feat: Phase 4 — LLM 集成，PR 描述生成与 AI 代码审查

2026-06-19 19:43:18 +08:00
parent 7ac3a3726b
commit afb270721c
8 changed files with 783 additions and 29 deletions
@@ -0,0 +1,151 @@
+package services
+
+import (
+	"database/sql"
+	"encoding/json"
+	"fmt"
+	"strings"
+
+	"github.com/go-git/go-git/v5/plumbing"
+	"github.com/go-git/go-git/v5/plumbing/object"
+	goopenai "github.com/sashabaranov/go-openai"
+)
+
+// PRDescription is the structured output from PR description generation.
+type PRDescription struct {
+	Title   string `json:"title"`
+	Type    string `json:"type"`
+	Summary string `json:"summary"`
+	Details string `json:"details"`
+	Impact  string `json:"impact"`
+}
+
+// GeneratePR generates a structured PR description from commit history and diff.
+// It streams progress via the callback and returns the parsed PR description.
+func GeneratePR(db *sql.DB, repoPath, base, head string, callback StreamCallback) (*PRDescription, error) {
+	// Read LLM config
+	config, err := GetLLMConfig(db)
+	if err != nil {
+		return nil, err
+	}
+
+	// Open repo
+	repo, err := OpenRepo(repoPath)
+	if err != nil {
+		return nil, fmt.Errorf("open repo: %w", err)
+	}
+
+	// Get commits between base and head
+	commits, err := GetCommitLog(repo, head, 100)
+	if err != nil {
+		return nil, fmt.Errorf("get commits: %w", err)
+	}
+
+	// Filter commits to only those reachable from head but not from base
+	baseHash, err := repo.ResolveRevision(plumbing.Revision(base))
+	if err == nil {
+		baseCommit, bcErr := repo.CommitObject(*baseHash)
+		if bcErr == nil {
+			baseSet := make(map[string]bool)
+			baseQueue := []*object.Commit{baseCommit}
+			for len(baseQueue) > 0 {
+				c := baseQueue[0]
+				baseQueue = baseQueue[1:]
+				if baseSet[c.Hash.String()] {
+					continue
+				}
+				baseSet[c.Hash.String()] = true
+				for _, p := range c.ParentHashes {
+					pc, err := repo.CommitObject(p)
+					if err == nil {
+						baseQueue = append(baseQueue, pc)
+					}
+				}
+			}
+			var filtered []CommitInfo
+			for _, ci := range commits {
+				if !baseSet[ci.Hash] {
+					filtered = append(filtered, ci)
+				}
+			}
+			if len(filtered) > 0 {
+				commits = filtered
+			}
+		}
+	}
+
+	// Get diff
+	diff, err := GetDiff(repo, base, head)
+	if err != nil {
+		return nil, fmt.Errorf("get diff: %w", err)
+	}
+
+	// Build prompt
+	var commitLines []string
+	for _, c := range commits {
+		commitLines = append(commitLines, fmt.Sprintf("- %s %s", c.ShortHash, c.Message))
+	}
+
+	commitStr := strings.Join(commitLines, "\n")
+	if commitStr == "" {
+		commitStr = "(no commits)"
+	}
+
+	// Truncate diff if too large (approx 60k chars to stay within token limits)
+	if len(diff) > 60000 {
+		diff = diff[:60000] + "\n\n... [diff truncated due to size]"
+	}
+
+	prompt := fmt.Sprintf(`你是一个专业的技术文档撰写助手。根据以下 Git 变更信息，生成一份结构化的 PR 描述。
+
+## Commit 记录
+%s
+
+## 代码变更 (Diff)
+%s
+
+请按以下 JSON 格式输出（直接输出 JSON，不要包含 markdown 代码块标记）：
+{
+  "title": "简洁的 PR 标题",
+  "type": "变更类型: feat|fix|refactor|docs|chore|style|test|perf",
+  "summary": "一段话概述变更内容",
+  "details": "详细的变更说明，按模块分组，使用 Markdown 格式",
+  "impact": "影响范围说明"
+}`, commitStr, diff)
+
+	messages := []goopenai.ChatCompletionMessage{
+		{Role: goopenai.ChatMessageRoleUser, Content: prompt},
+	}
+
+	// Call LLM with streaming
+	fullResponse, err := ChatStream(config, messages, callback)
+	if err != nil {
+		return nil, fmt.Errorf("LLM call: %w", err)
+	}
+
+	// Parse JSON response
+	jsonStr := extractJSON(fullResponse)
