pj0235-eai_agentplatform/eai_agentplatform/backend-go/internal/api/knowledge.go

package api

import (
	"encoding/json"
	"fmt"
	"os"
	"path/filepath"
	"strconv"
	"strings"
	"time"

	"github.com/gin-gonic/gin"

	"eai_agentplatform/backend/internal/middleware"
	"eai_agentplatform/backend/internal/model"
	"eai_agentplatform/backend/internal/repository"
	"eai_agentplatform/backend/internal/web"
)

// sourceRepo 知识源仓库（便于测试时覆写），包内共享。
var sourceRepo repository.KnowledgeSourceRepo

func init() {
	sourceRepo = repository.KnowledgeSourceRepo{}
}

// ============ 扫描 ============

// KnowledgeScan POST /api/knowledge/scan —— 扫描 knowledge_source 目录建 pending 记录
func KnowledgeScan(c *gin.Context) {
	entries, err := os.ReadDir(Cfg.KnowledgeSourceDir)
	if err != nil {
		web.Fail(c, web.NewBadRequest("知识源目录不存在："+Cfg.KnowledgeSourceDir))
		return
	}
	results := make([]gin.H, 0, len(entries))
	for _, e := range entries {
		if e.IsDir() || !strings.HasSuffix(strings.ToLower(e.Name()), ".md") {
			continue
		}
		if strings.HasPrefix(strings.ToUpper(e.Name()), "README") {
			continue
		}
		full := filepath.Join(Cfg.KnowledgeSourceDir, e.Name())
		data, err := os.ReadFile(full)
		if err != nil {
			continue
		}
		fm := parseFrontMatter(string(data))
		if fm["category"] == "" || fm["version"] == "" {
			continue // 非知识源文档，跳过
		}

		if existing, found := sourceRepo.GetByFilePath(e.Name()); found {
			results = append(results, gin.H{"file_path": e.Name(), "status": "skipped", "title": existing.Title})
			continue
		}
		src := model.KnowledgeSource{
			Title:             parseTitle(string(data)),
			FilePath:          e.Name(),
			Category:          fm["category"],
			Domain:            orDefault(fm["domain"], "product"),
			SourceVersion:     fm["version"],
			AuditStatus:       "pending",
			KnowledgeSpaceKey: ensureKnowledgeSpaceKeyOrDefault(orDefault(fm["knowledge_space_key"], inferKnowledgeSpaceKey(parseTitle(string(data)), fm["domain"], fm["category"]+" "+e.Name()))),
		}
		if !sourceRepo.Insert(&src) {
			results = append(results, gin.H{"file_path": e.Name(), "status": "error", "title": src.Title})
			continue
		}
		results = append(results, gin.H{"file_path": e.Name(), "status": "created", "title": src.Title, "knowledge_space_key": src.KnowledgeSpaceKey})
	}
	web.OK(c, gin.H{"results": results})
}

// KnowledgeAuditList GET /api/knowledge/audit-list?status=&page=&size=
func KnowledgeAuditList(c *gin.Context) {
	page, _ := strconv.Atoi(c.DefaultQuery("page", "1"))
	size, _ := strconv.Atoi(c.DefaultQuery("size", "20"))
	if page < 1 {
		page = 1
	}
	if size < 1 || size > 100 {
		size = 20
	}
	total, items := sourceRepo.ListForAudit(
		c.Query("status"), sanitizeSpaceKey(c.Query("knowledge_space_key")), page, size)
	web.OK(c, gin.H{"total": total, "items": items})
}

