Kris
ef5483dc1a
主要变更： 1. 修复速率限流器使用setdefault替代重复创建令牌桶 2. 重构交互注册表匹配逻辑，优化精确匹配查找 3. 重构去重缓存逻辑，移到适配器实例方法 4. 重构发送URL解析，增加合法性校验并拆分公共方法 5. 优化流式消息处理逻辑，简化flush_controller调用 6. 重构群聊类型判断代码，简化语法 7. 修复重连管理器对None类型关闭分类的处理 8. 新增消息缓存、线程模拟器、发送初始化模块 9. 重构凭证备份与会话存储逻辑，支持环境变量指定状态目录 10. 新增配置提示与向导二维码绑定功能 11. 优化媒体上传逻辑，增加重试机制与缓存 12. 新增审批键盘模板构建函数 13. 重构消息格式处理，修正媒体发送字段与长度限制 14. 修复令牌过期时间计算，使用time.time替代monotonic 15. 新增群组激活缓冲区与用户追踪器增强功能 16. 修复换行符问题，统一文件结尾格式
262 lines
8.4 KiB
Python
262 lines
8.4 KiB
Python
from __future__ import annotations
|
|
|
|
from yuxi.channels.models import ChannelResponse
|
|
|
|
from .constants import DM_CHAT_PREFIX, GROUP_CHAT_PREFIX
|
|
|
|
TEXT_CONTENT_LIMIT = 2000
|
|
MARKDOWN_CONTENT_LIMIT = 4096
|
|
EMBED_DESCRIPTION_LIMIT = 4096
|
|
EMBED_PROMPT_LIMIT = 200
|
|
EMBED_TITLE_LIMIT = 200
|
|
|
|
|
|
def build_text_payload(response: ChannelResponse) -> dict:
|
|
content = response.content[:TEXT_CONTENT_LIMIT]
|
|
chat_id = response.identity.channel_chat_id
|
|
|
|
payload: dict = {"content": content}
|
|
|
|
payload["msg_type"] = 0
|
|
|
|
if chat_id.startswith(GROUP_CHAT_PREFIX):
|
|
payload["group_openid"] = chat_id.replace(GROUP_CHAT_PREFIX, "")
|
|
elif not chat_id.startswith(DM_CHAT_PREFIX):
|
|
payload["channel_id"] = chat_id
|
|
|
|
if response.reply_to_message_id:
|
|
payload["msg_id"] = response.reply_to_message_id
|
|
|
|
return payload
|
|
|
|
|
|
def build_markdown_payload(response: ChannelResponse, template_id: str | None = None) -> dict:
|
|
md_template_id = response.metadata.get("markdown_template_id") or template_id
|
|
|
|
if md_template_id:
|
|
return {
|
|
"msg_type": 2,
|
|
"markdown": {
|
|
"template_id": md_template_id,
|
|
"params": [
|
|
{"key": "title", "values": [response.metadata.get("title", "")]},
|
|
{"key": "content", "values": [response.content[:MARKDOWN_CONTENT_LIMIT]]},
|
|
],
|
|
},
|
|
}
|
|
|
|
return {
|
|
"msg_type": 2,
|
|
"markdown": {
|
|
"content": response.content[:MARKDOWN_CONTENT_LIMIT],
|
|
},
|
|
}
|
|
|
|
|
|
def build_ark_payload(response: ChannelResponse) -> dict:
|
|
ark_template_id = response.metadata.get("ark_template_id")
|
|
ark_data = response.metadata.get("ark_data", {})
|
|
|
|
return {
|
|
"msg_type": 3,
|
|
"ark": {
|
|
"template_id": ark_template_id,
|
|
"kv": [{"key": k, "value": v} for k, v in ark_data.items()],
|
|
},
|
|
}
|
|
|
|
|
|
def build_embed_payload(response: ChannelResponse) -> dict:
|
|
embed_data = response.metadata.get("embed", {})
|
|
|
|
return {
|
|
"msg_type": 4,
|
|
"embed": {
|
|
"title": embed_data.get("title", "")[:EMBED_TITLE_LIMIT],
|
|
"description": response.content[:EMBED_DESCRIPTION_LIMIT],
|
|
"prompt": embed_data.get("prompt", response.content[:EMBED_PROMPT_LIMIT]),
|
|
"fields": embed_data.get("fields", []),
|
|
},
|
|
}
|
|
|
|
|
|
def build_media_generic_payload(response: ChannelResponse, file_id: str, msg_type: int = 7) -> dict:
|
|
chat_id = response.identity.channel_chat_id
|
|
payload: dict = {
|
|
"msg_type": msg_type,
|
|
"media": {"file_info": file_id},
|
|
}
|
|
|
|
if response.content:
|
|
payload["content"] = response.content[:TEXT_CONTENT_LIMIT]
|
|
|
|
if chat_id.startswith(GROUP_CHAT_PREFIX):
|
|
payload["group_openid"] = chat_id.replace(GROUP_CHAT_PREFIX, "")
|
|
elif not chat_id.startswith(DM_CHAT_PREFIX):
|
|
payload["channel_id"] = chat_id
|
|
|
|
return payload
|
|
|
|
|
|
def build_image_payload(response: ChannelResponse, file_id: str) -> dict:
|
|
chat_id = response.identity.channel_chat_id
|
|
payload: dict = {
|
|
"msg_type": 1,
|
|
"file_image": file_id,
|
|
}
|
|
|
|
if response.content:
|
|
payload["content"] = response.content[:TEXT_CONTENT_LIMIT]
|
|
|
|
if chat_id.startswith(GROUP_CHAT_PREFIX):
|
|
payload["group_openid"] = chat_id.replace(GROUP_CHAT_PREFIX, "")
|
|
elif not chat_id.startswith(DM_CHAT_PREFIX):
|
|
payload["channel_id"] = chat_id
|
|
|
|
return payload
|
|
|
|
|
|
def format_outbound(
|
|
response: ChannelResponse,
|
|
use_markdown: bool = False,
|
|
markdown_template_id: str | None = None,
|
|
) -> dict:
