shooting-event/backend/gemini.go

package main

import (
	"bytes"
	"encoding/json"
	"fmt"
	"io"
	"net/http"
	"net/url"
	"strings"
	"time"
)

type scoreAdviceModelResponse struct {
	AdvisedScore int    `json:"advisedScore"`
	Reason       string `json:"reason"`
}

type geminiGenerateRequest struct {
	Contents         []geminiContent          `json:"contents"`
	GenerationConfig geminiGenerationConfig   `json:"generationConfig,omitempty"`
	SafetySettings   []map[string]interface{} `json:"safetySettings,omitempty"`
}

type geminiContent struct {
	Role  string       `json:"role,omitempty"`
	Parts []geminiPart `json:"parts"`
}

type geminiPart struct {
	Text       string            `json:"text,omitempty"`
	InlineData *geminiInlineData `json:"inline_data,omitempty"`
}

type geminiInlineData struct {
	MimeType string `json:"mime_type"`
	Data     string `json:"data"`
}

type geminiGenerationConfig struct {
	Temperature      float64 `json:"temperature,omitempty"`
	ResponseMimeType string  `json:"responseMimeType,omitempty"`
	MaxOutputTokens  int     `json:"maxOutputTokens,omitempty"`
}

type geminiGenerateResponse struct {
	Candidates []struct {
		Content struct {
			Parts []struct {
				Text string `json:"text"`
			} `json:"parts"`
		} `json:"content"`
	} `json:"candidates"`
	Error *struct {
		Message string `json:"message"`
	} `json:"error,omitempty"`
}

func (a *App) generateScoreAdvice(stage string, playerID int, imageData string, currentScore int) (ScoreAdviceResponse, error) {
	mimeType, rawBase64, err := parseDataURI(imageData)
	if err != nil {
		return ScoreAdviceResponse{}, fmt.Errorf("invalid proof image data: %w", err)
	}

	model := strings.TrimSpace(a.cfg.GeminiModel)
	if model == "" {
		model = "gemini-2.0-flash"
	}

	requestPayload := geminiGenerateRequest{
		Contents: []geminiContent{
			{
				Role: "user",
				Parts: []geminiPart{
					{Text: scoreAdvicePrompt(stage, currentScore)},
					{
						InlineData: &geminiInlineData{
							MimeType: mimeType,
							Data:     rawBase64,
						},
					},
				},
			},
		},
		GenerationConfig: geminiGenerationConfig{
			Temperature:      0,
			ResponseMimeType: "application/json",
			MaxOutputTokens:  100,
		},
	}

	body, err := json.Marshal(requestPayload)
	if err != nil {
		return ScoreAdviceResponse{}, fmt.Errorf("marshal gemini request: %w", err)
	}

	endpoint := fmt.Sprintf(
		"https://generativelanguage.googleapis.com/v1beta/models/%s:generateContent?key=%s",
		url.PathEscape(model),
		url.QueryEscape(a.cfg.GeminiAPIKey),
	)

	req, err := http.NewRequest(http.MethodPost, endpoint, bytes.NewReader(body))
	if err != nil {
		return ScoreAdviceResponse{}, fmt.Errorf("create gemini request: %w", err)
	}
	req.Header.Set("Content-Type", "application/json")

	client := &http.Client{Timeout: 25 * time.Second}
	resp, err := client.Do(req)
	if err != nil {
		return ScoreAdviceResponse{}, fmt.Errorf("call gemini api: %w", err)
	}
	defer resp.Body.Close()

	respBody, err := io.ReadAll(io.LimitReader(resp.Body, 2_000_000))
	if err != nil {
		return ScoreAdviceResponse{}, fmt.Errorf("read gemini response: %w", err)
	}

	if resp.StatusCode >= 300 {
		return ScoreAdviceResponse{}, fmt.Errorf("gemini api status %d: %s", resp.StatusCode, strings.TrimSpace(string(respBody)))
	}

	var generated geminiGenerateResponse
	if err := json.Unmarshal(respBody, &generated); err != nil {
		return ScoreAdviceResponse{}, fmt.Errorf("decode gemini response: %w", err)
	}
	if generated.Error != nil && strings.TrimSpace(generated.Error.Message) != "" {
		return ScoreAdviceResponse{}, fmt.Errorf("gemini api error: %s", generated.Error.Message)
	}
	if len(generated.Candidates) == 0 || len(generated.Candidates[0].Content.Parts) == 0 {
		return ScoreAdviceResponse{}, fmt.Errorf("gemini returned no advice")
	}

	rawText := strings.TrimSpace(generated.Candidates[0].Content.Parts[0].Text)
	jsonText := extractJSONObject(rawText)
	if jsonText == "" {
		return ScoreAdviceResponse{}, fmt.Errorf("gemini response was not valid json")
	}

	var modelResponse scoreAdviceModelResponse
	if err := json.Unmarshal([]byte(jsonText), &modelResponse); err != nil {
		return ScoreAdviceResponse{}, fmt.Errorf("parse gemini advice json: %w", err)
	}

	return a.buildAdviceResponse(stage, playerID, modelResponse), nil
}

func parseDataURI(dataURI string) (string, string, error) {
	value := strings.TrimSpace(dataURI)
	if !strings.HasPrefix(value, "data:") {
		return "", "", fmt.Errorf("expected data uri")
	}

	parts := strings.SplitN(value, ",", 2)
	if len(parts) != 2 {
		return "", "", fmt.Errorf("invalid data uri format")
	}

	header := parts[0]
	payload := strings.TrimSpace(parts[1])
	if payload == "" {
		return "", "", fmt.Errorf("empty data uri payload")
	}

	mimeType := "image/jpeg"
	if semicolon := strings.Index(header, ";"); semicolon > len("data:") {
		mimeType = strings.TrimSpace(header[len("data:"):semicolon])
	}
	if mimeType == "" {
		mimeType = "image/jpeg"
	}

	return mimeType, payload, nil
}

func extractJSONObject(raw string) string {
	text := strings.TrimSpace(raw)
	if text == "" {
		return ""
	}
	if strings.HasPrefix(text, "```") {
		text = strings.TrimPrefix(text, "```json")
		text = strings.TrimPrefix(text, "```")
		text = strings.TrimSuffix(strings.TrimSpace(text), "```")
		text = strings.TrimSpace(text)
	}
	start := strings.Index(text, "{")
	end := strings.LastIndex(text, "}")
	if start == -1 || end == -1 || end <= start {
		return ""
	}
	return strings.TrimSpace(text[start : end+1])
}

func scoreAdvicePrompt(stage string, currentScore int) string {
	return fmt.Sprintf(`Target scoring assistant.
Stage: %s. Current score: %d.
Return STRICT JSON only:
{"advisedScore":<int 0..9999>,"reason":"<max 18 words>"}
Do not add markdown or extra fields.`, stage, currentScore)
}