chore(haystack-rag): remove entire RAG module
Some checks failed
Build and Push Docker / build-and-push (push) Failing after 2m44s
Some checks failed
Build and Push Docker / build-and-push (push) Failing after 2m44s
This commit is contained in:
gameloader
@ -1,5 +0,0 @@
|
|||||||
# haystack_rag module
|
|
||||||
from .main import run_chat_session
|
|
||||||
from .rag_pipeline import build_rag_pipeline
|
|
||||||
|
|
||||||
__all__ = ["run_chat_session", "build_rag_pipeline"]
|
|
||||||
@ -1,223 +0,0 @@
|
|||||||
# app.py
|
|
||||||
from fastapi import FastAPI, HTTPException, Depends
|
|
||||||
from pydantic import BaseModel
|
|
||||||
from typing import List, Optional
|
|
||||||
import logging
|
|
||||||
|
|
||||||
from haystack import Document
|
|
||||||
|
|
||||||
# Import necessary components from the provided code
|
|
||||||
from .data_handling import initialize_milvus_lite
|
|
||||||
from .main import initialize_document_embedder
|
|
||||||
from .retrieval import initialize_vector_retriever
|
|
||||||
from .embedding import initialize_text_embedder
|
|
||||||
|
|
||||||
# Setup logging
|
|
||||||
logging.basicConfig(level=logging.INFO)
|
|
||||||
logger = logging.getLogger(__name__)
|
|
||||||
|
|
||||||
# Initialize FastAPI app
|
|
||||||
app = FastAPI(title="Document Embedding and Retrieval API")
|
|
||||||
|
|
||||||
|
|
||||||
# Define request and response models
|
|
||||||
class EmbedRequest(BaseModel):
|
|
||||||
user_id: str
|
|
||||||
content: str
|
|
||||||
meta: Optional[dict] = {}
|
|
||||||
|
|
||||||
|
|
||||||
class RetrieveRequest(BaseModel):
|
|
||||||
user_id: str
|
|
||||||
query: str
|
|
||||||
|
|
||||||
|
|
||||||
class DocumentResponse(BaseModel):
|
|
||||||
content: str
|
|
||||||
score: Optional[float] = None
|
|
||||||
meta: Optional[dict] = {}
|
|
||||||
|
|
||||||
|
|
||||||
class RetrieveResponse(BaseModel):
|
|
||||||
documents: List[DocumentResponse]
|
|
||||||
query: str
|
|
||||||
answer: Optional[str] = None
|
|
||||||
|
|
||||||
|
|
||||||
# Helper functions
|
|
||||||
def get_document_embedder():
|
|
||||||
return initialize_document_embedder()
|
|
||||||
|
|
||||||
|
|
||||||
def get_document_store(user_id: str):
|
|
||||||
return initialize_milvus_lite(user_id)
|
|
||||||
|
|
||||||
|
|
||||||
@app.post("/embed", response_model=dict)
|
|
||||||
async def embed_document(
|
|
||||||
request: EmbedRequest, embedder=Depends(get_document_embedder)
|
|
||||||
):
|
|
||||||
"""
|
|
||||||
Embed content and store it in a Milvus collection for the specified user.
|
|
||||||
"""
|
|
||||||
try:
|
|
||||||
# Initialize document store for the user
|
|
||||||
document_store = get_document_store(request.user_id)
|
|
||||||
|
|
||||||
# Create a document with user content
|
|
||||||
meta = request.meta.copy()
|
|
||||||
meta["user_id"] = request.user_id # Ensure user_id is in meta
|
|
||||||
user_doc = Document(content=request.content, meta=meta)
|
|
||||||
|
|
||||||
# Embed the document
|
|
||||||
logger.info(f"Embedding document for user {request.user_id}")
|
|
||||||
embedding_result = embedder.run([user_doc])
|
|
||||||
embedded_docs = embedding_result.get("documents", [])
|
|
||||||
|
|
||||||
if not embedded_docs:
|
|
||||||
raise HTTPException(status_code=500, detail="Failed to embed document")
|
|
||||||
|
|
||||||
# Write to document store
|
|
||||||
logger.info(f"Writing embedded document to Milvus for user {request.user_id}")
|
|
||||||
document_store.write_documents(embedded_docs)
|
|
||||||
|
|
||||||
return {
|
|
||||||
"status": "success",
|
|
||||||
"message": f"Document embedded and stored for user {request.user_id}",
|
|
||||||
}
|
|
||||||
|
|
||||||
except Exception as e:
|
|
||||||
logger.error(f"Error embedding document: {str(e)}")
|
|
||||||
raise HTTPException(
|
|
||||||
status_code=500, detail=f"Error embedding document: {str(e)}"
|
|
||||||
)
|
|
||||||
|
|
||||||
|
|
||||||
@app.post("/retrieve", response_model=RetrieveResponse)
|
|
||||||
async def retrieve_documents(request: RetrieveRequest):
|
|
||||||
"""
|
|
||||||
Retrieve similar documents for a user based on a query without LLM generation.
|
|
||||||
Only retrieves documents using vector similarity.
