xk-ai-agent/internal/handler/kb_crawl_handler.go

package handler

import (
	"net/http"
	"strconv"

	"tcm-agent/internal/crawler"
	"tcm-agent/internal/dao"
	"tcm-agent/internal/service"

	"github.com/gin-gonic/gin"
)

// ========================================================================
//  KBCrawlHandler —— 药品抓取任务管理 API
//  ========================================================================
//  暴露 7 个端点（前缀 /api/v1/kb/admin/crawl，复用 KB 管理鉴权：
//  口令头或面板 JWT 均可）：
//
//    GET    /sources           可用抓取源列表（前端下拉）
//    GET    /tasks             任务列表（含实时 running 标记）
//    POST   /tasks             创建任务
//    PUT    /tasks/:id         更新任务（名称/目标库/调度/限量/启停）
//    DELETE /tasks/:id         软删除任务
//    POST   /tasks/:id/run     立即抓取（异步，立即返回）
//    GET    /tasks/:id/logs    任务运行历史（最近 N 条）
//  ========================================================================

// KBCrawlHandler 抓取任务处理器（无状态，直接调 dao/service）
type KBCrawlHandler struct{}

// NewKBCrawlHandler 构造
func NewKBCrawlHandler() *KBCrawlHandler { return &KBCrawlHandler{} }

// ListSources 可用抓取源
// GET /api/v1/kb/admin/crawl/sources
func (h *KBCrawlHandler) ListSources(c *gin.Context) {
	c.JSON(http.StatusOK, gin.H{"code": 200, "data": crawler.ListSources()})
}

// ListTasks 任务列表
// GET /api/v1/kb/admin/crawl/tasks
//
// 每行附加 running 字段（内存实时状态）——DB 的 last_status 有落库延迟，
// 前端旋转图标要跟内存状态走
func (h *KBCrawlHandler) ListTasks(c *gin.Context) {
	rows, err := dao.KBCrawlTaskList()
	if err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 500, "message": err.Error()})
		return
	}
	out := make([]gin.H, 0, len(rows))
	for _, t := range rows {
		out = append(out, gin.H{
			"id": t.ID, "name": t.Name, "source": t.Source,
			"library_id": t.LibraryID, "schedule_type": t.ScheduleType,
			"interval_hours": t.IntervalHours, "run_at_hour": t.RunAtHour,
			"run_at_weekday": t.RunAtWeekday, "items_per_run": t.ItemsPerRun,
			"progress_offset": t.ProgressOffset, "status": t.Status,
			"last_run_at": t.LastRunAt, "last_status": t.LastStatus,
			"last_message": t.LastMessage, "created_at": t.CreatedAt,
			"running": service.CrawlTaskRunning(t.ID),
		})
	}
	c.JSON(http.StatusOK, gin.H{"code": 200, "data": out})
}

// crawlTaskRequest 创建/更新任务的入参
type crawlTaskRequest struct {
	Name          string `json:"name"`
	Source        string `json:"source"`
	LibraryID     uint   `json:"library_id"`
	ScheduleType  string `json:"schedule_type"`
	IntervalHours int    `json:"interval_hours"`
	RunAtHour     int    `json:"run_at_hour"`
	RunAtWeekday  int    `json:"run_at_weekday"`
	ItemsPerRun   int    `json:"items_per_run"`
	Status        *int   `json:"status"` // 指针区分"没传"和"传了 0（禁用）"
}

// normalize 归一化 + 校验入参（创建和更新共用）
func (r *crawlTaskRequest) normalize() string {
	if r.Name == "" {
		return "任务名称不能为空"
	}
	if r.Source == "" {
		r.Source = "zhongyoo"
	}
	if _, ok := crawler.GetSource(r.Source); !ok {
		return "抓取源不存在: " + r.Source
	}
	if r.LibraryID == 0 {
		return "请选择目标知识库"
	}
	switch r.ScheduleType {
	case "", "manual":
		r.ScheduleType = "manual"
	case "interval":
		if r.IntervalHours < 1 {
			r.IntervalHours = 24
		}
	case "daily", "weekly":
		if r.RunAtHour < 0 || r.RunAtHour > 23 {
			r.RunAtHour = 3
		}
		if r.RunAtWeekday < 0 || r.RunAtWeekday > 6 {
			r.RunAtWeekday = 1
		}
	default:
		return "调度类型不合法: " + r.ScheduleType
	}
	if r.ItemsPerRun < 1 || r.ItemsPerRun > 500 {
		r.ItemsPerRun = 50
	}
	return ""
}

