Undefined-bot 3.15.0__tar.gz → 3.16.0__tar.gz

{undefined_bot-3.15.0 → undefined_bot-3.16.0}/CHANGELOG.md

@@ -1,3 +1,18 @@
+## v3.16.0 Bilibili 图文（opus）提取
+
+本版本为 B 站内容自动提取补齐图文（opus / 动态）类型，命中图文链接后发送一次合并转发：第一条是图文元数据、第二条起是正文，文字与图片按原顺序混排，正文超长自动分段；正文里的图文与视频卡片各自成为嵌套转发节点；新增的 `bilibili_opus` 工具还支持按字数区间或关键词读取正文。
+
+- 新增 Bilibili 图文自动提取管线。命中 `bilibili.com/opus/<id>`、`m.bilibili.com/opus/<id>`、`t.bilibili.com/<id>`、`b23.tv` 短链，以及 QQ 小程序与 news 分享卡片中的跳转链接后自动发送合并转发；与视频管线互不干扰，同一条消息同时含 BV 号与图文链接时两条管线各自发送。
+- 转发结构与视频一致：第一条节点是图文信息（封面、标题、UP主、时间、阅读 / 点赞 / 评论 / 转发、原文链接），第二条起是正文。正文按单节点 4000 字切分（节点名形如 `正文 1/3`），不截断、不丢内容，图片作为消息段保留在原文位置并可与文本共处同一节点；正文为空时以占位文案说明。
+- 正文里的卡片各自展开为独立节点：图文卡片再拉取一次递归展开，视频卡片真实下载视频文件并复用清晰度、时长、体积与超限策略限制且附带弹幕，链接、商品、直播、投票等其它卡片渲染为单个链接卡片节点。展开受 `opus_nested_depth`（默认 5 层）与 `opus_nested_max_cards`（默认 8 张）约束，超出边界降级为一行「标题 — 链接」；嵌套失败只降级该节点，不影响整条转发。
+- 图文详情兼容两个接口与两种响应结构：优先 `x/polymer/web-dynamic/v1/opus/detail`，风控（`-352`）、旧版专栏或动态不可见时回退 `x/polymer/web-dynamic/v1/detail`（带 WBI 签名）。`modules` 为列表或字典（含 `major.opus.summary` 摘要）都能解析，段落覆盖文本、图片、分割线、块引用、列表、代码与链接卡片，富文本节点覆盖 @、表情、公式、话题与 BV / CV / 网页链接；指向自身的图文卡片不再被当作嵌套卡片，避免自引用递归。
+- 新增 `bilibili_opus` 工具，`output_mode` 与 `bilibili_video` 对齐并额外提供 text 模式：`send` 发送合并转发，`uid` 把正文图片逐张登记为 `<attachment uid="pic_xxx"/>`（默认最多 9 张，单张失败不中断并逐条报告原因），`info` 只返回元信息，`text` 按字数读取正文。该工具已暴露给 `file_analysis_agent`，图文图片可直接做多模态分析。
+- `output_mode=text` 支持起始字、结束字与关键词：默认返回前 1000 字（`limit` 上限 20000），`keyword` 返回命中区间（最多 5 处）与前后各 60 字上下文并自动合并重叠片段。返回头给出正文总字数与本次范围，未读完时提示下一段用 `start=N` 继续；关键词未命中、正文为空、`start` 超出范围都有明确文案，非法参数返回可读错误而不是静默截断。
+- 新增 `[bilibili].opus_enabled`（默认开启，仍需 `auto_extract_enabled`）、`opus_max_items`（单条消息最多处理几篇图文）、`opus_nested_depth`、`opus_nested_max_cards`；群聊与私聊白名单复用视频侧的 `auto_extract_group_ids` / `auto_extract_private_ids`，旧 `config.toml` 不做改动也可直接运行。
+- 图文转发与视频转发同样写入历史：实时 AI 上下文只看到 `forward_...` UID，需要内容时按层调用 `messages.get_forward_msg` 读取，不会因为一张长图文把上下文撑爆。视频与图文的数字缩写、时长、时间戳与 4000 字切分逻辑共用同一份实现，视频原有的提取行为不变。
+
+---
+
 ## v3.15.0 按功能拆分 Embedding 配置与稳定性修复
 
