Remove deprected demo code
This commit is contained in:
yangdx
@@ -0,0 +1,98 @@
|
||||
"""
|
||||
Sometimes you need to switch a storage solution, but you want to save LLM token and time.
|
||||
This handy script helps you to copy the LLM caches from one storage solution to another.
|
||||
(Not all the storage impl are supported)
|
||||
"""
|
||||
|
||||
import asyncio
|
||||
import logging
|
||||
import os
|
||||
from dotenv import load_dotenv
|
||||
|
||||
from lightrag.kg.postgres_impl import PostgreSQLDB, PGKVStorage
|
||||
from lightrag.kg.json_kv_impl import JsonKVStorage
|
||||
from lightrag.namespace import NameSpace
|
||||
|
||||
load_dotenv()
|
||||
ROOT_DIR = os.environ.get("ROOT_DIR")
|
||||
WORKING_DIR = f"{ROOT_DIR}/dickens"
|
||||
|
||||
logging.basicConfig(format="%(levelname)s:%(message)s", level=logging.INFO)
|
||||
|
||||
if not os.path.exists(WORKING_DIR):
|
||||
os.mkdir(WORKING_DIR)
|
||||
|
||||
# AGE
|
||||
os.environ["AGE_GRAPH_NAME"] = "chinese"
|
||||
|
||||
postgres_db = PostgreSQLDB(
|
||||
config={
|
||||
"host": "localhost",
|
||||
"port": 15432,
|
||||
"user": "rag",
|
||||
"password": "rag",
|
||||
"database": "r2",
|
||||
}
|
||||
)
|
||||
|
||||
|
||||
async def copy_from_postgres_to_json():
|
||||
await postgres_db.initdb()
|
||||
|
||||
from_llm_response_cache = PGKVStorage(
|
||||
namespace=NameSpace.KV_STORE_LLM_RESPONSE_CACHE,
|
||||
global_config={"embedding_batch_num": 6},
|
||||
embedding_func=None,
|
||||
db=postgres_db,
|
||||
)
|
||||
|
||||
to_llm_response_cache = JsonKVStorage(
|
||||
namespace=NameSpace.KV_STORE_LLM_RESPONSE_CACHE,
|
||||
global_config={"working_dir": WORKING_DIR},
|
||||
embedding_func=None,
|
||||
)
|
||||
|
||||
kv = {}
|
||||
for c_id in await from_llm_response_cache.all_keys():
|
||||
print(f"Copying {c_id}")
|
||||
workspace = c_id["workspace"]
|
||||
mode = c_id["mode"]
|
||||
_id = c_id["id"]
|
||||
postgres_db.workspace = workspace
|
||||
obj = await from_llm_response_cache.get_by_mode_and_id(mode, _id)
|
||||
if mode not in kv:
|
||||
kv[mode] = {}
|
||||
kv[mode][_id] = obj[_id]
|
||||
print(f"Object {obj}")
|
||||
await to_llm_response_cache.upsert(kv)
|
||||
await to_llm_response_cache.index_done_callback()
|
||||
print("Mission accomplished!")
