from langchain_chroma import Chroma
from langchain_community.document_loaders import TextLoader
from langchain_community.embeddings import SentenceTransformerEmbeddings
embedding = SentenceTransformerEmbeddings(model_name="all-MiniLM-L6-v2")
def build_vector_db():
loader = TextLoader("data/runbooks.txt")
docs = loader.load()
vectordb = Chroma.from_documents(
docs,
embedding,
persist_directory="./chroma_db"
)
vectordb.persist()
def search_docs(query):
vectordb = Chroma(
persist_directory="./chroma_db",
embedding_function=embedding
)
docs = vectordb.similarity_search(query, k=2)
return [d.page_content for d in docs]
