Agent-System/internal/server/chat_document_reader.go

package server

import (
	"context"
	"fmt"
	"strings"
	"unicode"

	"chat-agent-system/internal/auth"
	"chat-agent-system/internal/chat"
	"chat-agent-system/internal/coordinator"
	"chat-agent-system/internal/docgov"
	"chat-agent-system/internal/intent"
	"chat-agent-system/internal/projectinstance"
)

// bindSessionDocumentSummaryReader 将正文访问绑定到已认证的用户与项目。
// 读取前先应用与文档列表/正文 API 相同的阶段权限过滤。
func (s *Server) bindSessionDocumentSummaryReader(sess *chat.Session, user *auth.UserAccount) {
	if sess == nil || sess.UserCtx == nil {
		return
	}
	sess.UserCtx.DocumentSummaryReader = nil
	projectID := sess.Project()
	userID := int64(0)
	if user != nil {
		userID = user.ID
	}
	if projectID <= 0 || userID <= 0 || s.gov == nil {
		return
	}

	sess.UserCtx.DocumentSummaryReader = func(ctx context.Context, target string) (*coordinator.ReadableDocument, error) {
		if s.authorizer == nil {
			return nil, fmt.Errorf("project authorization unavailable")
		}
		if err := s.authorizer.RequireProjectAccessOrAdmin(ctx, projectID, userID, s.isSystemAdminUser(ctx, user)); err != nil {
			return nil, err
		}
		docs, err := s.gov.ListWorkflowDocumentsByProject(ctx, projectID)
		if err != nil {
			return nil, err
		}
		docs = s.filterWorkflowDocsByStageAccess(ctx, projectID, userID, docs)
		// 工作流作用域：仅在激活工作流的文档池内解析正文；跨流点名返回可解释的引导错误。
		// 激活键由 enrichSessionWorkflowStage 每轮刷新；任务集合按 live 实例快照解析。
		workflowKey, scopeTaskKeys := s.activeWorkflowScopeForDocs(ctx, projectID)
		if sess.UserCtx != nil && sess.UserCtx.Project.ActiveWorkflowKey != "" {
			workflowKey = sess.UserCtx.Project.ActiveWorkflowKey
		}
		doc, err := selectReadableWorkflowDocument(docs, target, workflowKey, scopeTaskKeys)
		if err != nil {
			return nil, err
		}
		content, err := s.loadControlledDocumentContent(ctx, doc.DocID)
		if err != nil {
			return nil, err
		}
		return &coordinator.ReadableDocument{
			DocID:          doc.DocID,
			Title:          doc.Title,
			TaskKey:        doc.TaskKey,
			LifecycleState: string(doc.LifecycleState),
			Version:        doc.Version,
			Markdown:       content.Markdown,
			IsTemplateOnly: doc.IsTemplateOnly || content.IsTemplateOnly,
		}, nil
	}
}

// activeWorkflowScopeForDocs 读取项目激活工作流键与任务集合；失败时返回空（= 不过滤）。
func (s *Server) activeWorkflowScopeForDocs(ctx context.Context, projectID int64) (string, map[string]bool) {
	if s == nil || s.instSvc == nil || !s.instSvc.Available() || projectID <= 0 {
		return "", nil
	}
	workflowKey, taskKeys, err := s.instSvc.ActiveWorkflowScope(ctx, projectID)
	if err != nil {
		return "", nil
	}
	return workflowKey, taskKeys
}

// taskKeyOutsideActiveWorkflow 判断任务键是否落在激活工作流之外。
// 未激活 / 任务键未知归属时放行（存量兼容）；跨流时返回引导话术。
func (s *Server) taskKeyOutsideActiveWorkflow(ctx context.Context, projectID int64, taskKey string) (bool, string) {
	taskKey = strings.TrimSpace(taskKey)
	if taskKey == "" || projectID <= 0 {
		return false, ""
	}
	workflowKey, scopeTaskKeys := s.activeWorkflowScopeForDocs(ctx, projectID)
	if workflowKey == "" {
		return false, ""
	}
	if len(scopeTaskKeys) > 0 {
		if projectinstance.TaskKeyInScope(scopeTaskKeys, taskKey) {
			return false, ""
		}
		owner := intent.TaskKeyWorkflowOwner(taskKey)
		if owner == "" || owner == workflowKey {
			return false, ""
		}
		return true, intent.WorkflowScopeGuidanceQuestion(taskKey, []string{owner}, workflowKey)
	}
	owner := intent.TaskKeyWorkflowOwner(taskKey)
	if owner == "" || owner == workflowKey {
		return false, ""
	}
	return true, intent.WorkflowScopeGuidanceQuestion(taskKey, []string{owner}, workflowKey)
}

// workflowKeyForDocument 为列表项标注归属工作流：优先激活流任务集合，其次任务键前缀推断。
func workflowKeyForDocument(taskKey, activeKey string, scopeTaskKeys map[string]bool) string {
	if activeKey != "" && len(scopeTaskKeys) > 0 && projectinstance.TaskKeyInScope(scopeTaskKeys, taskKey) {
		return activeKey
	}
	if owner := intent.TaskKeyWorkflowOwner(taskKey); owner != "" {
		return owner
	}
	return activeKey
}