+	var pr PRDescription
+	if err := json.Unmarshal([]byte(jsonStr), &pr); err != nil {
+		return nil, fmt.Errorf("parse PR description: %w (raw: %s)", err, truncateString(fullResponse, 200))
+	}
+
+	// Send structured events
+	if callback != nil {
+		callback("title", map[string]interface{}{"content": pr.Title})
+		callback("type", map[string]interface{}{"content": pr.Type})
+		callback("summary", map[string]interface{}{"content": pr.Summary})
+		callback("detail", map[string]interface{}{"content": pr.Details})
+		callback("impact", map[string]interface{}{"content": pr.Impact})
+		callback("done", map[string]interface{}{"content": ""})
+	}
+
+	return &pr, nil
+}
+
+func truncateString(s string, maxLen int) string {
+	if len(s) <= maxLen {
+		return s
+	}
+	return s[:maxLen] + "..."
+}
@@ -0,0 +1,152 @@
+package services
+
+import (
+	"context"
+	"database/sql"
+	"fmt"
+	"io"
+	"strings"
+
+	goopenai "github.com/sashabaranov/go-openai"
+)
+
+// LLMConfig holds LLM API configuration.
+type LLMConfig struct {
+	Endpoint string
+	APIKey   string
+	Model    string
+}
+
+// GetLLMConfig reads LLM settings from the database.
+func GetLLMConfig(db *sql.DB) (LLMConfig, error) {
+	config := LLMConfig{}
+
+	rows, err := db.Query(`SELECT key, value FROM settings WHERE key IN ('llm.endpoint', 'llm.api_key', 'llm.model')`)
+	if err != nil {
+		return config, fmt.Errorf("read settings: %w", err)
+	}
+	defer rows.Close()
+
+	for rows.Next() {
+		var key, value string
+		if rows.Scan(&key, &value) == nil {
+			switch key {
+			case "llm.endpoint":
+				config.Endpoint = value
+			case "llm.api_key":
+				config.APIKey = value
+			case "llm.model":
+				config.Model = value
+			}
+		}
+	}
+
+	if config.APIKey == "" {
+		return config, fmt.Errorf("LLM API key not configured — please set it in the settings page")
+	}
+
+	if config.Model == "" {
+		config.Model = "gpt-4o"
+	}
+
+	return config, nil
+}
+
+// StreamCallback is called for each SSE event during LLM streaming.
+type StreamCallback func(event string, data interface{})
+
+// ChatStream sends a streaming chat completion request to an OpenAI-compatible API.
+// It calls callback with "content" events for each chunk received.
+// Returns the full concatenated response text.
+func ChatStream(config LLMConfig, messages []goopenai.ChatCompletionMessage, callback StreamCallback) (string, error) {
+	clientConfig := goopenai.DefaultConfig(config.APIKey)
+	if config.Endpoint != "" {
+		clientConfig.BaseURL = config.Endpoint
+	}
+	client := goopenai.NewClientWithConfig(clientConfig)
+
+	ctx := context.Background()
+	stream, err := client.CreateChatCompletionStream(ctx, goopenai.ChatCompletionRequest{
+		Model:    config.Model,
+		Messages: messages,
+		Stream:   true,
+	})
+	if err != nil {
+		return "", fmt.Errorf("create stream: %w", err)
+	}
+	defer stream.Close()
+
+	var fullResponse strings.Builder
+	for {
+		response, err := stream.Recv()
+		if err == io.EOF {
+			break
+		}
+		if err != nil {
+			return fullResponse.String(), fmt.Errorf("stream recv: %w", err)
+		}
+		if len(response.Choices) > 0 {
+			content := response.Choices[0].Delta.Content
+			if content != "" {
+				fullResponse.WriteString(content)
+				if callback != nil {
+					callback("content", map[string]interface{}{"content": content})
+				}
+			}
+		}
+	}
+
+	return fullResponse.String(), nil
+}
+
+// extractJSON attempts to extract a JSON object or array from a string
+// that may contain markdown code blocks or extra text.