|
||||
|
||||
|
||||
async def copy_from_json_to_postgres():
|
||||
await postgres_db.initdb()
|
||||
|
||||
from_llm_response_cache = JsonKVStorage(
|
||||
namespace=NameSpace.KV_STORE_LLM_RESPONSE_CACHE,
|
||||
global_config={"working_dir": WORKING_DIR},
|
||||
embedding_func=None,
|
||||
)
|
||||
|
||||
to_llm_response_cache = PGKVStorage(
|
||||
namespace=NameSpace.KV_STORE_LLM_RESPONSE_CACHE,
|
||||
global_config={"embedding_batch_num": 6},
|
||||
embedding_func=None,
|
||||
db=postgres_db,
|
||||
)
|
||||
|
||||
for mode in await from_llm_response_cache.all_keys():
|
||||
print(f"Copying {mode}")
|
||||
caches = await from_llm_response_cache.get_by_id(mode)
|
||||
for k, v in caches.items():
|
||||
item = {mode: {k: v}}
|
||||
print(f"\tCopying {item}")
|
||||
await to_llm_response_cache.upsert(item)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
asyncio.run(copy_from_json_to_postgres())
|
||||
59
examples/unofficial-sample/lightrag_bedrock_demo.py
Normal file
59
examples/unofficial-sample/lightrag_bedrock_demo.py
Normal file
@@ -0,0 +1,59 @@
|
||||
"""
|
||||
LightRAG meets Amazon Bedrock ⛰️
|
||||
"""
|
||||
|
||||
import os
|
||||
import logging
|
||||
|
||||
from lightrag import LightRAG, QueryParam
|
||||
from lightrag.llm.bedrock import bedrock_complete, bedrock_embed
|
||||
from lightrag.utils import EmbeddingFunc
|
||||
from lightrag.kg.shared_storage import initialize_pipeline_status
|
||||
|
||||
import asyncio
|
||||
import nest_asyncio
|
||||
|
||||
nest_asyncio.apply()
|
||||
|
||||
logging.getLogger("aiobotocore").setLevel(logging.WARNING)
|
||||
|
||||
WORKING_DIR = "./dickens"
|
||||
if not os.path.exists(WORKING_DIR):
|
||||
os.mkdir(WORKING_DIR)
|
||||
|
||||
|
||||
async def initialize_rag():
|
||||
rag = LightRAG(
|
||||
working_dir=WORKING_DIR,
|
||||
llm_model_func=bedrock_complete,
|
||||
llm_model_name="Anthropic Claude 3 Haiku // Amazon Bedrock",
|
||||
embedding_func=EmbeddingFunc(
|
||||
embedding_dim=1024, max_token_size=8192, func=bedrock_embed
|
||||
),
|
||||
)
|
||||
|
||||
await rag.initialize_storages()
|
||||
await initialize_pipeline_status()
|
||||
|
||||
return rag
|
||||
|
||||
|
||||
def main():
|
||||
rag = asyncio.run(initialize_rag())
|
||||
|
||||
with open("./book.txt", "r", encoding="utf-8") as f:
|
||||
rag.insert(f.read())
|
||||
|
||||
for mode in ["naive", "local", "global", "hybrid"]:
|
||||
print("\n+-" + "-" * len(mode) + "-+")
|
||||
print(f"| {mode.capitalize()} |")
|
||||
print("+-" + "-" * len(mode) + "-+\n")
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode=mode)
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
82
examples/unofficial-sample/lightrag_hf_demo.py
Normal file
82
examples/unofficial-sample/lightrag_hf_demo.py
Normal file
@@ -0,0 +1,82 @@
|
||||
import os
|
||||
|
||||
from lightrag import LightRAG, QueryParam
|
||||
from lightrag.llm.hf import hf_model_complete, hf_embed
|
||||
from lightrag.utils import EmbeddingFunc
|
||||
from transformers import AutoModel, AutoTokenizer
|
||||
from lightrag.kg.shared_storage import initialize_pipeline_status
|
||||
|
||||
import asyncio
|
||||
import nest_asyncio
|
||||
|
||||
nest_asyncio.apply()
|
||||
|
||||
WORKING_DIR = "./dickens"
|
||||
|
||||
if not os.path.exists(WORKING_DIR):
|
||||
os.mkdir(WORKING_DIR)
|
||||
|
||||
|
||||
async def initialize_rag():
|
||||
rag = LightRAG(
|
||||
working_dir=WORKING_DIR,
|
||||
llm_model_func=hf_model_complete,
|
||||
llm_model_name="meta-llama/Llama-3.1-8B-Instruct",
|
||||
embedding_func=EmbeddingFunc(
|
||||
embedding_dim=384,
|
||||
max_token_size=5000,
|
||||
func=lambda texts: hf_embed(
|
||||
texts,
|
||||
tokenizer=AutoTokenizer.from_pretrained(
|
||||