|
|
msg_type = response.metadata.get("qq_msg_type", "")
|
|
chat_id = response.identity.channel_chat_id
|
|
|
|
if msg_type == "markdown" or (use_markdown and not msg_type):
|
|
return build_markdown_payload(response, markdown_template_id)
|
|
elif msg_type == "ark" and response.metadata.get("ark_template_id"):
|
|
return build_ark_payload(response)
|
|
elif msg_type == "embed":
|
|
if chat_id.startswith(DM_CHAT_PREFIX) or chat_id.startswith(GROUP_CHAT_PREFIX):
|
|
return build_markdown_payload(response, markdown_template_id)
|
|
return build_embed_payload(response)
|
|
elif msg_type == "image" and response.attachments:
|
|
file_id = response.attachments[0].file_id or response.attachments[0].url or ""
|
|
return build_image_payload(response, file_id)
|
|
elif msg_type in ("voice", "video", "file") and response.attachments:
|
|
file_id_list = [
|
|
response.metadata.get("media_file_id", ""),
|
|
response.attachments[0].file_id or "",
|
|
response.attachments[0].url or "",
|
|
]
|
|
file_id = next((fid for fid in file_id_list if fid), "")
|
|
return build_media_generic_payload(response, file_id, msg_type=7)
|
|
else:
|
|
return build_text_payload(response)
|
|
|
|
|
|
class MarkdownChunker:
|
|
MAX_CHARS = 4096
|
|
CHUNK_OVERLAP = 200
|
|
|
|
def __init__(self, max_chars: int = MAX_CHARS, chunk_overlap: int = CHUNK_OVERLAP):
|
|
self._max_chars = max_chars
|
|
self._chunk_overlap = chunk_overlap
|
|
|
|
def chunk(self, text: str) -> list[str]:
|
|
if len(text) <= self._max_chars:
|
|
return [text]
|
|
|
|
paragraphs = self._split_paragraphs(text)
|
|
chunks: list[str] = []
|
|
current_chunk: list[str] = []
|
|
current_len = 0
|
|
|
|
for para in paragraphs:
|
|
para_len = len(para)
|
|
if current_len + para_len <= self._max_chars:
|
|
current_chunk.append(para)
|
|
current_len += para_len
|
|
else:
|
|
if current_chunk:
|
|
chunks.append("".join(current_chunk))
|
|
if para_len > self._max_chars:
|
|
sub_chunks = self._force_split(para)
|
|
if current_chunk:
|
|
for i, sc in enumerate(sub_chunks):
|
|
chunks.append(sc)
|
|
else:
|
|
chunks.extend(sub_chunks)
|
|
current_chunk = []
|
|
current_len = 0
|
|
else:
|
|
current_chunk = [para]
|
|
current_len = para_len
|
|
|
|
if current_chunk:
|
|
chunks.append("".join(current_chunk))
|
|
|
|
return chunks
|
|
|
|
def _split_paragraphs(self, text: str) -> list[str]:
|
|
sections: list[str] = []
|
|
in_code_block = False
|
|
current: list[str] = []
|
|
lines = text.splitlines(keepends=True)
|
|
|
|
for line in lines:
|
|
stripped = line.strip()
|
|
if stripped.startswith("```"):
|
|
if current:
|
|
sections.append("".join(current))
|
|
current = []
|
|
if in_code_block:
|
|
sections.append(line)
|
|
in_code_block = False
|
|
else:
|
|
in_code_block = True
|
|
current.append(line)
|
|
continue
|
|
|
|
if in_code_block:
|
|
current.append(line)
|
|
if stripped.endswith("```"):
|
|
sections.append("".join(current))
|
|
current = []
|
|
in_code_block = False
|
|
continue
|
|
|
|
if not stripped:
|
|
if current:
|
|
sections.append("".join(current))
|
|
current = []
|
|
sections.append(line)
|
|
elif (
|
|
stripped.startswith(("#", "-", "*", ">", "|"))
|
|
and current
|
|
and not current[-1].strip().startswith(("#", "-", "*", ">", "|", "1.", "2.", "3."))
|
|
):
|
|
if current:
|
|
sections.append("".join(current))
|
|
current = []
|
|
current.append(line)
|
|
else:
|
|
current.append(line)
|
|
|
|
if current:
|
|
sections.append("".join(current))
|
|
|
|
result: list[str] = []
|
|
buffer: list[str] = []
|
|
for s in sections:
|
|
stripped = s.strip()
|
|
if not stripped and buffer:
|
|
result.append("".join(buffer))
|
|
buffer = []
|
|
buffer.append(s)
|
|
|
|
if buffer:
|
|
content = "".join(buffer)
|
|
if content.strip():
|
|
result.append(content)
|
|
|
|
return result or [text]
|
|
|
|
def _force_split(self, text: str) -> list[str]:
|
|
chunks: list[str] = []
|
|
for i in range(0, len(text), self._max_chars - self._chunk_overlap):
|
|
chunks.append(text[i : i + self._max_chars])
|
|
return chunks
|