MCP OI-Wiki

from pymilvus import MilvusClient from fastembed import TextEmbedding import sys import os import json def main( docs_dir : str = './OI-wiki/docs', db_path : str = './db/oi-wiki.db', embedding_model : str = 'BAAI/bge-small-zh-v1.5' ): client = MilvusClient(db_path) embedding = TextEmbedding(embedding_model) collection_name = "oiwiki" if client.has_collection(collection_name) : client.drop_collection(collection_name) contents, paths = [], [] with open('result.jsonl', 'r') as f: for line in f.readlines() : line = line.strip('\n') data = json.loads(line) path = data["custom_id"] paths.append(path) contents.append(data["response"]["body"]["choices"][0]["message"]["content"]) with open(os.path.join(docs_dir, path), 'r') as raw_f : raw = raw_f.read(512) paths.append(path) contents.append(raw) vectors = list(embedding.embed(contents)) dimension = len(vectors[0]) client.create_collection( collection_name=collection_name, dimension=dimension, ) data = [ {"id": i, "vector": vectors[i], "path": paths[i]} for i in range(len(vectors)) ] client.insert( collection_name=collection_name, data=data ) if __name__ == "__main__" : main(*sys.argv[1:])