 本版本将嵌入模型配置按功能拆分，知识库、认知记忆与表情包可以共用默认模型或各自覆写连接、维度与指令；同时修复侧写并发合并丢失更新、Agent 技能无法加载、热更新静默失效、SIGTERM 停机缺失等多处影响长期运行的问题。

{undefined_bot-3.15.0 → undefined_bot-3.16.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: Undefined-bot
-Version: 3.15.0
+Version: 3.16.0
 Summary: QQ bot platform with cognitive memory architecture and multi-agent Skills, via OneBot V11.
 Project-URL: Repository, https://github.com/69gg/Undefined
 Project-URL: Issues, https://github.com/69gg/Undefined/issues

{undefined_bot-3.15.0 → undefined_bot-3.16.0}/config.toml.example

@@ -1473,6 +1473,18 @@ auto_extract_group_ids = []
 # zh: 自动提取功能的私聊白名单（空=跟随全局 access.allowed_private_ids）。
 # en: Private chat allowlist for auto-extraction (empty = follow global access.allowed_private_ids).
 auto_extract_private_ids = []
+# zh: 是否在自动提取中处理 B 站图文（opus / 动态）链接。仍需 auto_extract_enabled 打开。
+# en: Whether Bilibili opus/dynamic links are handled during auto-extraction. auto_extract_enabled must also be on.
+opus_enabled = true
+# zh: 单条消息最多自动处理几篇图文。<=0 回退 3，>10 截断到 10。
+# en: Max number of opus posts to auto-process from one message. <=0 falls back to 3, >10 is capped to 10.
+opus_max_items = 3
+# zh: 图文正文里的图文/视频卡片最多再展开几层（每层都是嵌套合并转发）。0=不展开，只给链接；上限 5。
+# en: Max nesting depth for opus/video cards inside a post body (each level is a nested forward). 0=link only; capped at 5.
+opus_nested_depth = 5
+# zh: 单篇图文最多展开多少张嵌套卡片，超出部分降级为一行文本+链接。上限 20。
+# en: Max nested cards expanded per post; extras degrade to a single text line with the link. Capped at 20.
+opus_nested_max_cards = 8
 
 # zh: 抖音视频自动提取配置。
 # en: Douyin video auto-extraction settings.

{undefined_bot-3.15.0 → undefined_bot-3.16.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "Undefined-bot"
-version = "3.15.0"
+version = "3.16.0"
 description = "QQ bot platform with cognitive memory architecture and multi-agent Skills, via OneBot V11."
 readme = "README.md"
 authors = [
@@ -103,7 +103,7 @@ exclude = [
 python_version = "3.12"
 strict = true
 ignore_missing_imports = true
-exclude = ["^code/", "^apps/.*/src-tauri/target/"]
+exclude = ["^code/", "^apps/.*/src-tauri/target/", "^data/cache/uv/"]
 
 [tool.ruff]
 exclude = ["code"]

{undefined_bot-3.15.0 → undefined_bot-3.16.0}/src/Undefined/__init__.py

@@ -24,7 +24,7 @@ if TYPE_CHECKING:
     from .skills.registry import BaseRegistry as BaseRegistry
     from .skills.tools import ToolRegistry as ToolRegistry
 
-__version__: str = "3.15.0"
+__version__: str = "3.16.0"
 
