knowledge-governance/internal/httpapi/source_available.go

package httpapi

import (
	"context"
	"errors"
	"net/http"
	"strconv"
	"strings"
	"time"

	"knowledge-governance/internal/dingtalkdiscovery"
)

// dingtalkDocLister 是本路由需要的文档能力面：列在线文档/表格、并真实读一次确认可读。
// 与知识库/钉盘共用同一份 dws 凭据，注入方在 cmd/api。
type dingtalkDocLister interface {
	ListDingtalkDocuments(ctx context.Context, query string, limit int) (dingtalkdiscovery.DocResult, error)
	ValidateDingtalkDocument(ctx context.Context, node string) (dingtalkdiscovery.DocumentNode, error)
}

// availableItem 是「添加数据源」向导第一步展示的候选条目，四类来源同一形状。
type availableItem struct {
	Ref         string `json:"ref"`
	Name        string `json:"name"`
	Kind        string `json:"kind"`
	ContentType string `json:"content_type,omitempty"`
	URL         string `json:"url,omitempty"`
	UpdatedAt   string `json:"updated_at,omitempty"`
}

var availableSourceTypes = map[string]bool{"wiki": true, "drive": true, "doc": true, "pp": true}

// listAvailableSources 处理 GET /api/v1/source-catalog/available?type=&q=&limit=：
// 向导用它列「可加入的数据源」。类型不同，形状相同——前端不需要为每类写一套。
func listAvailableSources(w http.ResponseWriter, r *http.Request, config Config) {
	if _, ok := requireSessionIdentity(w, r); !ok {
		return
	}
	kind := strings.TrimSpace(r.URL.Query().Get("type"))
	if !availableSourceTypes[kind] {
		writeJSON(w, http.StatusBadRequest, map[string]string{"error": "type 必须是 wiki、drive、doc 或 pp"})
		return
	}
	ctx, cancel := context.WithTimeout(r.Context(), 30*time.Second)
	defer cancel()
	limit := 50
	if raw := strings.TrimSpace(r.URL.Query().Get("limit")); raw != "" {
		if parsed, err := strconv.Atoi(raw); err == nil && parsed > 0 && parsed <= 200 {
			limit = parsed
		}
	}
	query := strings.TrimSpace(r.URL.Query().Get("q"))
	items := make([]availableItem, 0, 16)
	notice := ""
	switch kind {
	case "wiki", "drive":
		lister := config.DingtalkSources
		if lister == nil {
			writeJSON(w, http.StatusNotImplemented, map[string]string{"error": "dingtalk source discovery is not wired"})
			return
		}
		result, err := lister.ListAvailableDingtalkSources(ctx)
		if err != nil {
			writeDiscoveryFailure(w, err)
			return
		}
		if kind == "wiki" {
			for _, base := range result.KnowledgeBases {
				items = append(items, availableItem{Ref: base.URL, Name: base.Name, Kind: "wiki", URL: base.URL})
			}
			notice = result.KnowledgeNotice
		} else {
			for _, space := range result.DriveSpaces {
				items = append(items, availableItem{Ref: dingtalkdiscovery.SpaceURL(space.SpaceID), Name: space.Name, Kind: "drive", URL: dingtalkdiscovery.SpaceURL(space.SpaceID)})
			}
			notice = result.DriveNotice
		}
	case "doc":
		lister := config.DingtalkDocuments
		if lister == nil {
			writeJSON(w, http.StatusNotImplemented, map[string]string{"error": "dingtalk document discovery is not wired"})
			return
		}
		result, err := lister.ListDingtalkDocuments(ctx, query, limit)
		if err != nil {
			writeDiscoveryFailure(w, err)
			return
		}
		for _, doc := range result.Documents {
			items = append(items, availableItem{
				Ref: doc.NodeID, Name: doc.Name, Kind: "doc",
				ContentType: doc.ContentType, URL: doc.URL, UpdatedAt: doc.UpdatedAt,
			})
		}
		notice = result.Notice
	case "pp":
		// PP 是公开只读目录：没有可枚举的账号级列表，向导要求直接给资料中心地址，
		// 由第 4 步的校验去真打一次。这里返回空列表并说明原因，不伪装成"没权限"。
		notice = "PP 资料下载中心没有可列举的账号级列表，请在第 2 步填写资料中心地址。"
	}
	writeJSON(w, http.StatusOK, map[string]any{"data": map[string]any{
		"type": kind, "items": items, "notice": notice, "fetched_at": time.Now().UTC(),
	}})
}