"sentence-transformers/all-MiniLM-L6-v2"
|
||||
),
|
||||
embed_model=AutoModel.from_pretrained(
|
||||
"sentence-transformers/all-MiniLM-L6-v2"
|
||||
),
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
await rag.initialize_storages()
|
||||
await initialize_pipeline_status()
|
||||
|
||||
return rag
|
||||
|
||||
|
||||
def main():
|
||||
rag = asyncio.run(initialize_rag())
|
||||
|
||||
with open("./book.txt", "r", encoding="utf-8") as f:
|
||||
rag.insert(f.read())
|
||||
|
||||
# Perform naive search
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="naive")
|
||||
)
|
||||
)
|
||||
|
||||
# Perform local search
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="local")
|
||||
)
|
||||
)
|
||||
|
||||
# Perform global search
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="global")
|
||||
)
|
||||
)
|
||||
|
||||
# Perform hybrid search
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="hybrid")
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
142
examples/unofficial-sample/lightrag_llamaindex_direct_demo.py
Normal file
142
examples/unofficial-sample/lightrag_llamaindex_direct_demo.py
Normal file
@@ -0,0 +1,142 @@
|
||||
import os
|
||||
from lightrag import LightRAG, QueryParam
|
||||
from lightrag.llm.llama_index_impl import (
|
||||
llama_index_complete_if_cache,
|
||||
llama_index_embed,
|
||||
)
|
||||
from lightrag.utils import EmbeddingFunc
|
||||
from llama_index.llms.openai import OpenAI
|
||||
from llama_index.embeddings.openai import OpenAIEmbedding
|
||||
import asyncio
|
||||
import nest_asyncio
|
||||
|
||||
nest_asyncio.apply()
|
||||
|
||||
from lightrag.kg.shared_storage import initialize_pipeline_status
|
||||
|
||||
# Configure working directory
|
||||
WORKING_DIR = "./index_default"
|
||||
print(f"WORKING_DIR: {WORKING_DIR}")
|
||||
|
||||
# Model configuration
|
||||
LLM_MODEL = os.environ.get("LLM_MODEL", "gpt-4")
|
||||
print(f"LLM_MODEL: {LLM_MODEL}")
|
||||
EMBEDDING_MODEL = os.environ.get("EMBEDDING_MODEL", "text-embedding-3-large")
|
||||
print(f"EMBEDDING_MODEL: {EMBEDDING_MODEL}")
|
||||
EMBEDDING_MAX_TOKEN_SIZE = int(os.environ.get("EMBEDDING_MAX_TOKEN_SIZE", 8192))
|
||||
print(f"EMBEDDING_MAX_TOKEN_SIZE: {EMBEDDING_MAX_TOKEN_SIZE}")
|
||||
|
||||
# OpenAI configuration
|
||||
OPENAI_API_KEY = os.environ.get("OPENAI_API_KEY", "your-api-key-here")
|
||||
|
||||
if not os.path.exists(WORKING_DIR):
|
||||
print(f"Creating working directory: {WORKING_DIR}")
|
||||
os.mkdir(WORKING_DIR)
|
||||
|
||||
|
||||
# Initialize LLM function
|
||||
async def llm_model_func(prompt, system_prompt=None, history_messages=[], **kwargs):
|
||||
try:
|
||||
# Initialize OpenAI if not in kwargs
|
||||
if "llm_instance" not in kwargs:
|
||||
llm_instance = OpenAI(
|
||||
model=LLM_MODEL,
|
||||
api_key=OPENAI_API_KEY,
|
||||
temperature=0.7,
|
||||
)
|
||||
kwargs["llm_instance"] = llm_instance
|
||||
|
||||
response = await llama_index_complete_if_cache(
|
||||
kwargs["llm_instance"],
|
||||
prompt,
|
||||
system_prompt=system_prompt,
|
||||
history_messages=history_messages,
|
||||
**kwargs,
|
||||
)
|
||||
return response
|
||||
except Exception as e:
|
||||
print(f"LLM request failed: {str(e)}")
|
||||
raise
|
||||
|
||||
|
||||
# Initialize embedding function
|
||||
async def embedding_func(texts):
|
||||
try:
|
||||
embed_model = OpenAIEmbedding(
|
||||
model=EMBEDDING_MODEL,
|
||||
api_key=OPENAI_API_KEY,
|
||||
)
|
||||
return await llama_index_embed(texts, embed_model=embed_model)
|
||||
except Exception as e:
|
||||
print(f"Embedding failed: {str(e)}")
|
||||
raise
|
||||
|
||||
|
||||
# Get embedding dimension
|
||||
async def get_embedding_dim():
|
||||
test_text = ["This is a test sentence."]