// errDocWrongWorkflow 目标文档存在但不属于当前激活工作流。
type errDocWrongWorkflow struct {
	Title         string
	TaskKey       string
	OwnerWorkflow string
	ActiveWorkflow string
}

func (e *errDocWrongWorkflow) Error() string {
	if e == nil {
		return "document belongs to another workflow"
	}
	owners := []string{}
	if e.OwnerWorkflow != "" {
		owners = []string{e.OwnerWorkflow}
	}
	doc := e.Title
	if doc == "" {
		doc = e.TaskKey
	}
	return intent.WorkflowScopeGuidanceQuestion(doc, owners, e.ActiveWorkflow)
}

func (e *errDocWrongWorkflow) AsCoordinatorErr() error {
	return &coordinator.ErrWrongWorkflow{Message: e.Error()}
}

// selectReadableWorkflowDocument 在可读文档池中解析目标。
// scopeTaskKeys 非空时先在激活流池内匹配；未命中但精确/归一化匹配到被剔除的他流文档时返回 ErrWrongWorkflow。
// scopeTaskKeys 为空（未激活 / planning）时保持现状：在全量池内匹配。
func selectReadableWorkflowDocument(
	docs []docgov.WorkflowDocumentSummary,
	target string,
	activeWorkflow string,
	scopeTaskKeys map[string]bool,
) (*docgov.WorkflowDocumentSummary, error) {
	if len(docs) == 0 {
		return nil, fmt.Errorf("no readable workflow documents")
	}
	targetKey := normalizeDocumentLookupKey(target)
	if targetKey == "" {
		inScope := filterDocsByTaskKeys(docs, scopeTaskKeys)
		pool := inScope
		if len(scopeTaskKeys) == 0 {
			pool = docs
		}
		if len(pool) == 1 {
			return &pool[0], nil
		}
		return nil, fmt.Errorf("document target required")
	}

	inScope := filterDocsByTaskKeys(docs, scopeTaskKeys)
	pool := inScope
	if len(scopeTaskKeys) == 0 {
		pool = docs
	}
	if hit := matchDocumentInPool(pool, target, targetKey); hit != nil {
		return hit, nil
	}
	// 激活流池未命中：检查是否精确/归一化命中被剔除的他流文档。
	if len(scopeTaskKeys) > 0 {
		outOfScope := filterDocsOutsideTaskKeys(docs, scopeTaskKeys)
		if hit := matchDocumentInPool(outOfScope, target, targetKey); hit != nil {
			owner := intent.TaskKeyWorkflowOwner(hit.TaskKey)
			return nil, (&errDocWrongWorkflow{
				Title:          hit.Title,
				TaskKey:        hit.TaskKey,
				OwnerWorkflow:  owner,
				ActiveWorkflow: activeWorkflow,
			}).AsCoordinatorErr()
		}
	}
	return nil, fmt.Errorf("document not found in readable scope")
}

func filterDocsByTaskKeys(docs []docgov.WorkflowDocumentSummary, taskKeys map[string]bool) []docgov.WorkflowDocumentSummary {
	if len(taskKeys) == 0 {
		return docs
	}
	out := make([]docgov.WorkflowDocumentSummary, 0, len(docs))
	for _, d := range docs {
		if projectinstance.TaskKeyInScope(taskKeys, d.TaskKey) {
			out = append(out, d)
		}
	}
	return out
}

func filterDocsOutsideTaskKeys(docs []docgov.WorkflowDocumentSummary, taskKeys map[string]bool) []docgov.WorkflowDocumentSummary {
	if len(taskKeys) == 0 {
		return nil
	}
	out := make([]docgov.WorkflowDocumentSummary, 0)
	for _, d := range docs {
		if !projectinstance.TaskKeyInScope(taskKeys, d.TaskKey) {
			out = append(out, d)
		}
	}
	return out
}

func matchDocumentInPool(docs []docgov.WorkflowDocumentSummary, target, targetKey string) *docgov.WorkflowDocumentSummary {
	for i := range docs {
		if target == docs[i].DocID || target == docs[i].TaskKey || target == docs[i].Title {
			return &docs[i]
		}
	}
	for i := range docs {
		keys := []string{
			normalizeDocumentLookupKey(docs[i].DocID),
			normalizeDocumentLookupKey(docs[i].TaskKey),
			normalizeDocumentLookupKey(docs[i].Title),
		}
		for _, key := range keys {
			if key != "" && (key == targetKey || strings.Contains(key, targetKey) || strings.Contains(targetKey, key)) {
				return &docs[i]
			}
		}
	}
	return nil
}

func normalizeDocumentLookupKey(value string) string {
	value = strings.ToLower(strings.TrimSpace(value))
	value = strings.NewReplacer("收集", "", "采集", "", "正文", "", "文档", "", "《", "", "》", "").Replace(value)
	return strings.Map(func(r rune) rune {
		if unicode.IsSpace(r) || unicode.IsPunct(r) {
			return -1
		}
		return r
	}, value)
}