Cyrene_v2/plugins/channels/qq_obv11/message.go

package qq_obv11

// This file carries OneBot v11 protocol types and the conversions between the
// native event/segment model and the normalized adapter.UnifiedMessage /
// UnifiedResponse model. It understands both array-format messages
// ([]{"type","data"}) and the legacy CQ-code string format
// ([CQ:image,file=...]text), and maps QQ media (image/record/video/file) into
// adapter.Media and QQ mentions into adapter.Mention.

import (
	"encoding/json"
	"fmt"
	"strconv"
	"strings"
	"time"

	"git.yeij.top/AskaEth/Cyrene_v2/internal/channel/adapter"
)

const platformName = "qq-obv11"

// OneBot v11 event as delivered over WebSocket.
type obEvent struct {
	PostType    string          `json:"post_type"`
	MessageType string          `json:"message_type"`
	Time        int64           `json:"time"`
	SelfID      int64           `json:"self_id"`
	MessageID   int64           `json:"message_id"`
	UserID      int64           `json:"user_id"`
	GroupID     int64           `json:"group_id"`
	GroupName   string          `json:"group_name"`
	SubType     string          `json:"sub_type"`
	RawMessage  string          `json:"raw_message"`
	Message     json.RawMessage `json:"message"`
	Sender      *obSender       `json:"sender"`
}

// obSender is the sender identity carried by a message event.
type obSender struct {
	UserID   int64  `json:"user_id"`
	Nickname string `json:"nickname"`
	Card     string `json:"card"`
	Role     string `json:"role"`
	Sex      string `json:"sex"`
}

// obSegment is one element of an array-format message.
type obSegment struct {
	Type string         `json:"type"`
	Data map[string]any `json:"data"`
}

// obSendMsg is the outbound OneBot v11 API request (action + params).
type obSendMsg struct {
	Action string       `json:"action"`
	Params obSendParams `json:"params"`
	Echo   string       `json:"echo,omitempty"`
}

// obSendParams carries the send_msg parameters.
type obSendParams struct {
	MessageType string      `json:"message_type"`
	UserID      int64       `json:"user_id,omitempty"`
	GroupID     int64       `json:"group_id,omitempty"`
	Message     []obSegment `json:"message"`
	AutoEscape  bool        `json:"auto_escape"`
}

// parseEvent decodes a raw WebSocket text frame into a OneBot event. If the
// payload is not a message-style event the returned error is non-nil and the
// caller should attempt to fall back to another interpretation.
func parseEvent(payload []byte) (*obEvent, error) {
	var ev obEvent
	if err := json.Unmarshal(payload, &ev); err != nil {
		return nil, fmt.Errorf("qq-obv11: decode event: %w", err)
	}
	return &ev, nil
}

// toUnified converts a parsed message event into a normalized UnifiedMessage.
// raw preserves the original event payload for downstream tracing.
func (a *Adapter) toUnified(ev *obEvent, raw []byte) adapter.UnifiedMessage {
	text, media, mentions, replyID, replyText, hasText := a.parseMessage(ev)

	senderID := senderIDOf(ev)
	senderName := "unknown"
	role := adapter.RoleUnknown
	if ev.Sender != nil {
		senderName = firstNonEmpty(ev.Sender.Card, ev.Sender.Nickname, senderID)
		role = roleFromOneBot(ev.Sender.Role)
	}

	msg := adapter.UnifiedMessage{
		Platform:    platformName,
		SenderID:    senderID,
		SenderName:  senderName,
		Content:     text,
		ContentType: contentOf(media, hasText),
		Media:       media,
		Mentions:    mentions,
		ReplyToID:   replyID,
		ReplyToText: replyText,
		MessageID:   formatID(ev.MessageID),
		RawData:     raw,
		Timestamp:   time.Unix(ev.Time, 0),
		Sender: adapter.Actor{
			UserID:      senderID,
			DisplayName: senderName,
			Role:        role,
		},
	}