|
|
||||||
"""
|
|
||||||
try:
|
|
||||||
# Get document store for the user
|
|
||||||
document_store = get_document_store(request.user_id)
|
|
||||||
|
|
||||||
# Initialize text embedder for query embedding
|
|
||||||
text_embedder = initialize_text_embedder()
|
|
||||||
|
|
||||||
# Initialize retriever
|
|
||||||
retriever = initialize_vector_retriever(document_store)
|
|
||||||
|
|
||||||
# Embed the query
|
|
||||||
logger.info(f"Embedding query for user {request.user_id}: '{request.query}'")
|
|
||||||
embedding_result = text_embedder.run(text=request.query)
|
|
||||||
query_embedding = embedding_result.get("embedding")
|
|
||||||
|
|
||||||
if not query_embedding:
|
|
||||||
raise HTTPException(status_code=500, detail="Failed to embed query")
|
|
||||||
|
|
||||||
# Retrieve similar documents
|
|
||||||
logger.info(f"Retrieving documents for query: '{request.query}'")
|
|
||||||
retriever_result = retriever.run(query_embedding=query_embedding)
|
|
||||||
retrieved_docs = retriever_result.get("documents", [])
|
|
||||||
|
|
||||||
# Convert to response format
|
|
||||||
documents = []
|
|
||||||
for doc in retrieved_docs:
|
|
||||||
documents.append(
|
|
||||||
DocumentResponse(
|
|
||||||
content=doc.content,
|
|
||||||
score=doc.score if hasattr(doc, "score") else None,
|
|
||||||
meta=doc.meta,
|
|
||||||
)
|
|
||||||
)
|
|
||||||
|
|
||||||
return RetrieveResponse(documents=documents, query=request.query, answer=None)
|
|
||||||
|
|
||||||
except Exception as e:
|
|
||||||
logger.error(f"Error retrieving documents: {str(e)}")
|
|
||||||
raise HTTPException(
|
|
||||||
status_code=500, detail=f"Error retrieving documents: {str(e)}"
|
|
||||||
)
|
|
||||||
|
|
||||||
|
|
||||||
# Add these imports at the top if not already there
|
|
||||||
from typing import List, Optional, Literal
|
|
||||||
|
|
||||||
|
|
||||||
# Define message models for the new endpoint
|
|
||||||
class Message(BaseModel):
|
|
||||||
content: str
|
|
||||||
role: str
|
|
||||||
name: Optional[str] = None
|
|
||||||
|
|
||||||
|
|
||||||
class EmbedMessagesRequest(BaseModel):
|
|
||||||
user_id: str
|
|
||||||
messages: List[Message]
|
|
||||||
meta: Optional[dict] = {}
|
|
||||||
|
|
||||||
|
|
||||||
@app.post("/embed_messages", response_model=dict)
|
|
||||||
async def embed_messages(
|
|
||||||
request: EmbedMessagesRequest, embedder=Depends(get_document_embedder)
|
|
||||||
):
|
|
||||||
"""
|
|
||||||
Process a messages array, extract content from user messages,
|
|
||||||
concatenate them with newlines, then embed and store in Milvus.
|
|
||||||
"""
|
|
||||||
try:
|
|
||||||
# Initialize document store for the user
|
|
||||||
document_store = get_document_store(request.user_id)
|
|
||||||
|
|
||||||
# Filter messages to keep only those with role="user"
|
|
||||||
user_messages = [msg for msg in request.messages if msg.role == "user"]
|
|
||||||
|
|
||||||
# Extract content from each user message
|
|
||||||
user_contents = [msg.content for msg in user_messages]
|
|
||||||
|
|
||||||
# Join contents with newline character
|
|
||||||
concatenated_content = "\n".join(user_contents)
|
|
||||||
|
|
||||||
if not concatenated_content.strip():
|
|
||||||
return {
|
|
||||||
"status": "warning",
|
|
||||||
"message": "No user messages found or all user messages were empty",
|
|
||||||
}
|
|
||||||
|
|
||||||
# Create a document with concatenated content
|
|
||||||
meta = request.meta.copy()
|
|
||||||
meta["user_id"] = request.user_id # Ensure user_id is in meta
|
|
||||||
user_doc = Document(content=concatenated_content, meta=meta)
|
|
||||||
|
|
||||||
# Embed the document
|
|
||||||
logger.info(f"Embedding concatenated user messages for user {request.user_id}")
|
|
||||||
embedding_result = embedder.run([user_doc])
|
|
||||||
embedded_docs = embedding_result.get("documents", [])
|
|
||||||
|
|
||||||
if not embedded_docs:
|
|
||||||
raise HTTPException(status_code=500, detail="Failed to embed document")
|
|
||||||
|
|
||||||
# Write to document store
|
|
||||||
logger.info(f"Writing embedded document to Milvus for user {request.user_id}")
|
|
||||||
document_store.write_documents(embedded_docs)
|
|
||||||
|
|
||||||
return {
|
|
||||||
"status": "success",
|
|
||||||
"message": f"User messages embedded and stored for user {request.user_id}",
|
|
||||||
"processed_messages_count": len(user_messages),
|
|
||||||
"concatenated_length": len(concatenated_content),
|
|
||||||
}
|
|
||||||
|
|
||||||
except Exception as e:
|
|
||||||
logger.error(f"Error embedding messages: {str(e)}")
|
|
||||||
raise HTTPException(
|
|
||||||
status_code=500, detail=f"Error embedding messages: {str(e)}"
|
|
||||||
)
|
|
||||||
|
|
||||||
|
|
||||||
if __name__ == "__main__":
|
|
||||||
import uvicorn
|
|
||||||
|
|
||||||
uvicorn.run(app, host="0.0.0.0", port=7999)
|
|
||||||
@ -1,96 +0,0 @@
|
|||||||
# data_handling.py
|
|
||||||
import os
|
|
||||||
from pathlib import Path
|
|
||||||
from typing import List, Optional
|
|
||||||
import logging # Added logging
|
|
||||||
|
|
||||||
from haystack import Document
|
|
||||||
from milvus_haystack import MilvusDocumentStore
|
|
||||||
|
|
||||||
# Import config variables needed
|
|
||||||
from config import (
|
|
||||||
OPENAI_EMBEDDING_DIM, # Keep for logging/validation if desired, but not passed to init
|
|
||||||
USER_ID_PREFIX,
|
|
||||||
MILVUS_PERSIST_BASE_DIR,
|
|
||||||
MILVUS_INDEX_PARAMS,
|
|
||||||
MILVUS_SEARCH_PARAMS,
|
|
||||||
MILVUS_STAND_URI,
|
|
||||||
)
|
|
||||||
|
|
||||||
logger = logging.getLogger(__name__) # Use logger
|
|
||||||
|
|
||||||
|
|
||||||
# get_user_milvus_path function remains the same
|
|
||||||
def get_user_milvus_path(user_id: str, base_dir: Path = MILVUS_PERSIST_BASE_DIR) -> str:
|
|
||||||
"""
|
|
||||||
获取指定用户的 Milvus Lite 数据库文件路径。
|
|
||||||
该函数会执行以下操作：
|
|
||||||
1. 基于- `base_dir` 和 `user_id` 构建一个用户专属的目录路径。
|
|
||||||
2. 确保该目录存在，如果不存在则会创建它。
|
|
||||||
3. 将目录路径与固定的数据库文件名 "milvus_lite.db" 组合。
|
|
||||||
4. 返回最终的完整文件路径（字符串格式）。
|
|
||||||
Args:
|
|
||||||
user_id (str): 用户的唯一标识符。
|
|
||||||
base_dir (Path, optional): Milvus 数据持久化的根目录.