+func extractJSON(s string) string {
+	// Try to find JSON in markdown code block
+	if idx := strings.Index(s, "```json"); idx >= 0 {
+		start := idx + 7
+		if end := strings.Index(s[start:], "```"); end >= 0 {
+			return strings.TrimSpace(s[start : start+end])
+		}
+	}
+	if idx := strings.Index(s, "```"); idx >= 0 {
+		start := idx + 3
+		if nl := strings.Index(s[start:], "\n"); nl >= 0 {
+			start += nl + 1
+		}
+		if end := strings.Index(s[start:], "```"); end >= 0 {
+			return strings.TrimSpace(s[start : start+end])
+		}
+	}
+
+	// Find first { or [
+	startObj := strings.Index(s, "{")
+	startArr := strings.Index(s, "[")
+
+	var start int
+	var endChar byte
+	if startObj >= 0 && (startArr < 0 || startObj < startArr) {
+		start = startObj
+		endChar = '}'
+	} else if startArr >= 0 {
+		start = startArr
+		endChar = ']'
+	} else {
+		return s
+	}
+
+	// Find matching closing bracket
+	depth := 0
+	for i := start; i < len(s); i++ {
+		if s[i] == '{' || s[i] == '[' {
+			depth++
+		} else if s[i] == '}' || s[i] == ']' {
+			depth--
+			if depth == 0 && s[i] == endChar {
+				return s[start : i+1]
+			}
+		}
+	}
+
+	return s[start:]
+}
@@ -0,0 +1,298 @@
+package services
+
+import (
+	"database/sql"
+	"encoding/json"
+	"fmt"
+	"sort"
+	"strings"
+
+	goopenai "github.com/sashabaranov/go-openai"
+)
+
+// ReviewSuggestion is a single review finding for a file.
+type ReviewSuggestion struct {
+	Severity    string `json:"severity"`
+	Description string `json:"description"`
+	Suggestion  string `json:"suggestion"`
+	CodeExample string `json:"code_example,omitempty"`
+}
+
+// FileReview holds the review results for one file.
+type FileReview struct {
+	FileName    string             `json:"file_name"`
+	ChangeLines int                `json:"change_lines"`
+	Suggestions []ReviewSuggestion `json:"suggestions"`
+	RawReview   string             `json:"raw_review"`
+}
+
+// ReviewSummary is the overall assessment after reviewing all files.
+type ReviewSummary struct {
+	Score           int    `json:"score"`
+	Overall         string `json:"overall"`
+	Findings        string `json:"findings"`
+	Recommendations string `json:"recommendations"`
+}
+
+// countDiffLines counts the number of added/removed lines in a diff patch.
+func countDiffLines(patch string) int {
+	count := 0
+	for _, line := range strings.Split(patch, "\n") {
+		if strings.HasPrefix(line, "+") || strings.HasPrefix(line, "-") {
+			if !strings.HasPrefix(line, "+++") && !strings.HasPrefix(line, "---") {
+				count++
+			}
+		}
+	}
+	return count
+}
+
+// GenerateReview performs AI code review on diff files with Top-N strategy.
+// It streams events (file_start, suggestion, file_end, summary, done) via callback.