|
||||
embedding = await embedding_func(test_text)
|
||||
embedding_dim = embedding.shape[1]
|
||||
print(f"embedding_dim={embedding_dim}")
|
||||
return embedding_dim
|
||||
|
||||
|
||||
async def initialize_rag():
|
||||
embedding_dimension = await get_embedding_dim()
|
||||
|
||||
rag = LightRAG(
|
||||
working_dir=WORKING_DIR,
|
||||
llm_model_func=llm_model_func,
|
||||
embedding_func=EmbeddingFunc(
|
||||
embedding_dim=embedding_dimension,
|
||||
max_token_size=EMBEDDING_MAX_TOKEN_SIZE,
|
||||
func=embedding_func,
|
||||
),
|
||||
)
|
||||
|
||||
await rag.initialize_storages()
|
||||
await initialize_pipeline_status()
|
||||
|
||||
return rag
|
||||
|
||||
|
||||
def main():
|
||||
# Initialize RAG instance
|
||||
rag = asyncio.run(initialize_rag())
|
||||
|
||||
# Insert example text
|
||||
with open("./book.txt", "r", encoding="utf-8") as f:
|
||||
rag.insert(f.read())
|
||||
|
||||
# Test different query modes
|
||||
print("\nNaive Search:")
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="naive")
|
||||
)
|
||||
)
|
||||
|
||||
print("\nLocal Search:")
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="local")
|
||||
)
|
||||
)
|
||||
|
||||
print("\nGlobal Search:")
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="global")
|
||||
)
|
||||
)
|
||||
|
||||
print("\nHybrid Search:")
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="hybrid")
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
145
examples/unofficial-sample/lightrag_llamaindex_litellm_demo.py
Normal file
145
examples/unofficial-sample/lightrag_llamaindex_litellm_demo.py
Normal file
@@ -0,0 +1,145 @@
|
||||
import os
|
||||
from lightrag import LightRAG, QueryParam
|
||||
from lightrag.llm.llama_index_impl import (
|
||||
llama_index_complete_if_cache,
|
||||
llama_index_embed,
|
||||
)
|
||||
from lightrag.utils import EmbeddingFunc
|
||||
from llama_index.llms.litellm import LiteLLM
|
||||
from llama_index.embeddings.litellm import LiteLLMEmbedding
|
||||
import asyncio
|
||||
import nest_asyncio
|
||||
|
||||
nest_asyncio.apply()
|
||||
|
||||
from lightrag.kg.shared_storage import initialize_pipeline_status
|
||||
|
||||
# Configure working directory
|
||||
WORKING_DIR = "./index_default"
|
||||
print(f"WORKING_DIR: {WORKING_DIR}")
|
||||
|
||||
# Model configuration
|
||||
LLM_MODEL = os.environ.get("LLM_MODEL", "gpt-4")
|
||||
print(f"LLM_MODEL: {LLM_MODEL}")
|
||||
EMBEDDING_MODEL = os.environ.get("EMBEDDING_MODEL", "text-embedding-3-large")
|
||||
print(f"EMBEDDING_MODEL: {EMBEDDING_MODEL}")
|
||||
EMBEDDING_MAX_TOKEN_SIZE = int(os.environ.get("EMBEDDING_MAX_TOKEN_SIZE", 8192))
|
||||
print(f"EMBEDDING_MAX_TOKEN_SIZE: {EMBEDDING_MAX_TOKEN_SIZE}")
|
||||
|
||||
# LiteLLM configuration
|
||||
LITELLM_URL = os.environ.get("LITELLM_URL", "http://localhost:4000")
|
||||
print(f"LITELLM_URL: {LITELLM_URL}")
|
||||
LITELLM_KEY = os.environ.get("LITELLM_KEY", "sk-1234")
|
||||
|
||||
if not os.path.exists(WORKING_DIR):
|
||||
os.mkdir(WORKING_DIR)
|
||||
|
||||
|
||||
# Initialize LLM function
|
||||
async def llm_model_func(prompt, system_prompt=None, history_messages=[], **kwargs):
|
||||
try:
|
||||
# Initialize LiteLLM if not in kwargs
|
||||
if "llm_instance" not in kwargs:
|
||||
llm_instance = LiteLLM(
|
||||
model=f"openai/{LLM_MODEL}", # Format: "provider/model_name"
|
||||
api_base=LITELLM_URL,
|
||||
api_key=LITELLM_KEY,
|
||||
temperature=0.7,
|
||||
)
|
||||
kwargs["llm_instance"] = llm_instance
|
||||
|
||||
response = await llama_index_complete_if_cache(
|
||||
kwargs["llm_instance"],
|
||||
prompt,
|
||||
system_prompt=system_prompt,
|
||||
history_messages=history_messages,
|
||||
**kwargs,
|
||||
)
|
||||
return response
|
||||
except Exception as e:
|
||||
print(f"LLM request failed: {str(e)}")
|
||||
raise
|
||||
|
||||
|
||||
# Initialize embedding function
|
||||
async def embedding_func(texts):
|
||||
try:
|
||||
embed_model = LiteLLMEmbedding(
|
||||
model_name=f"openai/{EMBEDDING_MODEL}",
|
||||
api_base=LITELLM_URL,
|
||||
api_key=LITELLM_KEY,
|
||||
)
|
||||
return await llama_index_embed(texts, embed_model=embed_model)
|
||||
except Exception as e:
|
||||
print(f"Embedding failed: {str(e)}")
|
||||
raise
|
||||
|
||||
|
||||
# Get embedding dimension
|
||||
async def get_embedding_dim():
|
||||
test_text = ["This is a test sentence."]