|
|
||||||
默认为 MILVUS_PERSIST_BASE_DIR.
|
|
||||||
Returns:
|
|
||||||
str: 指向用户 Milvus 数据库文件的完整路径字符串。
|
|
||||||
"""
|
|
||||||
user_db_dir = base_dir / user_id
|
|
||||||
user_db_dir.mkdir(parents=True, exist_ok=True)
|
|
||||||
return str(user_db_dir / "milvus_lite.db")
|
|
||||||
|
|
||||||
|
|
||||||
def initialize_milvus_lite(user_id: str) -> MilvusDocumentStore:
|
|
||||||
"""
|
|
||||||
Initializes Milvus Lite DocumentStore for a user using milvus-haystack.
|
|
||||||
Dimension is inferred by Milvus upon first write, not passed here.
|
|
||||||
"""
|
|
||||||
print(f"Initializing Milvus Lite store for user: {user_id}")
|
|
||||||
milvus_uri = get_user_milvus_path(user_id)
|
|
||||||
print(f"Milvus Lite URI: {milvus_uri}")
|
|
||||||
# Log the dimension expected based on config, even if not passed directly
|
|
||||||
print(f"Expecting Embedding Dimension (for first write): {OPENAI_EMBEDDING_DIM}")
|
|
||||||
|
|
||||||
document_store = MilvusDocumentStore(
|
|
||||||
connection_args={"uri": milvus_uri},
|
|
||||||
collection_name=user_id, # Default or customize
|
|
||||||
index_params=MILVUS_INDEX_PARAMS, # Pass index config
|
|
||||||
search_params=MILVUS_SEARCH_PARAMS, # Pass search config
|
|
||||||
drop_old=False, # Keep drop_old for testing convenience
|
|
||||||
)
|
|
||||||
# Note: The actual schema dimension is set when the first document with an embedding is written.
|
|
||||||
print(f"Milvus Lite store instance created for user {user_id} at {milvus_uri}")
|
|
||||||
return document_store
|
|
||||||
|
|
||||||
|
|
||||||
# add_user_document_to_store and get_user_documents can remain if needed for other purposes,
|
|
||||||
def add_user_document_to_store(
|
|
||||||
document_store: MilvusDocumentStore, user_id: str, text: str
|
|
||||||
):
|
|
||||||
doc = Document(content=text, meta={"user_id": user_id})
|
|
||||||
print(f"Adding document for user {user_id}: '{text[:50]}...'")
|
|
||||||
document_store.write_documents([doc])
|
|
||||||
|
|
||||||
|
|
||||||
# get_user_documents function remains the same
|
|
||||||
def get_user_documents(
|
|
||||||
document_store: MilvusDocumentStore, user_id: str
|
|
||||||
) -> List[Document]:
|
|
||||||
print(f"Retrieving all documents for user {user_id}...")
|
|
||||||
all_docs = document_store.get_all_documents()
|
|
||||||
print(f"Found {len(all_docs)} documents for user {user_id}.")
|
|
||||||
return all_docs
|
|
||||||
|
|
||||||
|
|
||||||
# Optional: Test code similar to before, but now using the OpenAI dimension
|
|
||||||
if __name__ == "__main__":
|
|
||||||
test_user = "test_user_openai_data"
|
|
||||||
store = initialize_milvus_lite(test_user)
|
|
||||||
# Add dummy docs (won't be embedded here, just stored)
|
|
||||||
add_user_document_to_store(store, test_user, "第一个文档，关于 OpenAI。")
|
|
||||||
add_user_document_to_store(store, test_user, "第二个文档，使用 API。")
|
|
||||||
docs = get_user_documents(store, test_user)
|
|
||||||
for d in docs:
|
|
||||||
print(f" - {d.content} (Meta: {d.meta})")
|
|
||||||
# Cleanup code similar to before
|
|
||||||
@ -1,63 +0,0 @@
|
|||||||
# embedding.py
|
|
||||||
from haystack.components.embedders import OpenAITextEmbedder
|
|
||||||
|
|
||||||
from haystack.utils import Secret
|
|
||||||
|
|
||||||
# 从 config 导入新的变量名
|
|
||||||
from config import (
|
|
||||||
OPENAI_EMBEDDING_MODEL,
|
|
||||||
OPENAI_API_KEY_FROM_CONFIG, # 使用配置中的 Key
|
|
||||||
OPENAI_API_BASE_URL_CONFIG, # 使用配置中的 Base URL
|
|
||||||
OPENAI_EMBEDDING_KEY,
|
|
||||||
OPENAI_EMBEDDING_BASE,
|
|
||||||
HUGGINGFACE_KEY,
|
|
||||||
HUGGINGFACE_EMBEDDING_MODEL,
|
|
||||||
OPENAI_EMBEDDING_DIM
|
|
||||||
)
|
|
||||||
|
|
||||||
|
|
||||||
def initialize_text_embedder() -> OpenAITextEmbedder:
|
|
||||||
"""
|
|
||||||
Initializes the Haystack OpenAITextEmbedder component.
