xinyin025/LangBot
1
0
Fork 0
mirror of https://github.com/langbot-app/LangBot.git synced 2026-06-13 09:16:04 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat: 重构图片消息传递逻辑 (#957, #955)

Browse Source
This commit is contained in:
Junyan Qin
2024-12-24 10:57:17 +08:00
parent 535c4a8a11
commit 12cfce3622
9 changed files with 70 additions and 54 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
pkg/provider/entities.py
View File

@@ -38,6 +38,8 @@ class ContentElement(pydantic.BaseModel):
image_url: typing.Optional[ImageURLContentObject] = None
image_base64: typing.Optional[str] = None
def __str__(self):
if self.type == 'text':
return self.text
@@ -53,6 +55,10 @@ class ContentElement(pydantic.BaseModel):
@classmethod
def from_image_url(cls, image_url: str):
return cls(type='image_url', image_url=ImageURLContentObject(url=image_url))
@classmethod
def from_image_base64(cls, image_base64: str):
return cls(type='image_base64', image_base64=image_base64)
class Message(pydantic.BaseModel):

1
pkg/provider/modelmgr/requester.py
View File

@@ -48,6 +48,7 @@ class LLMAPIRequester(metaclass=abc.ABCMeta):
@abc.abstractmethod
async def call(
self,
query: core_entities.Query,
model: modelmgr_entities.LLMModelInfo,
messages: typing.List[llm_entities.Message],
funcs: typing.List[tools_entities.LLMFunction] = None,

18
pkg/provider/modelmgr/requesters/anthropicmsgs.py
View File

@@ -2,6 +2,7 @@ from __future__ import annotations
import typing
import traceback
import base64
import anthropic
import httpx
@@ -39,6 +40,7 @@ class AnthropicMessages(requester.LLMAPIRequester):
async def call(
self,
query: core_entities.Query,
model: entities.LLMModelInfo,
messages: typing.List[llm_entities.Message],
funcs: typing.List[tools_entities.LLMFunction] = None,
@@ -70,28 +72,26 @@ class AnthropicMessages(requester.LLMAPIRequester):
if isinstance(m.content, str) and m.content.strip() != "":
req_messages.append(m.dict(exclude_none=True))
elif isinstance(m.content, list):
# m.content = [
# c for c in m.content if c.type == "text"
# ]
# if len(m.content) > 0:
# req_messages.append(m.dict(exclude_none=True))
msg_dict = m.dict(exclude_none=True)
for i, ce in enumerate(m.content):
if ce.type == "image_url":
base64_image, image_format = await image.qq_image_url_to_base64(ce.image_url.url)
if ce.type == "image_base64":
image_b64, image_format = await image.extract_b64_and_format(ce.image_base64)
alter_image_ele = {
"type": "image",
"source": {
"type": "base64",
"media_type": f"image/{image_format}",
"data": base64_image
"data": image_b64
}
}
msg_dict["content"][i] = alter_image_ele
print(msg_dict)
req_messages.append(msg_dict)
args["messages"] = req_messages

20
pkg/provider/modelmgr/requesters/chatcmpl.py
View File

@@ -65,6 +65,7 @@ class OpenAIChatCompletions(requester.LLMAPIRequester):
async def _closure(
self,
query: core_entities.Query,
req_messages: list[dict],
use_model: entities.LLMModelInfo,
use_funcs: list[tools_entities.LLMFunction] = None,
@@ -87,8 +88,12 @@ class OpenAIChatCompletions(requester.LLMAPIRequester):
for msg in messages:
if 'content' in msg and isinstance(msg["content"], list):
for me in msg["content"]:
if me["type"] == "image_url":
me["image_url"]['url'] = await self.get_base64_str(me["image_url"]['url'])
if me["type"] == "image_base64":
me["image_url"] = {
"url": me["image_base64"]
}
me["type"] = "image_url"
del me["image_base64"]
args["messages"] = messages
@@ -102,6 +107,7 @@ class OpenAIChatCompletions(requester.LLMAPIRequester):
async def call(
self,
query: core_entities.Query,
model: entities.LLMModelInfo,
messages: typing.List[llm_entities.Message],
funcs: typing.List[tools_entities.LLMFunction] = None,
@@ -118,7 +124,7 @@ class OpenAIChatCompletions(requester.LLMAPIRequester):
req_messages.append(msg_dict)
try:
return await self._closure(req_messages, model, funcs)
return await self._closure(query, req_messages, model, funcs)
except asyncio.TimeoutError:
raise errors.RequesterError('请求超时')
except openai.BadRequestError as e:
@@ -134,11 +140,3 @@ class OpenAIChatCompletions(requester.LLMAPIRequester):
raise errors.RequesterError(f'请求过于频繁或余额不足: {e.message}')
except openai.APIError as e:
raise errors.RequesterError(f'请求错误: {e.message}')
@async_lru.alru_cache(maxsize=128)
async def get_base64_str(
self,
original_url: str,
) -> str:
base64_image, image_format = await image.qq_image_url_to_base64(original_url)
return f"data:image/{image_format};base64,{base64_image}"