// validateSourceSelection 处理 POST /api/v1/source-catalog/validate：
// 添加前真读一次，读不到就不让进下一步——避免加入一堆永远同步不了的来源。
func validateSourceSelection(w http.ResponseWriter, r *http.Request, config Config) {
	if _, ok := requireSessionIdentity(w, r); !ok {
		return
	}
	var body struct {
		Type string   `json:"type"`
		Refs []string `json:"refs"`
	}
	if decodeJSONStrict(r, &body) != nil {
		writeJSON(w, http.StatusBadRequest, map[string]string{"error": "invalid_request_body"})
		return
	}
	kind := strings.TrimSpace(body.Type)
	if !availableSourceTypes[kind] || len(body.Refs) == 0 || len(body.Refs) > 50 {
		writeJSON(w, http.StatusBadRequest, map[string]string{"error": "type 无效，或一次校验的条目数不在 1–50 之间"})
		return
	}
	ctx, cancel := context.WithTimeout(r.Context(), 60*time.Second)
	defer cancel()
	switch kind {
	case "doc":
		lister := config.DingtalkDocuments
		if lister == nil {
			writeJSON(w, http.StatusNotImplemented, map[string]string{"error": "dingtalk document discovery is not wired"})
			return
		}
		verified := make([]map[string]string, 0, len(body.Refs))
		for _, ref := range body.Refs {
			doc, err := lister.ValidateDingtalkDocument(ctx, ref)
			if err != nil {
				writeDiscoveryFailure(w, err)
				return
			}
			verified = append(verified, map[string]string{"ref": doc.NodeID, "name": doc.Name, "content_type": doc.ContentType})
		}
		writeJSON(w, http.StatusOK, map[string]any{"data": map[string]any{
			"ok": true, "type": kind, "items": verified, "discovered": len(verified),
		}})
	case "wiki", "drive":
		lister := config.DingtalkSources
		if lister == nil {
			writeJSON(w, http.StatusNotImplemented, map[string]string{"error": "dingtalk source discovery is not wired"})
			return
		}
		result, err := lister.ListAvailableDingtalkSources(ctx)
		if err != nil {
			writeDiscoveryFailure(w, err)
			return
		}
		known := map[string]string{}
		if kind == "wiki" {
			for _, base := range result.KnowledgeBases {
				known[base.URL] = base.Name
			}
		} else {
			for _, space := range result.DriveSpaces {
				known[dingtalkdiscovery.SpaceURL(space.SpaceID)] = space.Name
			}
		}
		verified := make([]map[string]string, 0, len(body.Refs))
		for _, ref := range body.Refs {
			name, ok := known[strings.TrimSpace(ref)]
			if !ok {
				writeJSON(w, http.StatusConflict, map[string]string{"error": "当前钉钉账号已看不到该来源，请刷新后再选"})
				return
			}
			verified = append(verified, map[string]string{"ref": ref, "name": name})
		}
		writeJSON(w, http.StatusOK, map[string]any{"data": map[string]any{
			"ok": true, "type": kind, "items": verified, "discovered": len(verified),
		}})
	case "pp":
		// PP 的可用性由真正的下载探测在登记流程里判定，这里只做形式校验。
		for _, ref := range body.Refs {
			if !strings.HasPrefix(strings.TrimSpace(ref), "http") {
				writeJSON(w, http.StatusBadRequest, map[string]string{"error": "PP 需要 http(s) 资料中心地址"})
				return
			}
		}
		writeJSON(w, http.StatusOK, map[string]any{"data": map[string]any{"ok": true, "type": kind, "discovered": len(body.Refs)}})
	}
}

func writeDiscoveryFailure(w http.ResponseWriter, err error) {
	if errors.Is(err, dingtalkdiscovery.ErrUnavailable) {
		writeJSON(w, http.StatusServiceUnavailable, map[string]string{
			"error": "dingtalk_discovery_unavailable",
			"hint":  "dws 未登录或缺少读权限；请刷新钉钉会话，或让管理员检查 dws 凭据。",
		})
		return
	}
	writeJSON(w, http.StatusInternalServerError, map[string]string{"error": "could not list sources"})
}