|
|
||||||
Reads API Key and Base URL directly from config.py.
|
|
||||||
"""
|
|
||||||
|
|
||||||
# 检查从配置加载的 key 是否有效 (基础检查)
|
|
||||||
if not OPENAI_API_KEY_FROM_CONFIG or "YOUR_API_KEY" in OPENAI_API_KEY_FROM_CONFIG:
|
|
||||||
print("警告: OpenAI API Key 未在 config.py 中有效配置。")
|
|
||||||
# Consider raising an error here if the key is mandatory
|
|
||||||
# raise ValueError("OpenAI API Key not configured correctly in config.py")
|
|
||||||
|
|
||||||
print(f"Initializing OpenAI Text Embedder with model: {OPENAI_EMBEDDING_MODEL}")
|
|
||||||
|
|
||||||
# 使用配置中的 Base URL
|
|
||||||
if OPENAI_API_BASE_URL_CONFIG:
|
|
||||||
print(f"Using custom API base URL from config: {OPENAI_API_BASE_URL_CONFIG}")
|
|
||||||
else:
|
|
||||||
print("Using default OpenAI API base URL (None specified in config).")
|
|
||||||
|
|
||||||
text_embedder = OpenAITextEmbedder(
|
|
||||||
# 直接使用从 config.py 导入的 key 和 base_url
|
|
||||||
api_key=Secret.from_token(OPENAI_EMBEDDING_KEY),
|
|
||||||
api_base_url=OPENAI_EMBEDDING_BASE,
|
|
||||||
model=OPENAI_EMBEDDING_MODEL,
|
|
||||||
dimensions=OPENAI_EMBEDDING_DIM,
|
|
||||||
)
|
|
||||||
print("Text Embedder initialized.")
|
|
||||||
return text_embedder
|
|
||||||
|
|
||||||
|
|
||||||
# __main__ 部分也需要调整以反映不依赖环境变量
|
|
||||||
# Example usage
|
|
||||||
if __name__ == "__main__":
|
|
||||||
embedder = initialize_text_embedder()
|
|
||||||
sample_text = "这是一个示例文本，用于测试嵌入功能。"
|
|
||||||
try:
|
|
||||||
result = embedder.run(text=sample_text)
|
|
||||||
embedding = result["embedding"]
|
|
||||||
print(f"Sample text: '{sample_text}'")
|
|
||||||
# print(f"Generated embedding (first 5 dims): {embedding[:5]}")
|
|
||||||
print(f"Generated embedding dimension: {len(embedding)}")
|
|
||||||
print(f"Tokens used: {result['meta']['usage']['total_tokens']}")
|
|
||||||
except Exception as e:
|
|
||||||
print(f"Error during huggingface API call: {e}")
|
|
||||||
@ -1,73 +0,0 @@
|
|||||||
# llm_integration.py
|
|
||||||
from haystack.components.generators.openai import OpenAIGenerator
|
|
||||||
from haystack.components.builders import PromptBuilder
|
|
||||||
from haystack.utils import Secret
|
|
||||||
|
|
||||||
# 从 config 导入新的变量名
|
|
||||||
from config import (
|
|
||||||
OPENAI_LLM_MODEL,
|
|
||||||
DEFAULT_PROMPT_TEMPLATE,
|
|
||||||
OPENAI_API_KEY_FROM_CONFIG, # 使用配置中的 Key
|
|
||||||
OPENAI_API_BASE_URL_CONFIG, # 使用配置中的 Base URL
|
|
||||||
)
|
|
||||||
|
|
||||||
|
|
||||||
def initialize_llm_and_prompt_builder() -> tuple[OpenAIGenerator, PromptBuilder]:
|
|
||||||
"""
|
|
||||||
Initializes the OpenAI Generator and PromptBuilder components.
|
|
||||||
Reads API Key and Base URL directly from config.py.
|
|
||||||
"""
|
|
||||||
|
|
||||||
if not OPENAI_API_KEY_FROM_CONFIG or "YOUR_API_KEY" in OPENAI_API_KEY_FROM_CONFIG:
|
|
||||||
print("警告: OpenAI API Key 未在 config.py 中有效配置。")
|
|
||||||
# Consider raising an error
|
|
||||||
# raise ValueError("OpenAI API Key not configured correctly in config.py")
|
|
||||||
|
|
||||||
print(f"Initializing OpenAI Generator with model: {OPENAI_LLM_MODEL}")
|
|
||||||
|
|
||||||
if OPENAI_API_BASE_URL_CONFIG:
|
|
||||||
print(f"Using custom API base URL from config: {OPENAI_API_BASE_URL_CONFIG}")
|
|
||||||
else:
|
|
||||||
print("Using default OpenAI API base URL (None specified in config).")
|
|
||||||
|
|
||||||
llm_generator = OpenAIGenerator(
|
|
||||||
# 直接使用从 config.py 导入的 key 和 base_url
|
|
||||||
api_key=Secret.from_token(OPENAI_API_KEY_FROM_CONFIG),
|
|
||||||
model=OPENAI_LLM_MODEL,
|
|
||||||
api_base_url=OPENAI_API_BASE_URL_CONFIG,
|
|
||||||
)
|
|
||||||
print("OpenAI Generator initialized.")
|
|
||||||
|
|
||||||
print("Initializing Prompt Builder...")
|
|
||||||
prompt_builder = PromptBuilder(template=DEFAULT_PROMPT_TEMPLATE)
|
|
||||||
print("Prompt Builder initialized.")