 # symbol -> (module_path, attribute_name)；首次访问时才 importlib 加载
 _LAZY_IMPORTS: dict[str, tuple[str, str]] = {

undefined_bot-3.16.0/src/Undefined/bilibili/__init__.py

@@ -0,0 +1,22 @@
+"""Bilibili 视频提取模块
+
+提供 B 站视频标识符解析、视频下载和发送功能，以及图文（opus）解析。
+"""
+
+from Undefined.bilibili.opus_parser import (
+    extract_opus_from_json_message,
+    extract_opus_ids_with_shortlinks,
+)
+from Undefined.bilibili.parser import (
+    extract_bilibili_ids,
+    extract_from_json_message,
+    normalize_to_bvid,
+)
+
+__all__ = [
+    "extract_bilibili_ids",
+    "extract_from_json_message",
+    "extract_opus_from_json_message",
+    "extract_opus_ids_with_shortlinks",
+    "normalize_to_bvid",
+]

{undefined_bot-3.15.0 → undefined_bot-3.16.0}/src/Undefined/bilibili/api_client.py

@@ -2,11 +2,12 @@
 
 from __future__ import annotations
 
+import logging
 from typing import Any
 
 import httpx
 
-from Undefined.bilibili.errors import ApiResponseError
+from Undefined.bilibili.errors import ApiResponseError, OpusUnavailableError
 from Undefined.bilibili.models import VideoInfo, VideoStats
 from Undefined.bilibili.wbi import build_signed_params_sync, parse_cookie_string
 from Undefined.skills.http_config import build_httpx_client_kwargs
@@ -15,6 +16,39 @@ _BILIBILI_API_VIEW = "https://api.bilibili.com/x/web-interface/view"
 _BILIBILI_API_VIEW_WBI = "https://api.bilibili.com/x/web-interface/wbi/view"
 _BILIBILI_API_PLAYURL = "https://api.bilibili.com/x/player/playurl"
 _BILIBILI_API_PLAYURL_WBI = "https://api.bilibili.com/x/player/wbi/playurl"
+_BILIBILI_API_OPUS_DETAIL = (
+    "https://api.bilibili.com/x/polymer/web-dynamic/v1/opus/detail"
+)
+_BILIBILI_API_DYNAMIC_DETAIL = (
+    "https://api.bilibili.com/x/polymer/web-dynamic/v1/detail"
+)
+
+# 图文详情接口的 features 参数；htmlNewStyle 用于让旧版专栏返回新版正文结构
+OPUS_DETAIL_FEATURES: tuple[str, ...] = (
+    "onlyfansVote",
+    "onlyfansAssetsV2",
+    "decorationCard",
+    "htmlNewStyle",
+    "ugcDelete",
+    "editable",
+    "opusPrivateVisible",
+    "tribeeEdit",
+    "avatarAutoTheme",
+    "avatarTypeOpus",
+)
+
+# 通用动态详情接口的 features 参数；itemOpusStyle 让动态按图文风格返回
+DYNAMIC_DETAIL_FEATURES: tuple[str, ...] = (
+    "itemOpusStyle",
+    "opusBigCover",
+    "onlyfansVote",
+    "endFooterHidden",
+    "decorationCard",
+    "onlyfansAssetsV2",
+    "ugcDelete",
+    "onlyfansQaCard",
+    "commentsNewVersion",
+)
 
 DEFAULT_HEADERS: dict[str, str] = {
     "User-Agent": (
@@ -25,6 +59,8 @@ DEFAULT_HEADERS: dict[str, str] = {
     "Referer": "https://www.bilibili.com",
 }
 