|
||||
embedding = await embedding_func(test_text)
|
||||
embedding_dim = embedding.shape[1]
|
||||
print(f"embedding_dim={embedding_dim}")
|
||||
return embedding_dim
|
||||
|
||||
|
||||
async def initialize_rag():
|
||||
embedding_dimension = await get_embedding_dim()
|
||||
|
||||
rag = LightRAG(
|
||||
working_dir=WORKING_DIR,
|
||||
llm_model_func=llm_model_func,
|
||||
embedding_func=EmbeddingFunc(
|
||||
embedding_dim=embedding_dimension,
|
||||
max_token_size=EMBEDDING_MAX_TOKEN_SIZE,
|
||||
func=embedding_func,
|
||||
),
|
||||
)
|
||||
|
||||
await rag.initialize_storages()
|
||||
await initialize_pipeline_status()
|
||||
|
||||
return rag
|
||||
|
||||
|
||||
def main():
|
||||
# Initialize RAG instance
|
||||
rag = asyncio.run(initialize_rag())
|
||||
|
||||
# Insert example text
|
||||
with open("./book.txt", "r", encoding="utf-8") as f:
|
||||
rag.insert(f.read())
|
||||
|
||||
# Test different query modes
|
||||
print("\nNaive Search:")
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="naive")
|
||||
)
|
||||
)
|
||||
|
||||
print("\nLocal Search:")
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="local")
|
||||
)
|
||||
)
|
||||
|
||||
print("\nGlobal Search:")
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="global")
|
||||
)
|
||||
)
|
||||
|
||||
print("\nHybrid Search:")
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="hybrid")
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
110
examples/unofficial-sample/lightrag_lmdeploy_demo.py
Normal file
110
examples/unofficial-sample/lightrag_lmdeploy_demo.py
Normal file
@@ -0,0 +1,110 @@
|
||||
import os
|
||||
|
||||
from lightrag import LightRAG, QueryParam
|
||||
from lightrag.llm.lmdeploy import lmdeploy_model_if_cache
|
||||
from lightrag.llm.hf import hf_embed
|
||||
from lightrag.utils import EmbeddingFunc
|
||||
from transformers import AutoModel, AutoTokenizer
|
||||
from lightrag.kg.shared_storage import initialize_pipeline_status
|
||||
|
||||
import asyncio
|
||||
import nest_asyncio
|
||||
|
||||
nest_asyncio.apply()
|
||||
|
||||
WORKING_DIR = "./dickens"
|
||||
|
||||
if not os.path.exists(WORKING_DIR):
|
||||
os.mkdir(WORKING_DIR)
|
||||
|
||||
|
||||
async def lmdeploy_model_complete(
|
||||
prompt=None,
|
||||
system_prompt=None,
|
||||
history_messages=[],
|
||||
keyword_extraction=False,
|
||||
**kwargs,
|
||||
) -> str:
|
||||
model_name = kwargs["hashing_kv"].global_config["llm_model_name"]