// CreateTask 创建任务
// POST /api/v1/kb/admin/crawl/tasks
func (h *KBCrawlHandler) CreateTask(c *gin.Context) {
	var req crawlTaskRequest
	if err := c.ShouldBindJSON(&req); err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 400, "message": "参数格式错误: " + err.Error()})
		return
	}
	if msg := req.normalize(); msg != "" {
		c.JSON(http.StatusOK, gin.H{"code": 400, "message": msg})
		return
	}
	// 目标库必须真实存在（防手滑填错 ID，跑的时候才发现）
	if _, err := dao.KBGetLibrary(req.LibraryID); err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 400, "message": "目标知识库不存在，请先在知识库管理里创建"})
		return
	}
	row := &dao.KBCrawlTaskRow{
		Name: req.Name, Source: req.Source, LibraryID: req.LibraryID,
		ScheduleType: req.ScheduleType, IntervalHours: req.IntervalHours,
		RunAtHour: req.RunAtHour, RunAtWeekday: req.RunAtWeekday,
		ItemsPerRun: req.ItemsPerRun, Status: 1,
	}
	if err := dao.KBCrawlTaskCreate(row); err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 500, "message": err.Error()})
		return
	}
	c.JSON(http.StatusOK, gin.H{"code": 200, "data": row, "message": "任务已创建"})
}

// UpdateTask 更新任务
// PUT /api/v1/kb/admin/crawl/tasks/:id
func (h *KBCrawlHandler) UpdateTask(c *gin.Context) {
	id, err := parseUintParam(c, "id")
	if err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 400, "message": err.Error()})
		return
	}
	var req crawlTaskRequest
	if err := c.ShouldBindJSON(&req); err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 400, "message": "参数格式错误: " + err.Error()})
		return
	}
	if msg := req.normalize(); msg != "" {
		c.JSON(http.StatusOK, gin.H{"code": 400, "message": msg})
		return
	}
	if _, err := dao.KBGetLibrary(req.LibraryID); err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 400, "message": "目标知识库不存在"})
		return
	}
	fields := map[string]any{
		"name": req.Name, "source": req.Source, "library_id": req.LibraryID,
		"schedule_type": req.ScheduleType, "interval_hours": req.IntervalHours,
		"run_at_hour": req.RunAtHour, "run_at_weekday": req.RunAtWeekday,
		"items_per_run": req.ItemsPerRun,
	}
	if req.Status != nil {
		fields["status"] = *req.Status
	}
	if err := dao.KBCrawlTaskUpdate(id, fields); err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 500, "message": err.Error()})
		return
	}
	c.JSON(http.StatusOK, gin.H{"code": 200, "message": "任务已更新"})
}

// DeleteTask 软删除任务
// DELETE /api/v1/kb/admin/crawl/tasks/:id
func (h *KBCrawlHandler) DeleteTask(c *gin.Context) {
	id, err := parseUintParam(c, "id")
	if err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 400, "message": err.Error()})
		return
	}
	if service.CrawlTaskRunning(id) {
		c.JSON(http.StatusOK, gin.H{"code": 400, "message": "任务正在运行中，等本次跑完再删除"})
		return
	}
	if err := dao.KBCrawlTaskDelete(id); err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 500, "message": err.Error()})
		return
	}
	c.JSON(http.StatusOK, gin.H{"code": 200, "message": "任务已删除"})
}

// RunTask 立即抓取（异步触发）
// POST /api/v1/kb/admin/crawl/tasks/:id/run
func (h *KBCrawlHandler) RunTask(c *gin.Context) {
	id, err := parseUintParam(c, "id")
	if err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 400, "message": err.Error()})
		return
	}
	if err := service.TriggerCrawlTask(id, "manual"); err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 400, "message": err.Error()})
		return
	}
	c.JSON(http.StatusOK, gin.H{
		"code":    200,
		"message": "抓取已在后台启动，可在运行历史里查看进度（每批约需 1-3 分钟）",
	})
}

// ListLogs 任务运行历史
// GET /api/v1/kb/admin/crawl/tasks/:id/logs?limit=20
func (h *KBCrawlHandler) ListLogs(c *gin.Context) {
	id, err := parseUintParam(c, "id")
	if err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 400, "message": err.Error()})
		return
	}
	limit, _ := strconv.Atoi(c.DefaultQuery("limit", "20"))
	rows, err := dao.KBCrawlLogList(id, limit)
	if err != nil {
		c.JSON(http.StatusOK, gin.H{"code": 500, "message": err.Error()})
		return
	}
	c.JSON(http.StatusOK, gin.H{"code": 200, "data": rows})
}