代码理解

%pip install --upgrade --quiet langchain-openai tiktoken chromadb langchain git
 
# 设置环境变量 OPENAI_API_KEY 或从 .env 文件加载
# import dotenv
 
# dotenv.load_dotenv()

from git import Repo
from langchain_community.document_loaders.generic import GenericLoader
from langchain_community.document_loaders.parsers import LanguageParser
from langchain_text_splitters import Language

# Clone
repo_path = "/Users/jacoblee/Desktop/test_repo"
repo = Repo.clone_from("https://github.com/langchain-ai/langchain", to_path=repo_path)

# Load
loader = GenericLoader.from_filesystem(
    repo_path + "/libs/core/langchain_core",
    glob="**/*",
    suffixes=[".py"],
    exclude=["**/non-utf8-encoding.py"],
    parser=LanguageParser(language=Language.PYTHON, parser_threshold=500),
)
documents = loader.load()
len(documents)

295

from langchain_text_splitters import RecursiveCharacterTextSplitter
 
python_splitter = RecursiveCharacterTextSplitter.from_language(
    language=Language.PYTHON, chunk_size=2000, chunk_overlap=200
)
texts = python_splitter.split_documents(documents)
len(texts)

898

from langchain_community.vectorstores import Chroma
from langchain_openai import OpenAIEmbeddings
 
db = Chroma.from_documents(texts, OpenAIEmbeddings(disallowed_special=()))
retriever = db.as_retriever(
    search_type="mmr",  # 也可以尝试 "similarity"
    search_kwargs={"k": 8},
)

from langchain.chains import create_history_aware_retriever, create_retrieval_chain
from langchain.chains.combine_documents import create_stuff_documents_chain
from langchain_core.prompts import ChatPromptTemplate
from langchain_openai import ChatOpenAI
 
llm = ChatOpenAI(model="gpt-4")
 
# 首先，我们需要一个提示，可以将其传递到LLM中以生成此搜索查询
 
prompt = ChatPromptTemplate.from_messages(
    [
        ("placeholder", "{chat_history}"),
        ("user", "{input}"),
        (
            "user",
            "根据上面的对话，生成一个搜索查询，以获取与对话相关的信息",
        ),
    ]
)
 
retriever_chain = create_history_aware_retriever(llm, retriever, prompt)
 
prompt = ChatPromptTemplate.from_messages(
    [
        (
            "system",
            "根据以下上下文回答用户的问题：\n\n{context}",
        ),
        ("placeholder", "{chat_history}"),
        ("user", "{input}"),
    ]
)
document_chain = create_stuff_documents_chain(llm, prompt)
 
qa = create_retrieval_chain(retriever_chain, document_chain)

question = "什么是RunnableBinding？"
result = qa.invoke({"input": question})
result["answer"]

'RunnableBinding是LangChain库中用于将参数绑定到Runnable的类之一。当链中的可运行对象需要一个不在前一个可运行对象的输出中或用户输入中的参数时，这就很有用。它返回一个带有绑定参数和配置的新Runnable。RunnableBinding类中的bind方法用于执行此操作。'

questions = [
    "有哪些类是从Runnable类派生的？",
    "在与Runnable类相关的类层次结构的代码中，你提议的一种改进是什么？",
]
 
for question in questions:
    result = qa.invoke({"input": question})
    print(f"-> **问题**: {question} \n")
    print(f"**回答**: {result['answer']} \n")

-> **问题**: 有哪些类是从Runnable类派生的？ 

**回答**: 按照上下文中提到的，从`Runnable`类派生的类有：`RunnableLambda`，`RunnableLearnable`，`RunnableSerializable`，`RunnableWithFallbacks`。 

-> **问题**: 在与Runnable类相关的类层次结构的代码中，你提议的一种改进是什么？ 

**回答**: 一个潜在的改进可能是为不同类型的Runnable类引入抽象基类（ABC）或接口。目前，似乎有很多不同的Runnable类型，比如RunnableLambda，RunnableParallel等，每个类型都有自己的方法和属性。通过为所有这些类定义一个公共接口或ABC，我们可以确保一致性并更好地组织代码库。这也将使将来添加新类型的可运行类更容易，因为它们只需要实现接口或ABC中定义的方法即可。 

%pip install --upgrade --quiet langchain-community

from langchain_community.chat_models.ollama import ChatOllama
 
llm = ChatOllama(model="codellama")

response_message = llm.invoke(
    "In bash, how do I list all the text files in the current directory that have been modified in the last month?"
)
 
print(response_message.content)
print(response_message.response_metadata)

你可以使用`find`命令和`-mtime`选项来查找在当前目录中修改时间在最近一个月内的所有文本文件。下面是一个示例命令：
```bash
find . -type f -name "*.txt" -mtime -30

find . -type f -name "*.txt" -mmin -4320

ls -l | grep "*.txt"

看起来很合理！现在让我们使用之前加载的vectorstore进行设置。

我们省略对话方面，以使事情对于低功耗本地模型更易于管理：

```python
# from langchain.chains.question_answering import load_qa_chain

# # Prompt
# template = """Use the following pieces of context to answer the question at the end.
# If you don't know the answer, just say that you don't know, don't try to make up an answer.
# Use three sentences maximum and keep the answer as concise as possible.
# {context}
# Question: {question}
# Helpful Answer:"""
# QA_CHAIN_PROMPT = PromptTemplate(
#     input_variables=["context", "question"],
#     template=template,
# )

system_template = """
Answer the user's questions based on the below context.
If you don't know the answer, just say that you don't know, don't try to make up an answer. 
Use three sentences maximum and keep the answer as concise as possible:

{context}
"""

# First we need a prompt that we can pass into an LLM to generate this search query
prompt = ChatPromptTemplate.from_messages(
    [
        ("system", system_template),
        ("user", "{input}"),
    ]
)
document_chain = create_stuff_documents_chain(llm, prompt)

qa_chain = create_retrieval_chain(retriever, document_chain)

# Run, only returning the value under the answer key for readability
qa_chain.pick("answer").invoke({"input": "What is a RunnableBinding?"})