|
||||
return await lmdeploy_model_if_cache(
|
||||
model_name,
|
||||
prompt,
|
||||
system_prompt=system_prompt,
|
||||
history_messages=history_messages,
|
||||
## please specify chat_template if your local path does not follow original HF file name,
|
||||
## or model_name is a pytorch model on huggingface.co,
|
||||
## you can refer to https://github.com/InternLM/lmdeploy/blob/main/lmdeploy/model.py
|
||||
## for a list of chat_template available in lmdeploy.
|
||||
chat_template="llama3",
|
||||
# model_format ='awq', # if you are using awq quantization model.
|
||||
# quant_policy=8, # if you want to use online kv cache, 4=kv int4, 8=kv int8.
|
||||
**kwargs,
|
||||
)
|
||||
|
||||
|
||||
async def initialize_rag():
|
||||
rag = LightRAG(
|
||||
working_dir=WORKING_DIR,
|
||||
llm_model_func=lmdeploy_model_complete,
|
||||
llm_model_name="meta-llama/Llama-3.1-8B-Instruct", # please use definite path for local model
|
||||
embedding_func=EmbeddingFunc(
|
||||
embedding_dim=384,
|
||||
max_token_size=5000,
|
||||
func=lambda texts: hf_embed(
|
||||
texts,
|
||||
tokenizer=AutoTokenizer.from_pretrained(
|
||||
"sentence-transformers/all-MiniLM-L6-v2"
|
||||
),
|
||||
embed_model=AutoModel.from_pretrained(
|
||||
"sentence-transformers/all-MiniLM-L6-v2"
|
||||
),
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
await rag.initialize_storages()
|
||||
await initialize_pipeline_status()
|
||||
|
||||
return rag
|
||||
|
||||
|
||||
def main():
|
||||
# Initialize RAG instance
|
||||
rag = asyncio.run(initialize_rag())
|
||||
|
||||
# Insert example text
|
||||
with open("./book.txt", "r", encoding="utf-8") as f:
|
||||
rag.insert(f.read())
|
||||
|
||||
# Test different query modes
|
||||
print("\nNaive Search:")
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="naive")
|
||||
)
|
||||
)
|
||||
|
||||
print("\nLocal Search:")
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="local")
|
||||
)
|
||||
)
|
||||
|
||||
print("\nGlobal Search:")
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="global")
|
||||
)
|
||||
)
|
||||
|
||||
print("\nHybrid Search:")
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="hybrid")
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
171
examples/unofficial-sample/lightrag_nvidia_demo.py
Normal file
171
examples/unofficial-sample/lightrag_nvidia_demo.py
Normal file
@@ -0,0 +1,171 @@
|
||||
import os
|
||||
import asyncio
|
||||
import nest_asyncio
|
||||
|
||||
from lightrag import LightRAG, QueryParam
|
||||
from lightrag.llm import (
|
||||
openai_complete_if_cache,
|
||||
nvidia_openai_embed,
|
||||
)
|
||||
from lightrag.utils import EmbeddingFunc
|
||||
import numpy as np
|
||||
from lightrag.kg.shared_storage import initialize_pipeline_status
|
||||
|
||||
# for custom llm_model_func
|
||||
from lightrag.utils import locate_json_string_body_from_string
|
||||
|
||||
nest_asyncio.apply()
|
||||
|
||||
WORKING_DIR = "./dickens"
|
||||
|
||||
if not os.path.exists(WORKING_DIR):
|
||||
os.mkdir(WORKING_DIR)
|
||||
|
||||
# some method to use your API key (choose one)
|
||||
# NVIDIA_OPENAI_API_KEY = os.getenv("NVIDIA_OPENAI_API_KEY")
|
||||
NVIDIA_OPENAI_API_KEY = "nvapi-xxxx" # your api key
|
||||
|
||||
# using pre-defined function for nvidia LLM API. OpenAI compatible
|
||||
# llm_model_func = nvidia_openai_complete
|
||||
|
||||
|
||||
# If you trying to make custom llm_model_func to use llm model on NVIDIA API like other example:
|
||||
async def llm_model_func(
|
||||
prompt, system_prompt=None, history_messages=[], keyword_extraction=False, **kwargs
|
||||
) -> str:
|
||||
result = await openai_complete_if_cache(
|
||||