|
|
||||||
|
|
||||||
return llm_generator, prompt_builder
|
|
||||||
|
|
||||||
|
|
||||||
# __main__ 部分也需要调整
|
|
||||||
|
|
||||||
# Example Usage
|
|
||||||
if __name__ == "__main__":
|
|
||||||
from haystack import Document
|
|
||||||
|
|
||||||
llm, builder = initialize_llm_and_prompt_builder()
|
|
||||||
|
|
||||||
sample_question = "Haystack 是什么？"
|
|
||||||
sample_docs = [
|
|
||||||
Document(content="Haystack 是一个用于构建 NLP 应用程序的开源框架。"),
|
|
||||||
Document(content="你可以使用 Haystack 连接不同的组件。"),
|
|
||||||
]
|
|
||||||
|
|
||||||
prompt_builder_output = builder.run(question=sample_question, documents=sample_docs)
|
|
||||||
prompt = prompt_builder_output["prompt"]
|
|
||||||
print("\n--- Generated Prompt ---")
|
|
||||||
print(prompt)
|
|
||||||
|
|
||||||
print("\n--- Running OpenAI LLM ---")
|
|
||||||
try:
|
|
||||||
# Note: OpenAIGenerator expects 'prompt' as input key by default
|
|
||||||
llm_output = llm.run(prompt=prompt)
|
|
||||||
print("LLM Output:", llm_output)
|
|
||||||
except Exception as e:
|
|
||||||
print(f"Error during OpenAI API call: {e}")
|
|
||||||
@ -1,151 +0,0 @@
|
|||||||
# main.py
|
|
||||||
import sys
|
|
||||||
import logging
|
|
||||||
from haystack import Document
|
|
||||||
|
|
||||||
# 需要 OpenAIDocumentEmbedder 来嵌入要写入的文档
|
|
||||||
from haystack.components.embedders import OpenAIDocumentEmbedder
|
|
||||||
from haystack.utils import Secret
|
|
||||||
|
|
||||||
# 设置logger
|
|
||||||
logger = logging.getLogger(__name__)
|
|
||||||
|
|
||||||
# 导入所需的配置和构建函数
|
|
||||||
from config import (
|
|
||||||
DEFAULT_USER_ID,
|
|
||||||
OPENAI_API_KEY_FROM_CONFIG,
|
|
||||||
OPENAI_API_BASE_URL_CONFIG,
|
|
||||||
OPENAI_EMBEDDING_MODEL,
|
|
||||||
OPENAI_EMBEDDING_KEY,
|
|
||||||
OPENAI_EMBEDDING_BASE,
|
|
||||||
)
|
|
||||||
from .rag_pipeline import build_rag_pipeline # 构建 RAG 查询管道
|
|
||||||
|
|
||||||
|
|
||||||
# 辅助函数：初始化 Document Embedder (与 embedding.py 中的类似)
|
|
||||||
def initialize_document_embedder() -> OpenAIDocumentEmbedder:
|
|
||||||
"""初始化用于嵌入文档的 OpenAIDocumentEmbedder。"""
|
|
||||||
if not OPENAI_API_KEY_FROM_CONFIG or "YOUR_API_KEY" in OPENAI_API_KEY_FROM_CONFIG:
|
|
||||||
print("警告: OpenAI API Key 未在 config.py 中有效配置。")
|
|
||||||
# raise ValueError("OpenAI API Key not configured correctly in config.py")
|
|
||||||
|
|
||||||
print(f"Initializing OpenAI Document Embedder with model: {OPENAI_EMBEDDING_MODEL}")
|
|
||||||
if OPENAI_API_BASE_URL_CONFIG:
|
|
||||||
print(f"Using custom API base URL from config: {OPENAI_API_BASE_URL_CONFIG}")
|
|
||||||
else:
|
|
||||||
print("Using default OpenAI API base URL (None specified in config).")
|
|
||||||
|
|
||||||
document_embedder = OpenAIDocumentEmbedder(
|
|
||||||
api_key=Secret.from_token(OPENAI_EMBEDDING_KEY),
|
|
||||||
model=OPENAI_EMBEDDING_MODEL,
|
|
||||||
api_base_url=OPENAI_EMBEDDING_BASE,
|
|
||||||
# meta_fields_to_embed=["name"] # 如果需要嵌入元数据字段
|
|
||||||
# embedding_batch_size=10 # 可以调整批处理大小
|
|
||||||
)
|
|
||||||
print("OpenAI Document Embedder initialized.")
|
|
||||||
return document_embedder
|
|
||||||
|
|
||||||
|
|
||||||
def run_chat_session(user_id: str):
|
|
||||||
"""
|
|
||||||
运行 RAG 聊天会话主循环。
|
|
||||||
每次用户输入时，先将其嵌入并添加到 Milvus，然后运行 RAG 管道生成回复。
|
|
||||||
"""
|
|
||||||
print(f"--- Starting Chat Session for User: {user_id} ---")
|
|
||||||
|
|
||||||
# 构建 RAG 查询管道和获取 DocumentStore 实例
|
|
||||||
rag_query_pipeline, document_store = build_rag_pipeline(user_id)
|
|
||||||
|
|
||||||
# 初始化用于写入用户输入的 Document Embedder
|
|
||||||
document_embedder = initialize_document_embedder()
|
|
||||||
|
|
||||||
print("\nChatbot is ready! Type your questions or 'exit' to quit.")
|
|
||||||
# 打印使用的模型信息
|
|
||||||
try:
|
|
||||||
pass
|
|
||||||
# print(f"Using LLM: {rag_query_pipeline.get_component('generator').model}")
|
|
||||||
# 注意 RAG pipeline 中 query embedder 的名字是 'text_embedder'
|
|
||||||
# print(f"Using Query Embedder: {rag_query_pipeline.get_component('text_embedder').model}")
|
|
||||||
# print(f"Using Document Embedder (for writing): {document_embedder.model}")
|
|
||||||
except Exception as e:
|
|
||||||
print(f"Warning: Could not get component model names - {e}")
|
|
||||||
|
|
||||||
while True:
|
|
||||||
try:
|
|
||||||
query = input(f"[{user_id}] You: ")
|
|
||||||
if query.lower() == "exit":
|
|
||||||
print("Exiting chat session. Goodbye!")