// KnowledgeAudit POST /api/knowledge/audit/{source_id} —— 审批（approve→摄入 / reject→驳回）
func KnowledgeAudit(c *gin.Context) {
	u := middleware.CurrentUser(c)
	id, ok := parseID(c, "source_id")
	if !ok {
		return
	}
	src, found := sourceRepo.GetByID(id)
	if !found {
		web.Fail(c, web.NewNotFoundError("知识源不存在"))
		return
	}
	var req struct {
		Action       string `json:"action"`
		RejectReason string `json:"reject_reason"`
	}
	if err := c.ShouldBindJSON(&req); err != nil {
		web.Fail(c, web.NewBadRequest("请求参数错误"))
		return
	}
	now := time.Now()
	auditBy := u.ID
	switch req.Action {
	case "approve":
		if src.Ingested {
			web.OK(c, gin.H{"status": "approved", "source_id": src.ID, "products": 0, "chunks": 0, "questions": 0, "already_ingested": true})
			return
		}
		counts, err := ingestSource(&src)
		if err != nil {
			web.Fail(c, web.NewBadRequest("摄入失败："+err.Error()))
			return
		}
		src.AuditStatus = "approved"
		src.AuditBy = &auditBy
		src.AuditAt = &now
		src.RejectReason = ""
		src.Ingested = true
		if !sourceRepo.Update(&src) {
			web.Fail(c, web.NewBadRequest("审批失败"))
			return
		}
		triggerKnowledgeIndexRebuild()
		web.OK(c, gin.H{
			"status": "approved", "source_id": src.ID,
			"products": counts[0], "chunks": counts[1], "questions": counts[2],
		})
	case "reject":
		if strings.TrimSpace(req.RejectReason) == "" {
			web.Fail(c, web.NewBadRequest("驳回理由必填"))
			return
		}
		src.AuditStatus = "rejected"
		src.RejectReason = req.RejectReason
		src.AuditBy = &auditBy
		src.AuditAt = &now
		if !sourceRepo.Update(&src) {
			web.Fail(c, web.NewBadRequest("审批失败"))
			return
		}
		web.OK(c, gin.H{"status": "rejected", "source_id": src.ID})
	default:
		web.Fail(c, web.NewBadRequest("action 必须为 approve 或 reject"))
	}
}

// KnowledgeStatus GET /api/knowledge/status/{source_id}
func KnowledgeStatus(c *gin.Context) {
	id, ok := parseID(c, "source_id")
	if !ok {
		return
	}
	src, found := sourceRepo.GetByID(id)
	if !found {
		web.Fail(c, web.NewNotFoundError("知识源不存在"))
		return
	}
	web.OK(c, gin.H{
		"audit_status":        src.AuditStatus,
		"ingested":            src.Ingested,
		"reject_reason":       src.RejectReason,
		"knowledge_space_key": resolveKnowledgeSourceSpaceKey(src),
	})
}

// ============ md 解析 ============

func parseFrontMatter(content string) map[string]string {
	m := map[string]string{}
	parts := strings.SplitN(content, "---", 3)
	if len(parts) < 3 {
		return m
	}
	for _, line := range strings.Split(parts[1], "\n") {
		line = strings.TrimSpace(line)
		if line == "" || !strings.Contains(line, ":") {
			continue
		}
		kv := strings.SplitN(line, ":", 2)
		m[strings.TrimSpace(kv[0])] = strings.TrimSpace(kv[1])
	}
	return m
}

func parseTitle(content string) string {
	for _, line := range strings.Split(content, "\n") {
		line = strings.TrimSpace(line)
		if strings.HasPrefix(line, "# ") {
			return strings.TrimSpace(strings.TrimPrefix(line, "# "))
		}
	}
	return ""
}

type mdBlock struct {
	title string
	body  string
}

func splitSections(content string) map[string]string {
	sections := map[string]string{}
	lines := strings.Split(content, "\n")
	current := ""
	var buf strings.Builder
	for _, line := range lines {
		if strings.HasPrefix(line, "## ") {
			if current != "" {
				sections[current] = buf.String()
			}
			current = strings.TrimSpace(strings.TrimPrefix(line, "## "))
			buf.Reset()
		} else {
			buf.WriteString(line + "\n")
		}
	}
	if current != "" {
		sections[current] = buf.String()
	}
	return sections
}

func splitBlocks(body string) []mdBlock {
	var blocks []mdBlock
	current := ""
	var buf strings.Builder
	flush := func() {
		if current != "" || strings.TrimSpace(buf.String()) != "" {
			blocks = append(blocks, mdBlock{current, buf.String()})
		}
	}
	for _, line := range strings.Split(body, "\n") {
		if strings.HasPrefix(line, "### ") {
			flush()
			current = strings.TrimSpace(strings.TrimPrefix(line, "### "))
			buf.Reset()
		} else {
			buf.WriteString(line + "\n")
		}
	}
	flush()
	return blocks
}

func parseKV(body string) map[string]string {
	m := map[string]string{}
	for _, line := range strings.Split(body, "\n") {
		line = strings.TrimSpace(line)
		if !strings.HasPrefix(line, "- ") {
			continue
		}
		line = strings.TrimPrefix(line, "- ")
		if i := strings.Index(line, ":"); i >= 0 {
			m[strings.TrimSpace(line[:i])] = strings.TrimSpace(line[i+1:])
		}
	}
	return m
}

