fix: Async callback for langchain

pyproject.toml

@@ -10,7 +10,7 @@ classifiers = [
     "Operating System :: OS Independent",
     "Programming Language :: Python :: 3",
 ]
-requires-python = ">=3.7.1"
+requires-python = ">=3.8.1"
 dependencies = [
     "miservice_fork",
     "openai",
@@ -22,11 +22,11 @@ dependencies = [
     "EdgeGPT==0.1.26",
     "langchain==0.0.301",
     "datetime==5.2",
-    "bs4==0.0.1",
+    "beautifulsoup4>=4.12.0",
     "chardet==5.1.0",
     "typing==3.7.4.3",
-    "google-search-results==2.4.2",
-    "numexpr==2.8.6"
+    "google-search-results>=2.4.2",
+    "numexpr==2.8.6",
 ]
 license = {text = "MIT"}
 dynamic = ["version"]

requirements.txt

@@ -1,8 +1,9 @@
 # This file is @generated by PDM.
 # Please do not edit it manually.
 
-aiohttp==3.8.5
+aiohttp==3.8.4
 aiosignal==1.3.1
+annotated-types==0.6.0
 anyio==3.6.2
 async-timeout==4.0.2
 attrs==22.2.0
@@ -11,10 +12,13 @@ beautifulsoup4==4.12.2
 BingImageCreator==0.1.3
 browser-cookie3==0.19.1
 cachetools==4.2.4
-certifi==2023.7.22
+certifi==2022.12.7
+chardet==5.1.0
 charset-normalizer==3.1.0
 colorama==0.4.6
 dataclasses==0.6
+dataclasses-json==0.6.1
+datetime==5.2
 deep-translator==1.11.4
 edge-tts==6.1.3
 EdgeGPT==0.1.26
@@ -23,6 +27,7 @@ google-api-core==1.34.0
 google-auth==1.35.0
 google-cloud-core==1.7.3
 google-cloud-translate==2.0.1
+google-search-results==2.4.2
 googleapis-common-protos==1.59.1
 grpcio==1.56.2
 grpcio-status==1.48.2
@@ -33,38 +38,49 @@ httpcore==0.16.3
 httpx==0.24.1
 hyperframe==6.0.1
 idna==3.4
+jsonpatch==1.33
+jsonpointer==2.4
+langchain==0.0.301
+langsmith==0.0.52
 lz4==4.3.2
 markdown-it-py==2.2.0
+marshmallow==3.20.1
 mdurl==0.1.2
 miservice-fork==2.1.1
 multidict==6.0.4
+mypy-extensions==1.0.0
+numexpr==2.8.6
+numpy==1.24.4
 openai==0.27.2
+packaging==23.2
 prompt-toolkit==3.0.38
 protobuf==3.20.3
 pyasn1==0.5.0
 pyasn1-modules==0.3.0
 pycryptodomex==3.18.0
-pygments==2.15.0
+pydantic==2.4.2
+pydantic-core==2.10.1
+pygments==2.14.0
 PyJWT==2.8.0
+pytz==2023.3.post1
+PyYAML==6.0.1
 regex==2022.10.31
-requests==2.31.0
+requests==2.28.2
 rich==13.3.2
 rsa==4.9
 setuptools==68.0.0
 six==1.16.0
 sniffio==1.3.0
 soupsieve==2.4.1
+SQLAlchemy==2.0.22
+tenacity==8.2.3
 tqdm==4.65.0
+typing==3.7.4.3
+typing-extensions==4.8.0
+typing-inspect==0.9.0
 urllib3==1.26.15
 wcwidth==0.2.6
 websockets==11.0
 yarl==1.8.2
 zhipuai==1.0.7
-langchain==0.0.301
-datetime==5.2
-bs4==0.0.1
-chardet==5.1.0
-typing==3.7.4.3
-google-search-results==2.4.2
-numexpr==2.8.6
-
+zope-interface==6.1

xiaogpt/bot/langchain_bot.py

@@ -1,15 +1,14 @@
 from __future__ import annotations
 
-import openai
+import asyncio
+import os
+
 from rich import print
 
 from xiaogpt.bot.base_bot import BaseBot
-from xiaogpt.utils import split_sentences
-
+from xiaogpt.langchain.callbacks import AsyncIteratorCallbackHandler
 from xiaogpt.langchain.chain import agent_search
-from xiaogpt.langchain.stream_call_back import streaming_call_queue
-
-import os
+from xiaogpt.utils import split_sentences
 
 
 class LangChainBot(BaseBot):
@@ -42,18 +41,17 @@ def from_config(cls, config):
     async def ask(self, query, **options):
         # Todo，Currently only supports stream
         raise Exception(
-            "The bot does not support it. Please use 'ask_stream，add： --stream'"
+            "The bot does not support it. Please use 'ask_stream, add: --stream'"
         )
 