+func GenerateReview(db *sql.DB, repoPath, base, head string, topN int, callback StreamCallback) error {
+	// Read LLM config
+	config, err := GetLLMConfig(db)
+	if err != nil {
+		return err
+	}
+
+	// Open repo
+	repo, err := OpenRepo(repoPath)
+	if err != nil {
+		return fmt.Errorf("open repo: %w", err)
+	}
+
+	// Get diff files
+	files, err := GetDiffFiles(repo, base, head)
+	if err != nil {
+		return fmt.Errorf("get diff files: %w", err)
+	}
+
+	if len(files) == 0 {
+		if callback != nil {
+			callback("summary", map[string]interface{}{"content": "没有检测到代码变更。"})
+			callback("done", map[string]interface{}{"content": ""})
+		}
+		return nil
+	}
+
+	// Sort by change size (descending)
+	sort.Slice(files, func(i, j int) bool {
+		return countDiffLines(files[i].Patch) > countDiffLines(files[j].Patch)
+	})
+
+	totalFiles := len(files)
+
+	// Apply Top-N (0 means analyze all)
+	if topN > 0 && topN < len(files) {
+		files = files[:topN]
+	}
+
+	reviewedFiles := len(files)
+
+	// Notify frontend of file count info
+	if callback != nil {
+		callback("progress", map[string]interface{}{
+			"total_files":    totalFiles,
+			"reviewed_files": reviewedFiles,
+			"top_n":          topN,
+		})
+	}
+
+	// Review each file
+	var fileReviews []FileReview
+	for i, file := range files {
+		if callback != nil {
+			callback("file_start", map[string]interface{}{
+				"file":  file.Filename,
+				"index": i + 1,
+				"total": reviewedFiles,
+			})
+		}
+
+		changeLines := countDiffLines(file.Patch)
+
+		// Truncate per-file diff if too large
+		patch := file.Patch
+		if len(patch) > 30000 {
+			patch = patch[:30000] + "\n\n... [diff truncated due to size]"
+		}
+
+		// Build review prompt
+		prompt := fmt.Sprintf(`你是一个资深代码审查专家。请审查以下代码变更，给出专业的 Review 意见。
+
+## 文件: %s
+## 变更行数: +%d / -%d
+
+## Diff
+%s
+
+请按以下 JSON 格式输出审查意见（直接输出 JSON 数组，不要包含 markdown 代码块标记）：
+[
+  {
+    "severity": "critical 或 warning 或 info",
+    "description": "问题描述",
+    "suggestion": "建议的修改方案",
+    "code_example": "建议的代码（如有）"
+  }
+]
+
+严重程度说明：
+- critical: 严重问题（安全漏洞、数据丢失风险、崩溃风险）
+- warning: 建议改进（性能问题、代码规范、可维护性）
+- info: 提示信息（最佳实践、可选优化）
+
+如果代码没有问题，输出空数组 []。
+请用中文回复。`, file.Filename, changeLines/2+changeLines%2, changeLines/2, patch)
+
+		messages := []goopenai.ChatCompletionMessage{
+			{Role: goopenai.ChatMessageRoleUser, Content: prompt},
+		}
+
+		// Call LLM
+		fullResponse, err := ChatStream(config, messages, callback)
+		if err != nil {
+			// Continue with other files on error
+			if callback != nil {
+				callback("error", map[string]interface{}{
+					"file":    file.Filename,
+					"message": err.Error(),
+				})
+			}
+			fileReviews = append(fileReviews, FileReview{
+				FileName:    file.Filename,
+				ChangeLines: changeLines,
+				Suggestions: nil,
+				RawReview:   fmt.Sprintf("Error: %s", err.Error()),
+			})
+			continue
+		}
+
+		// Parse suggestions
+		jsonStr := extractJSON(fullResponse)
+		var suggestions []ReviewSuggestion
+		if err := json.Unmarshal([]byte(jsonStr), &suggestions); err != nil {
+			// If parsing fails, try single object
+			var single ReviewSuggestion
+			if err2 := json.Unmarshal([]byte(jsonStr), &single); err2 == nil {
+				suggestions = []ReviewSuggestion{single}
+			} else {
+				// Fall back to raw text as info suggestion
+				suggestions = []ReviewSuggestion{{
+					Severity:    "info",
+					Description: fullResponse,
+				}}
+			}
+		}
+
+		// Validate severity values
+		for i := range suggestions {
+			switch suggestions[i].Severity {
+			case "critical", "warning", "info":
+				// valid
+			default:
+				suggestions[i].Severity = "info"
+			}
+		}
+
+		// Send suggestion events
+		for _, s := range suggestions {
+			content := s.Description
+			if s.Suggestion != "" {
+				content += "\n\n**建议修改:** " + s.Suggestion
+			}
+			if s.CodeExample != "" {
+				content += "\n\n```\n" + s.CodeExample + "\n```"
+			}
+			if callback != nil {
+				callback("suggestion", map[string]interface{}{
+					"file":     file.Filename,
+					"severity": s.Severity,
+					"content":  content,
+				})
+			}
+		}
+
+		fileReviews = append(fileReviews, FileReview{
+			FileName:    file.Filename,
+			ChangeLines: changeLines,
+			Suggestions: suggestions,
+			RawReview:   fullResponse,
+		})
+
+		if callback != nil {
+			callback("file_end", map[string]interface{}{"file": file.Filename})
+		}
+	}
+
+	// Generate summary
+	if len(fileReviews) > 0 {
+		if callback != nil {
+			callback("progress", map[string]interface{}{"step": "generating_summary"})
+		}
+
+		var reviewParts []string
+		for _, fr := range fileReviews {
+			suggestionText := "没有发现问题"
+			if len(fr.Suggestions) > 0 {
+				var parts []string
+				for _, s := range fr.Suggestions {
+					parts = append(parts, fmt.Sprintf("[%s] %s", s.Severity, s.Description))
+				}
+				suggestionText = strings.Join(parts, "\n")
+			}
+			reviewParts = append(reviewParts, fmt.Sprintf("### %s (%d 行变更)\n%s",
+				fr.FileName, fr.ChangeLines, suggestionText))
+		}
+
+		summaryPrompt := fmt.Sprintf(`以下是多个文件的代码审查结果，请给出整体评估：
+
+%s
+
+请按以下 JSON 格式输出（直接输出 JSON，不要包含 markdown 代码块标记）：
+{
+  "score": 7,
+  "overall": "总体评价（2-3 句话）",
+  "findings": "按严重程度排序的主要发现汇总",
+  "recommendations": "改进建议优先级列表"
+}
+请用中文回复。`, strings.Join(reviewParts, "\n\n"))
+
+		messages := []goopenai.ChatCompletionMessage{
+			{Role: goopenai.ChatMessageRoleUser, Content: summaryPrompt},
+		}
+
+		summaryResponse, err := ChatStream(config, messages, callback)
+		if err == nil {
+			jsonStr := extractJSON(summaryResponse)
+			var summary ReviewSummary
+			if json.Unmarshal([]byte(jsonStr), &summary) == nil {
+				summaryText := fmt.Sprintf("整体评分: %d/10\n\n%s", summary.Score, summary.Overall)
+				if summary.Findings != "" {
+					summaryText += "\n\n**主要发现:**\n" + summary.Findings
+				}
+				if summary.Recommendations != "" {
+					summaryText += "\n\n**改进建议:**\n" + summary.Recommendations
+				}
+				if callback != nil {
+					callback("summary", map[string]interface{}{"content": summaryText})
+				}
+			} else {
+				// If parsing fails, send raw response as summary
+				if callback != nil {
+					callback("summary", map[string]interface{}{"content": summaryResponse})
+				}
+			}
+		} else {
+			if callback != nil {
+				callback("error", map[string]interface{}{"message": "生成汇总失败: " + err.Error()})
+			}
+		}
+	}
+
+	if callback != nil {
+		callback("done", map[string]interface{}{"content": ""})
+	}
+
+	return nil
+}