+logger = logging.getLogger(__name__)
+
 
 def _api_message(data: dict[str, Any]) -> str:
     return str(data.get("message") or data.get("msg") or "未知错误")
@@ -186,3 +222,76 @@ class BilibiliApiClient:
         if not isinstance(data, dict):
             raise ApiResponseError("播放流响应缺少 data")
         return data
+
+    def get_opus_item(self, dynamic_id: str) -> dict[str, Any]:
+        """获取图文 / 动态的 ``data.item`` 原始结构。
+
+        先请求 ``opus/detail``；失败（含 ``-352`` 风控、旧版专栏要求
+        ``htmlNewStyle``、以及图文详情接口未覆盖的动态）时回退到通用
+        ``web-dynamic/v1/detail``（带 WBI 签名）。
+        """
+        opus_id = str(dynamic_id).strip()
+        if not opus_id:
+            raise OpusUnavailableError("图文 ID 为空")
+
+        primary_error: Exception | None = None
+        try:
+            payload = self._request_json(
+                _BILIBILI_API_OPUS_DETAIL,
+                {
+                    "id": opus_id,
+                    "features": ",".join(OPUS_DETAIL_FEATURES),
+                    "timezone_offset": -480,
+                },
+            )
+            if int(payload.get("code", -1)) == 0:
+                return self._extract_opus_item(payload, opus_id)
+            primary_error = OpusUnavailableError(
+                f"获取图文信息失败: {_api_message(payload)}"
+            )
+        except OpusUnavailableError as exc:
+            primary_error = exc
+        except Exception as exc:  # 网络 / JSON 异常都走回退
+            primary_error = exc
+            logger.debug("[Bilibili] opus/detail 请求异常，尝试通用动态接口: %s", exc)
+
+        fallback_error: Exception | None = None
+        try:
+            payload = self.request_with_wbi_fallback(
+                endpoint=_BILIBILI_API_DYNAMIC_DETAIL,
+                params={
+                    "id": opus_id,
+                    "features": ",".join(DYNAMIC_DETAIL_FEATURES),
+                    "timezone_offset": -480,
+                },
+            )
+            if int(payload.get("code", -1)) == 0:
+                return self._extract_opus_item(payload, opus_id)
+            fallback_error = OpusUnavailableError(
+                f"获取动态信息失败: {_api_message(payload)}"
+            )
+        except OpusUnavailableError as exc:
+            fallback_error = exc
+        except Exception as exc:
+            fallback_error = exc
+
+        raise OpusUnavailableError(
+            f"图文 {opus_id} 不可用: {fallback_error or primary_error or '未知错误'}"
+        )
+
+    @staticmethod
+    def _extract_opus_item(payload: dict[str, Any], opus_id: str) -> dict[str, Any]:
+        """校验响应并返回 ``data.item``。"""
+        data = payload.get("data")
+        if not isinstance(data, dict):
+            raise OpusUnavailableError(f"图文 {opus_id} 响应缺少 data")
+        # data.fallback 非空表示该动态无法以图文形态返回（例如过旧的内容）
+        fallback = data.get("fallback")
+        item = data.get("item")
+        if not isinstance(item, dict) or not item:
+            detail = f"（fallback={fallback!r}）" if fallback else ""
+            raise OpusUnavailableError(f"图文 {opus_id} 没有可用内容{detail}")
+        modules = item.get("modules")
+        if not isinstance(modules, (list, dict)) or not modules:
+            raise OpusUnavailableError(f"图文 {opus_id} 响应缺少 modules")
+        return item

{undefined_bot-3.15.0 → undefined_bot-3.16.0}/src/Undefined/bilibili/downloader.py

@@ -6,6 +6,7 @@ import asyncio
 from functools import partial
 import logging
 from pathlib import Path
+from typing import Any
 import uuid
 
 from Undefined.bilibili.api_client import BilibiliApiClient
@@ -23,6 +24,7 @@ __all__ = [
     "VideoInfo",
     "cleanup_file",
     "download_video",
+    "get_opus_item",
     "get_video_info",
 ]
 
@@ -39,6 +41,21 @@ async def get_video_info(
     return await asyncio.to_thread(partial(_get_video_info_sync, bvid, cookie=cookie))
 
 
+async def get_opus_item(
+    dynamic_id: str,
+    cookie: str = "",
+) -> dict[str, Any]:
+    """获取图文 / 动态的原始 ``data.item`` 结构。"""
+    return await asyncio.to_thread(
+        partial(_get_opus_item_sync, dynamic_id, cookie=cookie)
+    )
+
+
+def _get_opus_item_sync(dynamic_id: str, *, cookie: str = "") -> dict[str, Any]:
+    with BilibiliApiClient(cookie=cookie, timeout=_DEFAULT_TIMEOUT_SECONDS) as client:
+        return client.get_opus_item(dynamic_id)
+
+
 def _get_video_info_sync(bvid: str, *, cookie: str = "") -> VideoInfo:
     with BilibiliApiClient(cookie=cookie, timeout=_DEFAULT_TIMEOUT_SECONDS) as client:
         return client.get_video_info(bvid)