"nvidia/llama-3.1-nemotron-70b-instruct",
|
||||
prompt,
|
||||
system_prompt=system_prompt,
|
||||
history_messages=history_messages,
|
||||
api_key=NVIDIA_OPENAI_API_KEY,
|
||||
base_url="https://integrate.api.nvidia.com/v1",
|
||||
**kwargs,
|
||||
)
|
||||
if keyword_extraction:
|
||||
return locate_json_string_body_from_string(result)
|
||||
return result
|
||||
|
||||
|
||||
# custom embedding
|
||||
nvidia_embed_model = "nvidia/nv-embedqa-e5-v5"
|
||||
|
||||
|
||||
async def indexing_embedding_func(texts: list[str]) -> np.ndarray:
|
||||
return await nvidia_openai_embed(
|
||||
texts,
|
||||
model=nvidia_embed_model, # maximum 512 token
|
||||
# model="nvidia/llama-3.2-nv-embedqa-1b-v1",
|
||||
api_key=NVIDIA_OPENAI_API_KEY,
|
||||
base_url="https://integrate.api.nvidia.com/v1",
|
||||
input_type="passage",
|
||||
trunc="END", # handling on server side if input token is longer than maximum token
|
||||
encode="float",
|
||||
)
|
||||
|
||||
|
||||
async def query_embedding_func(texts: list[str]) -> np.ndarray:
|
||||
return await nvidia_openai_embed(
|
||||
texts,
|
||||
model=nvidia_embed_model, # maximum 512 token
|
||||
# model="nvidia/llama-3.2-nv-embedqa-1b-v1",
|
||||
api_key=NVIDIA_OPENAI_API_KEY,
|
||||
base_url="https://integrate.api.nvidia.com/v1",
|
||||
input_type="query",
|
||||
trunc="END", # handling on server side if input token is longer than maximum token
|
||||
encode="float",
|
||||
)
|
||||
|
||||
|
||||
# dimension are same
|
||||
async def get_embedding_dim():
|
||||
test_text = ["This is a test sentence."]
|
||||
embedding = await indexing_embedding_func(test_text)
|
||||
embedding_dim = embedding.shape[1]
|
||||
return embedding_dim
|
||||
|
||||
|
||||
# function test
|
||||
async def test_funcs():
|
||||
result = await llm_model_func("How are you?")
|
||||
print("llm_model_func: ", result)
|
||||
|
||||
result = await indexing_embedding_func(["How are you?"])
|
||||
print("embedding_func: ", result)
|
||||
|
||||
|
||||
# asyncio.run(test_funcs())
|
||||
|
||||
|
||||
async def initialize_rag():
|
||||
embedding_dimension = await get_embedding_dim()
|
||||
print(f"Detected embedding dimension: {embedding_dimension}")
|
||||
|
||||
# lightRAG class during indexing
|
||||
rag = LightRAG(
|
||||
working_dir=WORKING_DIR,
|
||||
llm_model_func=llm_model_func,
|
||||
# llm_model_name="meta/llama3-70b-instruct", #un comment if
|
||||
embedding_func=EmbeddingFunc(
|
||||
embedding_dim=embedding_dimension,
|
||||
max_token_size=512, # maximum token size, somehow it's still exceed maximum number of token
|
||||
# so truncate (trunc) parameter on embedding_func will handle it and try to examine the tokenizer used in LightRAG
|
||||
# so you can adjust to be able to fit the NVIDIA model (future work)
|
||||
func=indexing_embedding_func,
|
||||
),
|
||||
)
|
||||
|
||||
await rag.initialize_storages()
|
||||
await initialize_pipeline_status()
|
||||
|
||||
return rag
|
||||
|
||||
|
||||
async def main():
|
||||
try:
|
||||
# Initialize RAG instance
|
||||
rag = await initialize_rag()
|
||||
|
||||
# reading file
|
||||
with open("./book.txt", "r", encoding="utf-8") as f:
|
||||
await rag.ainsert(f.read())
|
||||
|
||||
# Perform naive search
|
||||
print("==============Naive===============")
|
||||
print(
|
||||
await rag.aquery(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="naive")
|
||||
)
|
||||
)
|
||||
|
||||
# Perform local search
|
||||
print("==============local===============")
|
||||
print(
|
||||
await rag.aquery(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="local")
|
||||
)
|
||||
)
|
||||
|
||||
# Perform global search
|
||||
print("==============global===============")