     async def ask_stream(self, query, **options):
-        agent_search(query)
+        callback = AsyncIteratorCallbackHandler()
+        task = asyncio.create_task(agent_search(query, callback))
         try:
-            while True:
-                if not streaming_call_queue.empty():
-                    token = streaming_call_queue.get()
-                    print(token, end="")
-                    yield token
-                else:
-                    break
+            async for message in split_sentences(callback.aiter()):
+                yield message
         except Exception as e:
             print("An error occurred:", str(e))
+        finally:
+            print()
+            await task

xiaogpt/langchain/callbacks.py

@@ -0,0 +1,87 @@
+from __future__ import annotations
+
+import asyncio
+from typing import Any, AsyncIterator
+from uuid import UUID
+
+from langchain.callbacks.base import AsyncCallbackHandler
+
+
+class AsyncIteratorCallbackHandler(AsyncCallbackHandler):
+    """Callback handler that returns an async iterator."""
+
+    @property
+    def always_verbose(self) -> bool:
+        return True
+
+    def __init__(self) -> None:
+        self.queue = asyncio.Queue()
+        self.done = asyncio.Event()
+
+    async def on_chain_start(
+        self,
+        serialized: dict[str, Any],
+        inputs: dict[str, Any],
+        *,
+        run_id: UUID,
+        parent_run_id: UUID | None = None,
+        tags: list[str] | None = None,
+        metadata: dict[str, Any] | None = None,
+        **kwargs: Any,
+    ) -> None:
+        self.done.clear()
+
+    async def on_llm_new_token(self, token: str, **kwargs: Any) -> None:
+        if token is not None and token != "":
+            print(token, end="")
+            self.queue.put_nowait(token)
+
+    async def on_chain_end(
+        self,
+        outputs: dict[str, Any],
+        *,
+        run_id: UUID,
+        parent_run_id: UUID | None = None,
+        tags: list[str] | None = None,
+        **kwargs: Any,
+    ) -> None:
+        self.done.set()
+
+    async def on_chain_error(
+        self,
+        error: BaseException,
+        *,
+        run_id: UUID,
+        parent_run_id: UUID | None = None,
+        tags: list[str] | None = None,
+        **kwargs: Any,
+    ) -> None:
+        self.done.set()
+
+    async def aiter(self) -> AsyncIterator[str]:
+        while not self.queue.empty() or not self.done.is_set():
+            # Wait for the next token in the queue,
+            # but stop waiting if the done event is set
+            done, other = await asyncio.wait(
+                [
+                    # NOTE: If you add other tasks here, update the code below,
+                    # which assumes each set has exactly one task each
+                    asyncio.ensure_future(self.queue.get()),
+                    asyncio.ensure_future(self.done.wait()),
+                ],
+                return_when=asyncio.FIRST_COMPLETED,
+            )
+
+            # Cancel the other task
+            if other:
+                other.pop().cancel()
+
+            # Extract the value of the first completed task
+            token_or_done = done.pop().result()
+
+            # If the extracted value is the boolean True, the done event was set
+            if token_or_done is True:
+                break
+
+            # Otherwise, the extracted value is a token, which we yield
+            yield token_or_done

xiaogpt/langchain/chain.py

@@ -1,23 +1,15 @@
-from langchain.agents import initialize_agent, Tool
-from langchain.agents import AgentType
-from langchain.tools import BaseTool
-from langchain.llms import OpenAI
+from langchain.agents import AgentType, Tool, initialize_agent
+from langchain.callbacks.base import BaseCallbackHandler
 from langchain.chains import LLMMathChain
-from langchain.utilities import SerpAPIWrapper
 from langchain.chat_models import ChatOpenAI
-from langchain.memory import ChatMessageHistory
-from xiaogpt.langchain.stream_call_back import StreamCallbackHandler
-from langchain.agents.agent_toolkits import ZapierToolkit
-from langchain.utilities.zapier import ZapierNLAWrapper
-from langchain.memory import ConversationBufferMemory
+from langchain.utilities import SerpAPIWrapper
 
 
-def agent_search(query):
+async def agent_search(query: str, callback: BaseCallbackHandler) -> None:
     llm = ChatOpenAI(
         streaming=True,
         temperature=0,
         model="gpt-3.5-turbo-0613",
-        callbacks=[StreamCallbackHandler()],
     )
 
     # Initialization: search chain, mathematical calculation chain
@@ -35,4 +27,4 @@ def agent_search(query):
     )
 
     # query eg：'杭州亚运会中国队获得了多少枚金牌？' // '计算3的2次方'
-    agent.run(query)
+    await agent.arun(query, callbacks=[callback])

xiaogpt/utils.py

@@ -29,7 +29,7 @@ def calculate_tts_elapse(text: str) -> float:
     return len(_no_elapse_chars.sub("", text)) / speed
 
 
-_ending_punctuations = ("。", "？", "！", "；", ".", "?", "!", ";")
+_ending_punctuations = ("。", "？", "！", "；", "\n", "?", "!", ";")
 
 
 async def split_sentences(text_stream: AsyncIterator[str]) -> AsyncIterator[str]: