从RAG 到KAG

# set the HOME environment variable (only Windows users need to execute this command)
# set HOME=%USERPROFILE%
curl -sSL https://raw.githubusercontent.com/OpenSPG/openspg/refs/heads/master/dev/release/docker-compose-west.yml -o docker-compose-west.yml
docker compose -f docker-compose-west.yml up -d

# Create conda env: conda create -n kag-demo python=3.10 && conda activate kag-demo
# Clone code: git clone https://github.com/OpenSPG/KAG.git
# Install KAG: cd KAG && pip install -e .

# 安装 KAG: cd KAG & & pip Install-e。

{
“database”:”neo4j”, # default datahbase name, which will be replaced by namespace of knowledge base
“uri”:”neo4j://release-openspg-neo4j:7687", # neo4j server address, which can be replaced by customized neo4j server which is accessbile
“user”:”neo4j”, # neo4j username, default to neo4j
“password”:”neo4j@openspg”, # neo4j password, default to neo4j@openspg
}

{
“type”:”openai”, # KAG supports openai compatible interface of embedding service
“model”:”BAAI/bge-m3", # model name of embedding service
“base_url”:”https://api.siliconflow.cn/v1", # url of embedding service
“api_key”:”your api key”
}

{
“biz_scene”:”default”, # biz_scene for kag template
“language”:”en”, # en for english and zh for chinese
}

{
“model”: “deepseek-chat”,
“base_url”: “https://api.deepseek.com",
“api_key”: “deepseek api key”
}

$ vim ./builder/indexer.py

import os
import logging
from kag.common.registry import import_modules_from_path

from kag.builder.runner import BuilderChainRunner

logger = logging.getLogger(__name__)

def buildKB(file_path):
from kag.common.conf import KAG_CONFIG

runner = BuilderChainRunner.from_config(KAG_CONFIG.all_config[“kag_builder_pipeline”])
runner.invoke(file_path)

logger.info(f”\n\nbuildKB successfully for {file_path}\n\n”)

if __name__ == “__main__”:
import_modules_from_path(“.”)
dir_path = os.path.dirname(__file__)
# Set the `file_path` to the path of the previously prepared corpus file.
file_path = os.path.join(dir_path, “data/2wiki_sub_corpus.json”)

$ vim ./solver/qa.py

Paste the following content into qa.py.

import json
import logging
import os
import time
from concurrent.futures import ThreadPoolExecutor, as_completed

from tqdm import tqdm

from kag.common.benchmarks.evaluate import Evaluate
from kag.solver.logic.solver_pipeline import SolverPipeline
from kag.common.conf import KAG_CONFIG
from kag.common.registry import import_modules_from_path

from kag.common.checkpointer import CheckpointerManager

logger = logging.getLogger(__name__)

class EvaFor2wiki:
“””
init for kag client
“””

def __init__(self):
pass

“””
qa from knowledge base,
“””

def qa(self, query):
resp = SolverPipeline.from_config(KAG_CONFIG.all_config[“kag_solver_pipeline”])
answer, traceLog = resp.run(query)

logger.info(f”\n\nso the answer for ‘{query}’ is: {answer}\n\n”)
return answer, traceLog

if __name__ == “__main__”:
import_modules_from_path(“./prompt”)
evalObj = EvaFor2wiki()

evalObj.qa(“Which Stanford University professor works on Alzheimer’s?”)

$ vim ./kag_config.yaml$vim./kg_config.yaml

$ knext project update — proj_path ./$next 项目更新 - proj_path./