|
||||
print(
|
||||
await rag.aquery(
|
||||
"What are the top themes in this story?",
|
||||
param=QueryParam(mode="global"),
|
||||
)
|
||||
)
|
||||
|
||||
# Perform hybrid search
|
||||
print("==============hybrid===============")
|
||||
print(
|
||||
await rag.aquery(
|
||||
"What are the top themes in this story?",
|
||||
param=QueryParam(mode="hybrid"),
|
||||
)
|
||||
)
|
||||
except Exception as e:
|
||||
print(f"An error occurred: {e}")
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
asyncio.run(main())
|
||||
@@ -0,0 +1,112 @@
|
||||
import os
|
||||
import asyncio
|
||||
from lightrag import LightRAG, QueryParam
|
||||
from lightrag.llm.ollama import ollama_embed, openai_complete_if_cache
|
||||
from lightrag.utils import EmbeddingFunc
|
||||
from lightrag.kg.shared_storage import initialize_pipeline_status
|
||||
|
||||
# WorkingDir
|
||||
ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
|
||||
WORKING_DIR = os.path.join(ROOT_DIR, "myKG")
|
||||
if not os.path.exists(WORKING_DIR):
|
||||
os.mkdir(WORKING_DIR)
|
||||
print(f"WorkingDir: {WORKING_DIR}")
|
||||
|
||||
# redis
|
||||
os.environ["REDIS_URI"] = "redis://localhost:6379"
|
||||
|
||||
# neo4j
|
||||
BATCH_SIZE_NODES = 500
|
||||
BATCH_SIZE_EDGES = 100
|
||||
os.environ["NEO4J_URI"] = "bolt://117.50.173.35:7687"
|
||||
os.environ["NEO4J_USERNAME"] = "neo4j"
|
||||
os.environ["NEO4J_PASSWORD"] = "12345678"
|
||||
|
||||
# milvus
|
||||
os.environ["MILVUS_URI"] = "http://117.50.173.35:19530"
|
||||
os.environ["MILVUS_USER"] = "root"
|
||||
os.environ["MILVUS_PASSWORD"] = "Milvus"
|
||||
os.environ["MILVUS_DB_NAME"] = "lightrag"
|
||||
|
||||
|
||||
async def llm_model_func(
|
||||
prompt, system_prompt=None, history_messages=[], keyword_extraction=False, **kwargs
|
||||
) -> str:
|
||||
return await openai_complete_if_cache(
|
||||
"deepseek-chat",
|
||||
prompt,
|
||||
system_prompt=system_prompt,
|
||||
history_messages=history_messages,
|
||||
api_key="",
|
||||
base_url="",
|
||||
**kwargs,
|
||||
)
|
||||
|
||||
|
||||
embedding_func = EmbeddingFunc(
|
||||
embedding_dim=768,
|
||||
max_token_size=512,
|
||||
func=lambda texts: ollama_embed(
|
||||
texts, embed_model="shaw/dmeta-embedding-zh", host="http://117.50.173.35:11434"
|
||||
),
|
||||
)
|
||||
|
||||
|
||||
async def initialize_rag():
|
||||
rag = LightRAG(
|
||||
working_dir=WORKING_DIR,
|
||||
llm_model_func=llm_model_func,
|
||||
llm_model_max_token_size=32768,
|
||||
embedding_func=embedding_func,
|
||||
chunk_token_size=512,
|
||||
chunk_overlap_token_size=256,
|
||||
kv_storage="RedisKVStorage",
|
||||
graph_storage="Neo4JStorage",
|
||||
vector_storage="MilvusVectorDBStorage",
|
||||
doc_status_storage="RedisKVStorage",
|
||||
)
|
||||
|
||||
await rag.initialize_storages()
|
||||
await initialize_pipeline_status()
|
||||
|
||||
return rag
|
||||
|
||||
|
||||
def main():
|
||||
# Initialize RAG instance
|
||||
rag = asyncio.run(initialize_rag())
|
||||
|
||||
with open("./book.txt", "r", encoding="utf-8") as f:
|
||||
rag.insert(f.read())
|
||||
|
||||
# Perform naive search
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="naive")
|
||||
)
|
||||
)
|
||||
|
||||
# Perform local search
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="local")
|
||||
)
|
||||
)
|
||||
|
||||
# Perform global search
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="global")
|
||||
)
|
||||
)
|
||||
|
||||
# Perform hybrid search
|
||||
print(
|
||||
rag.query(
|
||||
"What are the top themes in this story?", param=QueryParam(mode="hybrid")
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
Reference in New Issue
Block a user