Python 和 Ollama 的问题

Question

你最近怎么样？

我在一台Mac M2上使用Python 3.11.7。我在一个虚拟环境（venv）中有一份依赖列表。

我在使用Ollama时遇到了一些问题。我在本地和使用Docker进行测试，但错误信息并没有帮助我找出问题所在。

我的想法是加载一个HTML文件，并能够在这个上下文中进行查询。

这是我的代码：

from langchain_community.llms import Ollama
from langchain_community.document_loaders import WebBaseLoader
from langchain_community.document_loaders import PyPDFLoader
from langchain_community.vectorstores import Chroma
from langchain_community import embeddings
from langchain_community.chat_models import ChatOllama
from langchain_core.runnables import RunnablePassthrough
from langchain_core.output_parsers import StrOutputParser
from langchain_core.prompts import ChatPromptTemplate
from langchain.output_parsers import PydanticOutputParser
from langchain.text_splitter import CharacterTextSplitter

model_local = Ollama(base_url="http://192.168.0.200:11434", model="mistral")

# 1. Split data into chunks
urls = [
    "https://es.wikipedia.org/wiki/The_A-Team",
]
docs = [WebBaseLoader(url).load() for url in urls]
docs_list = [item for sublist in docs for item in sublist]
text_splitter = CharacterTextSplitter.from_tiktoken_encoder(chunk_size=7500, chunk_overlap=100)
doc_splits = text_splitter.split_documents(docs_list)

# 2. Convert documents to Embeddings and store them
vectorstore = Chroma.from_documents(
    documents=doc_splits,
    collection_name="rag-chroma",
    embedding=embeddings.ollama.OllamaEmbeddings(model='nomic-embed-text'),
)
retriever = vectorstore.as_retriever()

# 3. Before RAG
print("Before RAG\n")
before_rag_template = "What is {topic}"
before_rag_prompt = ChatPromptTemplate.from_template(before_rag_template)
before_rag_chain = before_rag_prompt | model_local | StrOutputParser()
print(before_rag_chain.invoke({"topic": "Ollama"}))

# 4. After RAG
print("\n########\nAfter RAG\n")
after_rag_template = """Answer the question based only on the following context:
{context}
Question: {question}
"""
after_rag_prompt = ChatPromptTemplate.from_template(after_rag_template)
after_rag_chain = (
    {"context": retriever, "question": RunnablePassthrough()}
    | after_rag_prompt
    | model_local
    | StrOutputParser()
)
print(after_rag_chain.invoke("Quien integra Brigada A?"))

在右边我列出了模型，左边则是检查Ollama是否有回应。

错误

cd /Users/santiago/Proyects/OllamaURL ; /usr/bin/env /Users/santiago/Proyects/OllamaURL/env/bin/python /Users/santiago/.vscode/extensions/ms-python.debugpy-2024.2.0-darwin-arm64/bundled/libs/debugpy/adapter/../../debugpy/launcher 54637 -- /Users/santiago/Proyects/OllamaURL/rag.py
Traceback (most recent call last):
  File "/Users/santiago/Proyects/OllamaURL/rag.py", line 25, in <module>
    vectorstore = Chroma.from_documents(
                  ^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/santiago/Proyects/OllamaURL/env/lib/python3.11/site-packages/langchain_community/vectorstores/chroma.py", line 778, in from_documents
    return cls.from_texts(
           ^^^^^^^^^^^^^^^
  File "/Users/santiago/Proyects/OllamaURL/env/lib/python3.11/site-packages/langchain_community/vectorstores/chroma.py", line 736, in from_texts
    chroma_collection.add_texts(
  File "/Users/santiago/Proyects/OllamaURL/env/lib/python3.11/site-packages/langchain_community/vectorstores/chroma.py", line 275, in add_texts
    embeddings = self._embedding_function.embed_documents(texts)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/santiago/Proyects/OllamaURL/env/lib/python3.11/site-packages/langchain_community/embeddings/ollama.py", line 204, in embed_documents
    embeddings = self._embed(instruction_pairs)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/santiago/Proyects/OllamaURL/env/lib/python3.11/site-packages/langchain_community/embeddings/ollama.py", line 192, in _embed
    return [self._process_emb_response(prompt) for prompt in iter_]
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/santiago/Proyects/OllamaURL/env/lib/python3.11/site-packages/langchain_community/embeddings/ollama.py", line 192, in <listcomp>
    return [self._process_emb_response(prompt) for prompt in iter_]
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/santiago/Proyects/OllamaURL/env/lib/python3.11/site-packages/langchain_community/embeddings/ollama.py", line 166, in _process_emb_response
    raise ValueError(
ValueError: Error raised by inference API HTTP code: 404, {"error":"model 'nomic-embed-text' not found, try pulling it first"}

我希望你能指导我解决这个错误。

Answer 1

默认情况下，OllamaEmbeddings的base_url是http://localhost:11434。如果你想要把base_url设置成http://192.168.0.200:11434，可以这样做。

vectorstore = Chroma.from_documents(
    documents=doc_splits,
    collection_name="rag-chroma",
    embedding=embeddings.ollama.OllamaEmbeddings(
        base_url='http://192.168.0.200:11434',
        model='nomic-embed-text'
    ),
)

参考资料

1. OllamaEmbeddings (LangChain GitHub)

Answer 2

我对你的代码做了一些调整，并成功运行了。以下是代码

from langchain.text_splitter import CharacterTextSplitter
from langchain.schema.document import Document
text_splitter = CharacterTextSplitter(chunk_size=100, chunk_overlap=20)
text = "I am going to tell you a story about Tintin."
docs = [Document(page_content=x) for x in text_splitter.split_text(text)]


from langchain_community.vectorstores import Chroma
from langchain_community.llms import Ollama
from langchain_community import embeddings
persist_directory = "/tmp/chromadb"
vectorstore = Chroma.from_documents(
    documents=docs,
    collection_name="test",
    embedding=embeddings.ollama.OllamaEmbeddings(model='nomic-embed-text')
)
retriever = vectorstore.as_retriever()


from langchain_community.llms import Ollama
llm = Ollama(model="mistral")


from langchain_core.runnables import RunnablePassthrough
from langchain_core.output_parsers import StrOutputParser
from langchain_core.prompts import ChatPromptTemplate
template = """Answer the question based only on the following context:
{context}
Question: {question}
"""
prompt = ChatPromptTemplate.from_template(template)
rag_chain = (
    {"context": retriever, "question": RunnablePassthrough()}
    | prompt
    | llm
    | StrOutputParser()
)
print(rag_chain.invoke("tell me a story"))

我们可以看看你提供的错误信息

ValueError: Error raised by inference API HTTP code: 404, {"error":"model 'nomic-embed-text' not found, try pulling it first"}

你确定Ollama已经拉取了这样的嵌入模型吗？我通过运行以下命令拉取了它

ollama pull nomic-embed-text

Ollama嵌入参考 > https://python.langchain.com/docs/integrations/text_embedding/ollama