	switch ev.MessageType {
	case "group":
		// 群聊的渠道键是**群号**，发送者仍是 senderID（上一步解析）：两者绝不可互换。
		// 用群号当 SenderID 会让群里任何人都顶着群号去撞白名单（冒充主人）。事件里
		// 没有第二个群号来源，group_id 缺失时 ChannelID 就如实留空——下游
		// session.channelID() 会以 ErrInvalidChannel 拒掉这条消息（fail-closed），
		// 绝不退回用发送者号去编一个群会话键。
		gid := formatID(ev.GroupID)
		msg.ChannelType = adapter.ChannelGroup
		msg.ChannelID = gid
		msg.GroupID = gid
		name := firstNonEmpty(ev.GroupName, a.groupName(gid))
		msg.GroupName = name
		msg.Sender.GroupID = gid
		msg.Sender.IsGroup = true
		if name != "" {
			a.rememberGroup(gid, name)
		}
	case "private":
		msg.ChannelType = adapter.ChannelDirect
		// 私聊的对端**就是**发送者，所以这里的 ChannelID 可以安全地回落到 SenderID
		// （对端号 == 发送者号）。群聊绝不允许这样回落：群号不是发送者，用 ChannelID
		// 充当 SenderID 会让**任何群成员**都顶着群号去撞白名单——见 group 分支。
		msg.ChannelID = senderID
	}
	return msg
}

// parseMessage extracts the normalized parts of an event's message payload,
// handling both array-format segments and the CQ-code string format. hasText
// reports whether the message carries any real text (as opposed to only media /
// placeholder markers), which drives ContentType selection.
func (a *Adapter) parseMessage(ev *obEvent) (text string, media []adapter.Media, mentions []adapter.Mention, replyID, replyText string, hasText bool) {
	raw := ev.Message
	if len(raw) == 0 {
		raw = []byte(ev.RawMessage)
	}
	if len(raw) > 0 {
		switch raw[0] {
		case '[':
			return parseSegmentArray(raw)
		case '"':
			var s string
			if err := json.Unmarshal(raw, &s); err == nil {
				return parseCQMessage(s)
			}
		}
	}
	return parseCQMessage(ev.RawMessage)
}

// parseSegmentArray parses an array-format message.
func parseSegmentArray(raw []byte) (string, []adapter.Media, []adapter.Mention, string, string, bool) {
	var segs []obSegment
	_ = json.Unmarshal(raw, &segs)

	var b strings.Builder
	var media []adapter.Media
	var mentions []adapter.Mention
	var replyID, replyText string
	hasText := false

	for _, s := range segs {
		switch s.Type {
		case "text":
			t := segStr(s.Data, "text")
			b.WriteString(t)
			if t != "" {
				hasText = true
			}
		case "at":
			qq := segStr(s.Data, "qq")
			if qq != "" {
				mentions = append(mentions, adapter.Mention{UserID: qq})
				b.WriteString("@" + qq)
			}
		case "image":
			media = append(media, mediaFromSeg(s.Data, adapter.MediaImage))
			b.WriteString("[图片]")
		case "record":
			media = append(media, mediaFromSeg(s.Data, adapter.MediaVoice))
			b.WriteString("[语音]")
		case "video":
			media = append(media, mediaFromSeg(s.Data, adapter.MediaVideo))
			b.WriteString("[视频]")
		case "file":
			media = append(media, mediaFromSeg(s.Data, adapter.MediaFile))
			b.WriteString("[文件]")
		case "face":
			b.WriteString("[表情]")
		case "reply":
			if id := segStr(s.Data, "id"); id != "" {
				replyID = id
			}
			if t := segStr(s.Data, "text"); t != "" {
				replyText = t
			}
			b.WriteString("[回复]")
		case "json", "card":
			b.WriteString("[卡片消息]")
		default:
			b.WriteString("[未知消息]")
		}
	}
	return b.String(), media, mentions, replyID, replyText, hasText
}

// parseCQMessage parses a CQ-code string, yielding text, media, mentions, and a
// hasText flag indicating real (non-placeholder) content. It scans for
// [CQ:type,...] codes manually (no package-level regexp, per the no-global-var
// constraint).
func parseCQMessage(s string) (string, []adapter.Media, []adapter.Mention, string, string, bool) {
	var b strings.Builder
	var media []adapter.Media
	var mentions []adapter.Mention
	var replyID, replyText string
	hasText := false

	last := 0
	for i := 0; i < len(s); {
		if !strings.HasPrefix(s[i:], "[CQ:") {
			i++
			continue
		}
		end := strings.IndexByte(s[i:], ']')
		if end < 0 {
			break
		}
		fragment := s[last:i]
		b.WriteString(fragment)
		if fragment != "" {
			hasText = true
		}
		code := s[i : i+end+1]
		typ, params := decodeCQCode(code)
		switch typ {
		case "at":
			if qq := params["qq"]; qq != "" {
				mentions = append(mentions, adapter.Mention{UserID: qq})
				b.WriteString("@" + qq)
			}
		case "image":
			media = append(media, mediaFromParams(params, adapter.MediaImage))
			b.WriteString("[图片]")
		case "record":
			media = append(media, mediaFromParams(params, adapter.MediaVoice))
			b.WriteString("[语音]")
		case "video":
			media = append(media, mediaFromParams(params, adapter.MediaVideo))
			b.WriteString("[视频]")
		case "file":
			media = append(media, mediaFromParams(params, adapter.MediaFile))
			b.WriteString("[文件]")
		case "face":
			b.WriteString("[表情]")
		case "reply":
			if id := params["id"]; id != "" {
				replyID = id
			}
			replyText = params["text"]
			b.WriteString("[回复]")
		case "json", "card":
			b.WriteString("[卡片消息]")
		default:
			b.WriteString(code)
		}
		last = i + end + 1
		i = last
	}
	tail := s[last:]
	b.WriteString(tail)
	if tail != "" {
		hasText = true
	}
	return b.String(), media, mentions, replyID, replyText, hasText
}

func decodeCQCode(code string) (string, map[string]string) {
	inner := strings.TrimPrefix(code, "[CQ:")
	inner = strings.TrimSuffix(inner, "]")
	parts := strings.SplitN(inner, ",", 2)
	typ := parts[0]
	params := map[string]string{}
	if len(parts) == 2 && parts[1] != "" {
		for _, kv := range strings.Split(parts[1], ",") {
			kvParts := strings.SplitN(kv, "=", 2)
			if len(kvParts) != 2 {
				continue
			}
			params[kvParts[0]] = unescapeCQ(kvParts[1])
		}
	}
	return typ, params
}

func unescapeCQ(s string) string {
	s = strings.ReplaceAll(s, "&#44;", ",")
	s = strings.ReplaceAll(s, "&#91;", "[")
	s = strings.ReplaceAll(s, "&#93;", "]")
	s = strings.ReplaceAll(s, "&amp;", "&")
	s = strings.ReplaceAll(s, "&#40;", "(")
	s = strings.ReplaceAll(s, "&#41;", ")")
	return s
}

// mediaFromSeg builds a Media from an array-format segment's data map.
func mediaFromSeg(d map[string]any, kind adapter.MediaKind) adapter.Media {
	url := segStr(d, "url")
	file := segStr(d, "file")
	if url == "" {
		url = file
	}
	return adapter.Media{Kind: kind, URL: url, Name: file}
}

// mediaFromParams builds a Media from parsed CQ params.
func mediaFromParams(p map[string]string, kind adapter.MediaKind) adapter.Media {
	url := p["url"]
	file := p["file"]
	if url == "" {
		url = file
	}
	return adapter.Media{Kind: kind, URL: url, Name: file}
}