|
|
||||||
break
|
|
||||||
if not query.strip():
|
|
||||||
continue
|
|
||||||
|
|
||||||
# --- 步骤 1: 嵌入用户输入并写入 Milvus ---
|
|
||||||
# print(f"[Workflow] Embedding user input as a document...")
|
|
||||||
# 将用户输入包装成 Haystack Document
|
|
||||||
user_doc_to_write = Document(content=query, meta={"user_id": user_id})
|
|
||||||
|
|
||||||
# 使用 OpenAIDocumentEmbedder 运行嵌入
|
|
||||||
# 它需要一个列表作为输入，即使只有一个文档
|
|
||||||
embedding_result = document_embedder.run([user_doc_to_write])
|
|
||||||
embedded_docs = embedding_result.get(
|
|
||||||
"documents", []
|
|
||||||
) # 获取带有嵌入的文档列表
|
|
||||||
|
|
||||||
if embedded_docs:
|
|
||||||
# print(f"[Workflow] Writing embedded document to Milvus for user {user_id}...")
|
|
||||||
# 将带有嵌入的文档写入 DocumentStore
|
|
||||||
document_store.write_documents(embedded_docs)
|
|
||||||
# print("[Workflow] Document written to Milvus.")
|
|
||||||
else:
|
|
||||||
print("[Workflow] Warning: Failed to embed document, skipping write.")
|
|
||||||
# 可以在这里添加错误处理或日志记录
|
|
||||||
|
|
||||||
# --- 步骤 2: 使用 RAG 查询管道生成回复 ---
|
|
||||||
# print("[Workflow] Running RAG query pipeline...")
|
|
||||||
# 准备 RAG 管道的输入
|
|
||||||
# text_embedder 需要原始查询文本
|
|
||||||
# prompt_builder 也需要原始查询文本（在模板中用作 {{query}}）
|
|
||||||
pipeline_input = {
|
|
||||||
"text_embedder": {"text": query},
|
|
||||||
"prompt_builder": {"query": query},
|
|
||||||
}
|
|
||||||
|
|
||||||
# 运行 RAG 查询管道
|
|
||||||
results = rag_query_pipeline.run(pipeline_input)
|
|
||||||
|
|
||||||
# --- 步骤 3: 处理并打印结果 ---
|
|
||||||
# 根据文档示例，生成器的输出在 'generator' 组件的 'replies' 键中
|
|
||||||
if "llm" in results and results["llm"]["replies"]:
|
|
||||||
answer = results["llm"]["replies"][0]
|
|
||||||
# 尝试获取 token 使用量（可能在 meta 中）
|
|
||||||
total_tokens = "N/A"
|
|
||||||
try:
|
|
||||||
# meta 结构可能因版本或配置而异，需要检查确认
|
|
||||||
if (
|
|
||||||
"meta" in results["llm"]
|
|
||||||
and isinstance(results["llm"]["meta"], list)
|
|
||||||
and results["llm"]["meta"]
|
|
||||||
):
|
|
||||||
usage_info = results["llm"]["meta"][0].get("usage", {})
|
|
||||||
total_tokens = usage_info.get("total_tokens", "N/A")
|
|
||||||
except Exception:
|
|
||||||
pass # 忽略获取 token 的错误
|
|
||||||
|
|
||||||
print(f"Chatbot: {answer} (Tokens: {total_tokens})")
|
|
||||||
else:
|
|
||||||
print("Chatbot: Sorry, I couldn't generate an answer for that.")
|
|
||||||
logger.debug(f"Pipeline Results: {results}") # 记录调试信息
|
|
||||||
|
|
||||||
except KeyboardInterrupt:
|
|
||||||
print("\nExiting chat session. Goodbye!")
|
|
||||||
break
|
|
||||||
except Exception as e:
|
|
||||||
print(f"\nAn error occurred: {e}")
|
|
||||||
import traceback
|
|
||||||
|
|
||||||
traceback.print_exc() # 打印详细的回溯信息
|
|
||||||
|
|
||||||
|
|
||||||
if __name__ == "__main__":
|
|
||||||
current_user_id = DEFAULT_USER_ID
|
|
||||||
run_chat_session(current_user_id)
|
|
||||||
@ -1,195 +0,0 @@
|
|||||||
# rag_pipeline.py
|
|
||||||
from haystack import Pipeline
|
|
||||||
from haystack import Document # 导入 Document
|
|
||||||
|
|
||||||
from milvus_haystack import MilvusDocumentStore
|
|
||||||
from .data_handling import initialize_milvus_lite
|
|
||||||
from .embedding import initialize_text_embedder
|
|
||||||
from .retrieval import initialize_vector_retriever
|
|
||||||
from .llm_integration import initialize_llm_and_prompt_builder
|
|
||||||
from haystack.utils import Secret
|
|
||||||
|
|
||||||
|
|
||||||
def build_rag_pipeline(user_id: str) -> tuple[Pipeline, MilvusDocumentStore]:
|
|
||||||
"""
|
|
||||||
为指定用户构建并返回 RAG 查询管道和对应的 DocumentStore。
|
|
||||||
"""
|
|
||||||
print(f"\n--- Building RAG Pipeline for User: {user_id} ---")
|
|
||||||
|
|
||||||
# 1. 初始化该用户的 DocumentStore
|
|
||||||
document_store = initialize_milvus_lite(user_id)
|
|
||||||
|
|
||||||
# 2. 初始化共享组件（可以在应用启动时初始化一次，这里为简单起见每次都创建）
|
|
||||||
text_embedder = initialize_text_embedder()
|
|
||||||
vector_retriever = initialize_vector_retriever(document_store)
|
|
||||||
llm, prompt_builder = initialize_llm_and_prompt_builder()
|
|
||||||
|
|
||||||
# 3. 创建 Haystack Pipeline
|
|
||||||
rag_pipeline = Pipeline()
|
|
||||||
|
|
||||||
# 4. 向管道添加组件，并指定名称
|