22
pkg/provider/modelmgr/requesters/ollamachat.py
View File

@@ -6,6 +6,7 @@ import typing
from typing import Union, Mapping, Any, AsyncIterator
import uuid
import json
import base64
import async_lru
import ollama
@@ -13,7 +14,7 @@ import ollama
from .. import entities, errors, requester
from ... import entities as llm_entities
from ...tools import entities as tools_entities
from ....core import app
from ....core import app, entities as core_entities
from ....utils import image
REQUESTER_NAME: str = "ollama-chat"
@@ -43,7 +44,7 @@ class OllamaChatCompletions(requester.LLMAPIRequester):
**args
)
async def _closure(self, req_messages: list[dict], use_model: entities.LLMModelInfo,
async def _closure(self, query: core_entities.Query, req_messages: list[dict], use_model: entities.LLMModelInfo,
user_funcs: list[tools_entities.LLMFunction] = None) -> (
llm_entities.Message):
args: Any = self.request_cfg['args'].copy()
@@ -57,9 +58,9 @@ class OllamaChatCompletions(requester.LLMAPIRequester):
for me in msg["content"]:
if me["type"] == "text":
text_content.append(me["text"])
elif me["type"] == "image_url":
image_url = await self.get_base64_str(me["image_url"]['url'])
image_urls.append(image_url)
elif me["type"] == "image_base64":
image_urls.append(me["image_base64"])
msg["content"] = "\n".join(text_content)
msg["images"] = [url.split(',')[1] for url in image_urls]
if 'tool_calls' in msg: # LangBot 内部以 str 存储 tool_calls 的参数，这里需要转换为 dict
@@ -109,6 +110,7 @@ class OllamaChatCompletions(requester.LLMAPIRequester):
async def call(
self,
query: core_entities.Query,
model: entities.LLMModelInfo,
messages: typing.List[llm_entities.Message],
funcs: typing.List[tools_entities.LLMFunction] = None,
@@ -122,14 +124,6 @@ class OllamaChatCompletions(requester.LLMAPIRequester):
msg_dict["content"] = "\n".join(part["text"] for part in content)
req_messages.append(msg_dict)
try:
return await self._closure(req_messages, model, funcs)
return await self._closure(query, req_messages, model, funcs)
except asyncio.TimeoutError:
raise errors.RequesterError('请求超时')
@async_lru.alru_cache(maxsize=128)
async def get_base64_str(
self,
original_url: str,
) -> str:
base64_image, image_format = await image.qq_image_url_to_base64(original_url)
return f"data:image/{image_format};base64,{base64_image}"

4
pkg/provider/runners/localagent.py
View File

@@ -23,7 +23,7 @@ class LocalAgentRunner(runner.RequestRunner):
req_messages = query.prompt.messages.copy() + query.messages.copy() + [query.user_message]
# 首次请求
msg = await query.use_model.requester.call(query.use_model, req_messages, query.use_funcs)
msg = await query.use_model.requester.call(query, query.use_model, req_messages, query.use_funcs)
yield msg
@@ -61,7 +61,7 @@ class LocalAgentRunner(runner.RequestRunner):
req_messages.append(err_msg)
# 处理完所有调用，再次请求
msg = await query.use_model.requester.call(query.use_model, req_messages, query.use_funcs)
msg = await query.use_model.requester.call(query, query.use_model, req_messages, query.use_funcs)
yield msg