// outboundSegments converts a unified message into OneBot message segments.
func outboundSegments(msg adapter.UnifiedMessage) ([]obSegment, error) {
	var segs []obSegment
	if msg.Content != "" {
		segs = append(segs, obSegment{Type: "text", Data: map[string]any{"text": msg.Content}})
	}
	for _, m := range msg.Media {
		file := firstNonEmpty(m.URL, m.Name)
		switch m.Kind {
		case adapter.MediaImage:
			segs = append(segs, obSegment{Type: "image", Data: map[string]any{"file": file}})
		case adapter.MediaVoice:
			segs = append(segs, obSegment{Type: "record", Data: map[string]any{"file": file}})
		case adapter.MediaVideo:
			segs = append(segs, obSegment{Type: "video", Data: map[string]any{"file": file}})
		case adapter.MediaFile:
			segs = append(segs, obSegment{Type: "file", Data: map[string]any{"file": file}})
		}
	}
	if len(segs) == 0 {
		if msg.Content == "" {
			return nil, fmt.Errorf("qq-obv11: message has no content or media: %w", adapter.ErrUnsupported)
		}
		segs = append(segs, obSegment{Type: "text", Data: map[string]any{"text": msg.Content}})
	}
	return segs, nil
}

// resolveTarget determines where a reply should be delivered.
func resolveTarget(msg adapter.UnifiedMessage) (messageType string, userID, groupID int64, err error) {
	if msg.ChannelType == adapter.ChannelGroup {
		gid := firstNonEmpty(msg.GroupID, msg.ChannelID, msg.Sender.GroupID)
		if gid == "" {
			return "", 0, 0, fmt.Errorf("qq-obv11: group reply missing target: %w", adapter.ErrUnsupported)
		}
		g, convErr := parseInt64(gid)
		if convErr != nil {
			return "", 0, 0, fmt.Errorf("qq-obv11: group target %q: %w", gid, convErr)
		}
		return "group", 0, g, nil
	}
	uid := firstNonEmpty(msg.ChannelID, msg.SenderID, msg.Sender.UserID)
	if uid == "" {
		return "", 0, 0, fmt.Errorf("qq-obv11: private reply missing target: %w", adapter.ErrUnsupported)
	}
	u, convErr := parseInt64(uid)
	if convErr != nil {
		return "", 0, 0, fmt.Errorf("qq-obv11: user target %q: %w", uid, convErr)
	}
	return "private", u, 0, nil
}

func buildSendReq(messageType string, userID, groupID int64, segs []obSegment, echo string) ([]byte, error) {
	req := obSendMsg{
		Action: "send_msg",
		Params: obSendParams{
			MessageType: messageType,
			UserID:      userID,
			GroupID:     groupID,
			Message:     segs,
			AutoEscape:  false,
		},
		Echo: echo,
	}
	b, err := json.Marshal(req)
	if err != nil {
		return nil, fmt.Errorf("qq-obv11: encode send_msg: %w", err)
	}
	return b, nil
}

// contentOf picks the primary content type. When the message has no real text
// (only media / placeholder markers), it reflects the first media kind;
// otherwise text is the default.
func contentOf(media []adapter.Media, hasText bool) adapter.ContentType {
	if !hasText && len(media) > 0 {
		switch media[0].Kind {
		case adapter.MediaImage:
			return adapter.ContentImage
		case adapter.MediaVoice:
			return adapter.ContentVoice
		default:
			return adapter.ContentFile
		}
	}
	return adapter.ContentText
}

// roleFromOneBot 把 OneBot 上报的 sender.role 映射到 actor 的 Role 枚举。
// **群主（owner）与群管理员（admin）一律映射成 RoleMember**（裁决 06.4.2）：平台字段
// 只描述群内身份，不是主人车道凭据——主人只认白名单（owner.yaml + 每渠道 adminIds[]），
// 所以群主/管理员既绕不过限流与黑白名单，也进不了主会话；未知角色值保守归入 RoleUnknown。
func roleFromOneBot(raw string) adapter.Role {
	switch raw {
	case "owner", "admin", "member":
		return adapter.RoleMember
	default:
		return adapter.RoleUnknown
	}
}

func segStr(d map[string]any, key string) string {
	if d == nil {
		return ""
	}
	switch v := d[key].(type) {
	case string:
		return v
	case float64:
		return strconv.FormatInt(int64(v), 10)
	case json.Number:
		return v.String()
	default:
		return ""
	}
}