|
||||||
rag_pipeline.add_component(instance=text_embedder, name="text_embedder")
|
|
||||||
rag_pipeline.add_component(instance=vector_retriever, name="retriever")
|
|
||||||
rag_pipeline.add_component(instance=prompt_builder, name="prompt_builder")
|
|
||||||
rag_pipeline.add_component(instance=llm, name="llm")
|
|
||||||
|
|
||||||
# 5. 连接管道组件
|
|
||||||
# - 将用户问题文本输入到 text_embedder
|
|
||||||
# - 将 text_embedder 输出的嵌入向量连接到 retriever 的查询嵌入输入
|
|
||||||
# - 将 retriever 输出的文档连接到 prompt_builder 的文档输入
|
|
||||||
# - 将用户问题文本也连接到 prompt_builder 的问题输入
|
|
||||||
# - 将 prompt_builder 输出的完整提示连接到 llm 的提示输入
|
|
||||||
|
|
||||||
rag_pipeline.connect("text_embedder.embedding", "retriever.query_embedding")
|
|
||||||
rag_pipeline.connect("retriever.documents", "prompt_builder.documents")
|
|
||||||
rag_pipeline.connect("prompt_builder.prompt", "llm.prompt")
|
|
||||||
|
|
||||||
print("--- RAG Pipeline Built Successfully ---")
|
|
||||||
# 返回管道和文档存储实例，因为主程序需要用文档存储来写入数据
|
|
||||||
return rag_pipeline, document_store
|
|
||||||
|
|
||||||
|
|
||||||
# --- Corrected Test Block ---
|
|
||||||
if __name__ == "__main__":
|
|
||||||
import os # Needed for API Key check
|
|
||||||
|
|
||||||
# We need OpenAIDocumentEmbedder to index test documents
|
|
||||||
from haystack.components.embedders import OpenAIDocumentEmbedder
|
|
||||||
|
|
||||||
# Import necessary config for initializing the Document Embedder
|
|
||||||
from config import (
|
|
||||||
OPENAI_API_KEY_FROM_CONFIG,
|
|
||||||
OPENAI_API_BASE_URL_CONFIG,
|
|
||||||
OPENAI_EMBEDDING_MODEL,
|
|
||||||
)
|
|
||||||
|
|
||||||
# --- Configuration ---
|
|
||||||
test_user = "test_user"
|
|
||||||
test_query = "Haystack是什么？"
|
|
||||||
# Sample documents to index for testing
|
|
||||||
docs_to_index = [
|
|
||||||
Document(
|
|
||||||
content="Haystack是一个用于构建 NLP 应用程序（如问答系统、语义搜索）的开源框架。",
|
|
||||||
meta={"user_id": test_user, "source": "test_doc_1"},
|
|
||||||
),
|
|
||||||
Document(
|
|
||||||
content="你可以使用 Haystack 连接不同的组件，如文档存储、检索器和生成器。",
|
|
||||||
meta={"user_id": test_user, "source": "test_doc_2"},
|
|
||||||
),
|
|
||||||
Document(
|
|
||||||
content="Milvus 是一个流行的向量数据库，常用于 RAG 系统中存储嵌入。",
|
|
||||||
meta={"user_id": test_user, "source": "test_doc_3"},
|
|
||||||
),
|
|
||||||
]
|
|
||||||
|
|
||||||
print(f"--- Running Test for RAG Pipeline (User: {test_user}) ---")
|
|
||||||
|
|
||||||
# --- 1. Check API Key Availability ---
|
|
||||||
# Pipeline execution requires OpenAI API calls
|
|
||||||
api_key_configured = (
|
|
||||||
OPENAI_API_KEY_FROM_CONFIG and "YOUR_API_KEY" not in OPENAI_API_KEY_FROM_CONFIG
|
|
||||||
)
|
|
||||||
if not api_key_configured:
|
|
||||||
print("\n!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!")
|
|
||||||
print("! WARNING: OpenAI API Key not configured in config.py. !")
|
|
||||||
print("! Skipping RAG pipeline test execution. !")
|
|
||||||
print("!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!")
|
|
||||||
exit() # Exit script if key is missing for test run
|
|
||||||
else:
|
|
||||||
print("\n[Test Setup] OpenAI API Key found in config.")
|
|
||||||
|
|
||||||
# --- 2. Build the RAG Pipeline and get the Document Store ---
|
|
||||||
# This function initializes the store (potentially dropping old data)
|
|
||||||
# and builds the *querying* pipeline.
|
|
||||||
try:
|
|
||||||
pipeline, store = build_rag_pipeline(test_user)
|
|
||||||
except Exception as e:
|
|
||||||
print(f"\nError building RAG pipeline: {e}")
|
|
||||||
import traceback
|
|
||||||
|
|
||||||
traceback.print_exc()
|
|
||||||
exit()
|
|
||||||
|
|
||||||
# --- 3. Index Test Documents (with embeddings) ---
|
|
||||||
print("\n[Test Setup] Initializing Document Embedder for indexing test data...")
|
|
||||||
try:
|
|
||||||
# Initialize the Document Embedder directly here for the test
|
|
||||||
document_embedder = OpenAIDocumentEmbedder(
|
|
||||||
api_key=Secret.from_token(OPENAI_API_KEY_FROM_CONFIG),
|
|
||||||
model=OPENAI_EMBEDDING_MODEL,
|
|
||||||
api_base_url=OPENAI_API_BASE_URL_CONFIG,
|
|
||||||
)
|
|
||||||
print("[Test Setup] Document Embedder initialized.")
