s", mediaType, b64)

	reqBody := visionRequest{
		Model: "mimo-v2.5-free",
		Messages: []visionMessage{
			{
				Role: "user",
				Content: []map[string]interface{}{
					{"type": "text", "text": prompt},
					{"type": "image_url", "image_url": map[string]string{"url": dataURL}},
				},
			},
		},
	}

	body, _ := json.Marshal(reqBody)

	httpReq, err := http.NewRequest("POST", "https://opencode.ai/zen/v1/chat/completions", bytes.NewReader(body))
	if err != nil {
		fmt.Fprintf(os.Stderr, "Error creating request: %v\n", err)
		os.Exit(1)
	}

	httpReq.Header.Set("Content-Type", "application/json")
	httpReq.Header.Set("Authorization", "Bearer "+defaultAPIKey)
	httpReq.Header.Set("User-Agent", "vision-tool/1.0")
	httpReq.Header.Set("x-opencode-client", "cli")
	httpReq.Header.Set("x-opencode-project", "vision")
	httpReq.Header.Set("x-opencode-request", fmt.Sprintf("req_%d", time.Now().UnixNano()))
	httpReq.Header.Set("x-opencode-session", fmt.Sprintf("ses_%d", time.Now().UnixNano()))

	client := &http.Client{Timeout: 2 * time.Minute}
	resp, err := client.Do(httpReq)
	if err != nil {
		fmt.Fprintf(os.Stderr, "Error: %v\n", err)
		os.Exit(1)
	}
	defer resp.Body.Close()

	respData, _ := io.ReadAll(resp.Body)

	if resp.StatusCode != 200 {
		fmt.Fprintf(os.Stderr, "API error %d: %s\n", resp.StatusCode, string(respData))
		os.Exit(1)
	}

	var result visionResponse
	if err := json.Unmarshal(respData, &result); err != nil {
		fmt.Fprintf(os.Stderr, "Error parsing response: %v\n%s\n", err, string(respData))
		os.Exit(1)
	}

	if len(result.Choices) > 0 {
		switch c := result.Choices[0].Message.Content.(type) {
		case string:
			fmt.Println(c)
		default:
			b, _ := json.MarshalIndent(c, "", "  ")
			fmt.Println(string(b))
		}
	}
}