{undefined_bot-3.15.0 → undefined_bot-3.16.0}/src/Undefined/bilibili/errors.py

@@ -11,6 +11,10 @@ class ApiResponseError(BilibiliError):
     """B 站 API 返回失败或格式异常。"""
 
 
+class OpusUnavailableError(ApiResponseError):
+    """图文不存在、已删除，或响应缺少可用内容。"""
+
+
 class DownloadError(BilibiliError):
     """视频流下载或合并失败。"""
 

undefined_bot-3.16.0/src/Undefined/bilibili/format.py

@@ -0,0 +1,76 @@
+"""B 站内容展示用的共享格式化助手。
+
+这里只放纯函数：数字缩写、时长、时间戳、以及按长度切分文本。发送层
+（``bilibili.sender`` / ``bilibili.opus_sender``）与 ``utils/sender.py``
+共用同一套语义，避免各写一份。
+"""
+
+from __future__ import annotations
+
+from datetime import datetime, timedelta, timezone
+
+# QQ / OneBot 单条文本的保守上限；与 ``utils.sender.MAX_MESSAGE_LENGTH`` 保持一致。
+# 这里单独定义一份，避免 ``bilibili`` 包反向依赖发送器实现。
+MAX_TEXT_LENGTH = 4000
+
+# B 站展示时间统一使用北京时间（API 通过 timezone_offset=-480 对齐）
+BEIJING_TIMEZONE = timezone(timedelta(hours=8))
+
+
+def format_count(value: int) -> str:
+    """把计数格式化为中文缩写（1.2万 / 1.5亿）。"""
+    if value < 0:
+        value = 0
+    if value >= 100_000_000:
+        return f"{value / 100_000_000:.1f}亿"
+    if value >= 10_000:
+        return f"{value / 10_000:.1f}万"
+    return str(value)
+
+
+def format_duration(seconds: int) -> str:
+    """把秒数格式化为 mm:ss 或 h:mm:ss。"""
+    seconds = max(0, int(seconds))
+    hours, remainder = divmod(seconds, 3600)
+    minutes, secs = divmod(remainder, 60)
+    if hours:
+        return f"{hours}:{minutes:02d}:{secs:02d}"
+    return f"{minutes}:{secs:02d}"
+
+
+def format_progress(progress_ms: int) -> str:
+    """把弹幕时间戳（毫秒）格式化为 mm:ss。"""
+    return format_duration(max(0, int(progress_ms)) // 1000)
+
+
+def format_timestamp(timestamp: int) -> str:
+    """把 UNIX 秒级时间戳格式化为北京时间 ``YYYY-MM-DD HH:MM``。"""
+    if timestamp <= 0:
+        return ""
+    try:
+        moment = datetime.fromtimestamp(int(timestamp), tz=BEIJING_TIMEZONE)
+    except (OverflowError, OSError, ValueError):
+        return ""
+    return moment.strftime("%Y-%m-%d %H:%M")
+
+
+def split_text_chunks(text: str, limit: int = MAX_TEXT_LENGTH) -> list[str]:
+    """按长度分片，优先在换行后切分并保留原始文本。
+
+    与 ``utils.sender._split_text_chunks`` 同语义：任何字符都不会丢失，
+    无换行的超长段落会被硬切。
+    """
+    if not text:
+        return []
+    size = max(1, int(limit))
+    chunks: list[str] = []
+    start = 0
+    while start < len(text):
+        end = min(start + size, len(text))
+        if end < len(text):
+            newline = text.rfind("\n", start, end)
+            if newline >= start:
+                end = newline + 1
+        chunks.append(text[start:end])
+        start = end
+    return chunks