|
|
||||||
|
|
||||||
print("[Test Setup] Embedding test documents...")
|
|
||||||
embedding_result = document_embedder.run(docs_to_index)
|
|
||||||
embedded_docs = embedding_result.get("documents", [])
|
|
||||||
|
|
||||||
if embedded_docs:
|
|
||||||
print(
|
|
||||||
f"[Test Setup] Writing {len(embedded_docs)} embedded documents to Milvus..."
|
|
||||||
)
|
|
||||||
store.write_documents(embedded_docs)
|
|
||||||
print("[Test Setup] Test documents written successfully.")
|
|
||||||
# Optional: Verify count
|
|
||||||
# print(f"[Test Setup] Document count in store: {store.count_documents()}")
|
|
||||||
documents_indexed = True
|
|
||||||
else:
|
|
||||||
print("[Test Setup] ERROR: Failed to embed test documents.")
|
|
||||||
documents_indexed = False
|
|
||||||
|
|
||||||
except Exception as e:
|
|
||||||
print(f"\nError during test data indexing: {e}")
|
|
||||||
import traceback
|
|
||||||
|
|
||||||
traceback.print_exc()
|
|
||||||
documents_indexed = False
|
|
||||||
|
|
||||||
# --- 4. Run the RAG Pipeline (if setup succeeded) ---
|
|
||||||
if documents_indexed:
|
|
||||||
print(f"\n[Test Run] Running RAG pipeline for query: '{test_query}'")
|
|
||||||
|
|
||||||
# Prepare input for the RAG pipeline instance built by build_rag_pipeline
|
|
||||||
pipeline_input = {
|
|
||||||
"text_embedder": {"text": test_query}, # Input for the query embedder
|
|
||||||
"prompt_builder": {
|
|
||||||
"query": test_query
|
|
||||||
}, # Input for the prompt builder template
|
|
||||||
}
|
|
||||||
|
|
||||||
try:
|
|
||||||
results = pipeline.run(pipeline_input)
|
|
||||||
|
|
||||||
print("\n[Test Run] Pipeline Results:")
|
|
||||||
# Process and print the generator's answer
|
|
||||||
if "llm" in results and results["llm"]["replies"]:
|
|
||||||
answer = results["llm"]["replies"][0]
|
|
||||||
print(f"\nGenerated Answer: {answer}")
|
|
||||||
else:
|
|
||||||
print("\n[Test Run] Could not extract answer from generator.")
|
|
||||||
print(
|
|
||||||
"Full Pipeline Output:", results
|
|
||||||
) # Print full output for debugging
|
|
||||||
|
|
||||||
except Exception as e:
|
|
||||||
print(f"\n[Test Run] Error running RAG pipeline: {e}")
|
|
||||||
import traceback
|
|
||||||
|
|
||||||
traceback.print_exc()
|
|
||||||
else:
|
|
||||||
print("\n[Test Run] Skipping RAG pipeline execution due to indexing failure.")
|
|
||||||
|
|
||||||
# --- 5. Cleanup Note ---
|
|
||||||
# Optional: Add instructions or commented-out code for cleaning up the test Milvus data
|
|
||||||
print(
|
|
||||||
f"\n[Test Cleanup] Test finished. Consider manually removing data in: ./milvus_user_data_openai_fixed/{test_user}"
|
|
||||||
)
|
|
||||||
# import shutil
|
|
||||||
# from pathlib import Path
|
|
||||||
# from config import MILVUS_PERSIST_BASE_DIR
|
|
||||||
# test_db_path = MILVUS_PERSIST_BASE_DIR / test_user
|
|
||||||
# if test_db_path.exists():
|
|
||||||
# print(f"\nAttempting to clean up test data at {test_db_path}...")
|
|
||||||
# # shutil.rmtree(test_db_path) # Use with caution
|
|
||||||
|
|
||||||
print("\n--- RAG Pipeline Test Complete ---")
|
|
||||||
@ -1,30 +0,0 @@
|
|||||||
# retrieval.py
|
|
||||||
# --- 确认 Import 路径已更新 ---
|
|
||||||
from milvus_haystack import MilvusDocumentStore # 用于类型提示
|
|
||||||
from milvus_haystack.milvus_embedding_retriever import (
|
|
||||||
MilvusEmbeddingRetriever,
|
|
||||||
) # 使用正确的 integration import
|
|
||||||
|
|
||||||
# 从配置导入 top_k
|
|
||||||
from config import RETRIEVER_TOP_K
|
|
||||||
|
|
||||||
|
|
||||||
def initialize_vector_retriever(
|
|
||||||
document_store: MilvusDocumentStore,
|
|
||||||
) -> MilvusEmbeddingRetriever:
|
|
||||||
"""
|
|
||||||
Initializes the MilvusEmbeddingRetriever using milvus-haystack package.
|
|
||||||
Requires a correctly initialized MilvusDocumentStore instance.
|
|
||||||
"""
|
|
||||||
print(f"Initializing Milvus Embedding Retriever with top_k={RETRIEVER_TOP_K}")
|
|
||||||
|
|
||||||
# 初始化 MilvusEmbeddingRetriever 实例
|
|
||||||
# 它需要 document_store 实例来进行实际的搜索操作
|
|
||||||
# top_k 参数控制返回文档的数量
|
|
||||||
retriever = MilvusEmbeddingRetriever(
|
|
||||||
document_store=document_store,
|
|
||||||
top_k=RETRIEVER_TOP_K,
|
|
||||||
# 其他可选参数可以根据需要添加，例如 filters_policy
|
|
||||||
)
|
|
||||||
print("Milvus Embedding Retriever initialized.")
|
|
||||||
return retriever
|
|
||||||
Reference in New Issue
Block a user