// indexKey 选项索引 → 键（0→A, 1→B ...）
func indexKey(i int) string {
	if i < 0 || i > 25 {
		return "?"
	}
	return string(rune('A' + i))
}

func convertOptions(raw []string) []Option {
	opts := make([]Option, len(raw))
	for i, text := range raw {
		opts[i] = Option{Key: indexKey(i), Text: text}
	}
	return opts
}

func convertAnswer(raw []any) []string {
	out := make([]string, 0, len(raw))
	for _, a := range raw {
		switch v := a.(type) {
		case bool:
			out = append(out, strconv.FormatBool(v))
		case float64:
			out = append(out, indexKey(int(v)))
		case string:
			out = append(out, v)
		}
	}
	return out
}

// ============ 摄入 ============

// ingestSource 解析 md → product / knowledge_chunk / question，返回三表计数
//
// 口径提示：counts 记的是**从 md 解析出的条目数**，不是落库成功数 ——
// 单条写入失败不中断、也不扣减计数（沿用原有行为，未改）。
func ingestSource(src *model.KnowledgeSource) ([3]int, error) {
	var counts [3]int
	full := filepath.Join(Cfg.KnowledgeSourceDir, src.FilePath)
	data, err := os.ReadFile(full)
	if err != nil {
		return counts, fmt.Errorf("读取文件失败")
	}
	content := string(data)
	fm := parseFrontMatter(content)
	cat := orDefault(src.Category, fm["category"])
	sections := splitSections(content)

	// 1. 产品
	for _, b := range splitBlocks(sections["结构化产品数据"]) {
		kv := parseKV(b.body)
		if kv["code"] == "" || kv["name"] == "" {
			continue
		}
		p := model.Product{
			Code:                kv["code"],
			Name:                kv["name"],
			Category:            orDefault(kv["category"], cat),
			Tags:                kv["tags"],
			Description:         kv["description"],
			Pricing:             kv["pricing"],
			CommissionRecommend: kv["commission_recommend"],
			CommissionNegotiate: kv["commission_negotiate"],
			PublicCourseBonus:   kv["public_course_bonus"],
			VersionRisk:         kv["version_risk"],
			ReportRules:         kv["report_rules"],
			Status:              "active",
		}
		if existing, found := productRepo.GetByCode(p.Code); found {
			p.ID = existing.ID
			p.CreatedAt = existing.CreatedAt // Save 整字段覆盖，不回填会把创建时间写成零值
			productRepo.Update(&p)
		} else {
			productRepo.Insert(&p)
		}
		counts[0]++
	}

	// 2. AI 检索知识 → knowledge_chunk
	for i, b := range splitBlocks(sections["AI 检索知识"]) {
		text := strings.TrimSpace(b.body)
		if text == "" {
			continue
		}
		chunkRepo.Insert(&model.KnowledgeChunk{
			KnowledgeSourceID: &src.ID,
			SourceType:        "md",
			SourceID:          strconv.FormatUint(uint64(src.ID), 10),
			KnowledgeSpaceKey: resolveKnowledgeSourceSpaceKey(*src),
			ChunkIndex:        i,
			Content:           text,
		})
		counts[1]++
	}

	// 3. 考试题目 → question
	for _, b := range splitBlocks(sections["考试题目"]) {
		kv := parseKV(b.body)
		if kv["type"] == "" || kv["stem"] == "" {
			continue
		}
		var optsRaw []string
		_ = json.Unmarshal([]byte(kv["options"]), &optsRaw)
		var ansRaw []any
		_ = json.Unmarshal([]byte(kv["answer"]), &ansRaw)
		optsJSON, _ := json.Marshal(convertOptions(optsRaw))
		ansJSON, _ := json.Marshal(convertAnswer(ansRaw))
		q := model.Question{
			Domain:      orDefault(kv["domain"], src.Domain),
			Type:        kv["type"],
			Stem:        kv["stem"],
			Options:     string(optsJSON),
			Answer:      string(ansJSON),
			Explanation: kv["explanation"],
			Status:      "active",
		}
		questionRepo.Insert(&q)
		counts[2]++
	}

	return counts, nil
}

func orDefault(v, def string) string {
	if strings.TrimSpace(v) == "" {
		return def
	}
	return v
}