undefined_bot-3.16.0/src/Undefined/bilibili/models.py

@@ -0,0 +1,163 @@
+"""B 站下载数据模型。"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+from pathlib import Path
+
+
+@dataclass(slots=True, frozen=True)
+class VideoStats:
+    """视频互动统计。"""
+
+    view: int = 0
+    danmaku: int = 0
+    reply: int = 0
+    favorite: int = 0
+    coin: int = 0
+    share: int = 0
+    like: int = 0
+
+
+@dataclass(slots=True, frozen=True)
+class VideoInfo:
+    """视频基本信息。"""
+
+    bvid: str
+    aid: int
+    title: str
+    duration: int
+    cover_url: str
+    up_name: str
+    desc: str
+    cid: int
+    page_duration: int
+    stats: VideoStats
+
+    @property
+    def url(self) -> str:
+        """标准视频链接。"""
+        return f"https://www.bilibili.com/video/{self.bvid}"
+
+
+@dataclass(slots=True, frozen=True)
+class DownloadResult:
+    """视频下载结果。"""
+
+    path: Path
+    size_bytes: int
+    quality: int
+    quality_label: str
+    video_info: VideoInfo
+
+
+@dataclass(slots=True, frozen=True)
+class DanmakuItem:
+    """单条弹幕。"""
+
+    progress_ms: int
+    content: str
+    dmid: str = ""
+    mode: int = 0
+    pool: int = 0
+    ctime: int = 0
+    mid_hash: str = ""
+    color: int = 0
+    weight: int = 0
+
+
+@dataclass(slots=True, frozen=True)
+class OpusStats:
+    """图文互动统计。"""
+
+    view: int = 0
+    like: int = 0
+    comment: int = 0
+    repost: int = 0
+    coin: int = 0
+    favorite: int = 0
+
+
+@dataclass(slots=True, frozen=True)
+class OpusAuthor:
+    """图文作者信息。"""
+
+    mid: int = 0
+    name: str = ""
+    avatar_url: str = ""
+
+
+@dataclass(slots=True, frozen=True)
+class TextBlock:
+    """图文段落中的纯文本。"""
+
+    text: str
+
+
+@dataclass(slots=True, frozen=True)
+class ImageBlock:
+    """图文段落中的图片组。"""
+
+    urls: tuple[str, ...]
+
+
+@dataclass(slots=True, frozen=True)
+class VideoCardBlock:
+    """指向投稿视频的卡片段落。"""
+
+    bvid: str = ""
+    title: str = ""
+    cover_url: str = ""
+    jump_url: str = ""
+
+
+@dataclass(slots=True, frozen=True)
+class OpusCardBlock:
+    """指向另一篇图文的卡片段落。"""
+
+    opus_id: str = ""
+    title: str = ""
+    cover_url: str = ""
+    jump_url: str = ""
+
+
+@dataclass(slots=True, frozen=True)
+class LinkCardBlock:
+    """其它类型的卡片段落（商品 / 直播 / 投票 / 通用链接等）。"""
+
+    title: str = ""
+    jump_url: str = ""
+    cover_url: str = ""
+
+
+OpusBlock = TextBlock | ImageBlock | VideoCardBlock | OpusCardBlock | LinkCardBlock
+
+
+@dataclass(slots=True, frozen=True)
+class OpusInfo:
+    """图文（opus / 动态）基本信息。"""
+
+    opus_id: str
+    title: str
+    blocks: tuple[OpusBlock, ...]
+    author: OpusAuthor = OpusAuthor()
+    stats: OpusStats = OpusStats()
+    pub_ts: int = 0
+    cover_url: str = ""
+    dynamic_type_id: str = ""
+    is_forward: bool = False
+    forward_origin: OpusAuthor | None = None
+
+    @property
+    def url(self) -> str:
+        """标准图文链接。"""
+        return f"https://www.bilibili.com/opus/{self.opus_id}"
+
+    @property
+    def images(self) -> tuple[str, ...]:
+        """按顺序收集全部图片 URL。"""
+        urls: list[str] = []
+        for block in self.blocks:
+            if isinstance(block, ImageBlock):
+                urls.extend(block.urls)
+        return tuple(urls)