diff --git a/llmengine/base_collection.py b/llmengine/base_collection.py new file mode 100644 index 0000000..735fe00 --- /dev/null +++ b/llmengine/base_collection.py @@ -0,0 +1,23 @@ +from abc import ABC, abstractmethod +from typing import Dict + +model_pathMap = {} + +def collection_register(model_key, Klass): + """Register a model class for a given model key.""" + global model_pathMap + model_pathMap[model_key] = Klass + +def get_collection_class(model_path): + """Find the model class for a given model path.""" + for k, klass in model_pathMap.items(): + if len(model_path.split(k)) > 1: + return klass + print(f'{model_pathMap=}') + return None + +class BaseCollection(ABC): + @abstractmethod + def create_collection(self, db_type: str) -> Dict: + """Create a Milvus collection for the given db_type.""" + pass \ No newline at end of file diff --git a/llmengine/base_collection_delete.py b/llmengine/base_collection_delete.py new file mode 100644 index 0000000..b91da09 --- /dev/null +++ b/llmengine/base_collection_delete.py @@ -0,0 +1,23 @@ +from abc import ABC, abstractmethod +from typing import Dict + +model_pathMap = {} + +def collection_delete_register(model_key, Klass): + """Register a model class for a given model key.""" + global model_pathMap + model_pathMap[model_key] = Klass + +def get_collection_delete_class(model_path): + """Find the model class for a given model path.""" + for k, klass in model_pathMap.items(): + if len(model_path.split(k)) > 1: + return klass + print(f'{model_pathMap=}') + return None + +class BaseCollectionDelete(ABC): + @abstractmethod + def delete_collection(self, db_type: str) -> Dict: + """Delete a Milvus collection for the given db_type.""" + pass \ No newline at end of file diff --git a/llmengine/collection.py b/llmengine/collection.py new file mode 100644 index 0000000..8d08251 --- /dev/null +++ b/llmengine/collection.py @@ -0,0 +1,77 @@ +from traceback import format_exc +import os +import sys +import argparse +from llmengine.milvus_collection import * +from llmengine.base_collection import get_collection_class +from typing import Dict + +from appPublic.registerfunction import RegisterFunction +from appPublic.worker import awaitify +from appPublic.log import debug, exception +from ahserver.serverenv import ServerEnv +from ahserver.globalEnv import stream_response +from ahserver.webapp import webserver + +helptext = """Milvus Collection Creation API: + +1. Create Collection Endpoint: +path: /v1/collections +headers: { + "Content-Type": "application/json" +} +data: { + "db_type": "textdb" +} +response: { + "status": "success", + "collection_name": "ragdb_textdb", + "message": "集合 ragdb_textdb 创建成功" +} + +2. Docs Endpoint: +path: /v1/docs +response: This help text +""" + +def init(): + rf = RegisterFunction() + rf.register('collections', create_collection) + rf.register('docs', docs) + +async def docs(request, params_kw, *params, **kw): + return helptext + +async def create_collection(request, params_kw, *params, **kw): + debug(f'{params_kw=}') + se = ServerEnv() + engine = se.engine + f = awaitify(engine.create_collection) + db_type = params_kw.get('db_type') + if db_type is None: + e = exception(f'db_type is None') + raise e + result = await f(db_type) + debug(f'{result=}') + return result + +def main(): + parser = argparse.ArgumentParser(prog="Milvus Collection Service") + parser.add_argument('-w', '--workdir') + parser.add_argument('-p', '--port') + parser.add_argument('model_path') + args = parser.parse_args() + Klass = get_collection_class(args.model_path) + if Klass is None: + e = Exception(f'{args.model_path} has not mapping to a model class') + exception(f'{e}, {format_exc()}') + raise e + se = ServerEnv() + se.engine = Klass(args.model_path) + workdir = args.workdir or os.getcwd() + port = args.port + debug(f'{args=}') + webserver(init, workdir, port) + +if __name__ == '__main__': + main() \ No newline at end of file diff --git a/llmengine/collection_delete.py b/llmengine/collection_delete.py new file mode 100644 index 0000000..caeac16 --- /dev/null +++ b/llmengine/collection_delete.py @@ -0,0 +1,77 @@ +from traceback import format_exc +import os +import sys +import argparse +from llmengine.milvus_collection_delete import * +from llmengine.base_collection_delete import get_collection_delete_class +from typing import Dict + +from appPublic.registerfunction import RegisterFunction +from appPublic.worker import awaitify +from appPublic.log import debug, exception +from ahserver.serverenv import ServerEnv +from ahserver.globalEnv import stream_response +from ahserver.webapp import webserver + +helptext = """Milvus Collection Deletion API: + +1. Delete Collection Endpoint: +path: /v1/collections/delete +headers: { + "Content-Type": "application/json" +} +data: { + "db_type": "textdb" +} +response: { + "status": "success", + "collection_name": "ragdb_textdb", + "message": "集合 ragdb_textdb 删除成功" +} + +2. Docs Endpoint: +path: /v1/docs +response: This help text +""" + +def init(): + rf = RegisterFunction() + rf.register('collections/delete', delete_collection) + rf.register('docs', docs) + +async def docs(request, params_kw, *params, **kw): + return helptext + +async def delete_collection(request, params_kw, *params, **kw): + debug(f'{params_kw=}') + se = ServerEnv() + engine = se.engine + f = awaitify(engine.delete_collection) + db_type = params_kw.get('db_type') + if db_type is None: + e = exception(f'db_type is None') + raise e + result = await f(db_type) + debug(f'{result=}') + return result + +def main(): + parser = argparse.ArgumentParser(prog="Milvus Collection Delete Service") + parser.add_argument('-w', '--workdir') + parser.add_argument('-p', '--port') + parser.add_argument('model_path') + args = parser.parse_args() + Klass = get_collection_delete_class(args.model_path) + if Klass is None: + e = Exception(f'{args.model_path} has not mapping to a model class') + exception(f'{e}, {format_exc()}') + raise e + se = ServerEnv() + se.engine = Klass(args.model_path) + workdir = args.workdir or os.getcwd() + port = args.port + debug(f'{args=}') + webserver(init, workdir, port) + +if __name__ == '__main__': + main() \ No newline at end of file diff --git a/llmengine/milvus_collection.py b/llmengine/milvus_collection.py new file mode 100644 index 0000000..1c062af --- /dev/null +++ b/llmengine/milvus_collection.py @@ -0,0 +1,144 @@ +import os +import logging +from typing import Dict +from pymilvus import connections, utility, Collection, CollectionSchema, FieldSchema, DataType +from llmengine.base_collection import BaseCollection, collection_register + +logger = logging.getLogger(__name__) + +CONFIG_PATH = os.getenv('CONFIG_PATH', '/share/wangmeihua/rag/conf/milvusconfig.yaml') +try: + with open(CONFIG_PATH, 'r', encoding='utf-8') as f: + import yaml + config = yaml.safe_load(f) + MILVUS_DB_PATH = config['database']['milvus_db_path'] +except Exception as e: + logger.error(f"加载配置文件 {CONFIG_PATH} 失败: {str(e)}") + raise RuntimeError(f"无法加载配置文件: {str(e)}") + +class MilvusCollection(BaseCollection): + def __init__(self, model_id): + self.model_id = model_id + self.db_path = MILVUS_DB_PATH + + def ensure_milvus_directory(self) -> None: + """确保 Milvus 数据库目录存在""" + db_dir = os.path.dirname(self.db_path) + if not os.path.exists(db_dir): + os.makedirs(db_dir, exist_ok=True) + logger.debug(f"创建 Milvus 目录: {db_dir}") + if not os.access(db_dir, os.W_OK): + raise RuntimeError(f"Milvus 目录 {db_dir} 不可写") + + def initialize_milvus_connection(self) -> None: + """初始化 Milvus 连接""" + try: + if not connections.has_connection("default"): + connections.connect("default", uri=self.db_path) + logger.debug(f"已连接到 Milvus Lite,路径: {self.db_path}") + else: + logger.debug("已存在 Milvus 连接,跳过重复连接") + except Exception as e: + logger.error(f"连接 Milvus 失败: {str(e)}") + raise RuntimeError(f"连接 Milvus 失败: {str(e)}") + + def cleanup_milvus_connection(self) -> None: + """清理 Milvus 连接""" + try: + if connections.has_connection("default"): + connections.disconnect("default") + logger.debug("已断开 Milvus 连接") + except Exception as e: + logger.warning(f"断开 Milvus 连接失败: {str(e)}") + + def create_collection(self, db_type: str) -> Dict: + """ + 为指定 db_type 创建 Milvus 集合。 + 返回创建结果,包含集合名称和状态。 + """ + try: + # 参数验证 + if not db_type: + raise ValueError("db_type 不能为空") + if "_" in db_type: + raise ValueError("db_type 不能包含下划线") + if len(db_type) > 100: + raise ValueError("db_type 的长度应小于 100") + + self.ensure_milvus_directory() + self.initialize_milvus_connection() + + # 集合名称 + collection_name = f"ragdb_{db_type}" + if len(collection_name) > 255: + raise ValueError(f"集合名称 {collection_name} 超过 255 个字符") + logger.debug(f"集合名称: {collection_name}") + + # 如果集合已存在,检查 schema 兼容性 + if utility.has_collection(collection_name): + try: + collection = Collection(collection_name) + collection.load() + logger.debug(f"集合 {collection_name} 已存在并加载成功") + return { + "status": "success", + "collection_name": collection_name, + "message": f"集合 {collection_name} 已存在" + } + except Exception as e: + logger.error(f"加载集合 {collection_name} 失败: {str(e)}") + raise RuntimeError(f"加载集合失败: {str(e)}") + + # 定义 schema + fields = [ + FieldSchema(name="pk", dtype=DataType.VARCHAR, is_primary=True, max_length=36, auto_id=True), + FieldSchema(name="userid", dtype=DataType.VARCHAR, max_length=100), + FieldSchema(name="document_id", dtype=DataType.VARCHAR, max_length=36), + FieldSchema(name="text", dtype=DataType.VARCHAR, max_length=65535), + FieldSchema(name="vector", dtype=DataType.FLOAT_VECTOR, dim=1024), + FieldSchema(name="filename", dtype=DataType.VARCHAR, max_length=255), + FieldSchema(name="file_path", dtype=DataType.VARCHAR, max_length=1024), + FieldSchema(name="upload_time", dtype=DataType.VARCHAR, max_length=64), + FieldSchema(name="file_type", dtype=DataType.VARCHAR, max_length=64), + ] + schema = CollectionSchema( + fields=fields, + description=f"{db_type} 数据集合,跨用户使用,包含 document_id 和元数据字段", + auto_id=True, + primary_field="pk", + ) + + # 创建集合 + try: + collection = Collection(collection_name, schema) + collection.create_index( + field_name="vector", + index_params={"index_type": "AUTOINDEX", "metric_type": "COSINE"} + ) + for field in ["userid", "document_id", "filename", "file_path", "upload_time", "file_type"]: + collection.create_index( + field_name=field, + index_params={"index_type": "INVERTED"} + ) + collection.load() + logger.debug(f"成功创建并加载集合: {collection_name}") + return { + "status": "success", + "collection_name": collection_name, + "message": f"集合 {collection_name} 创建成功" + } + except Exception as e: + logger.error(f"创建集合 {collection_name} 失败: {str(e)}") + raise RuntimeError(f"创建集合失败: {str(e)}") + + except Exception as e: + logger.error(f"创建集合失败: {str(e)}") + return { + "status": "error", + "collection_name": collection_name, + "message": str(e) + } + finally: + self.cleanup_milvus_connection() + +collection_register('Milvus', MilvusCollection) \ No newline at end of file diff --git a/llmengine/milvus_collection_delete.py b/llmengine/milvus_collection_delete.py new file mode 100644 index 0000000..1eb6aa1 --- /dev/null +++ b/llmengine/milvus_collection_delete.py @@ -0,0 +1,107 @@ +import os +import logging +from typing import Dict +from pymilvus import connections, utility +from llmengine.base_collection_delete import BaseCollectionDelete, collection_delete_register + +logger = logging.getLogger(__name__) + +CONFIG_PATH = os.getenv('CONFIG_PATH', '/share/wangmeihua/rag/conf/milvusconfig.yaml') +try: + with open(CONFIG_PATH, 'r', encoding='utf-8') as f: + import yaml + config = yaml.safe_load(f) + MILVUS_DB_PATH = config['database']['milvus_db_path'] +except Exception as e: + logger.error(f"加载配置文件 {CONFIG_PATH} 失败: {str(e)}") + raise RuntimeError(f"无法加载配置文件: {str(e)}") + +class MilvusCollectionDelete(BaseCollectionDelete): + def __init__(self, model_id): + self.model_id = model_id + self.db_path = MILVUS_DB_PATH + + def ensure_milvus_directory(self) -> None: + """确保 Milvus 数据库目录存在""" + db_dir = os.path.dirname(self.db_path) + if not os.path.exists(db_dir): + os.makedirs(db_dir, exist_ok=True) + logger.debug(f"创建 Milvus 目录: {db_dir}") + if not os.access(db_dir, os.W_OK): + raise RuntimeError(f"Milvus 目录 {db_dir} 不可写") + + def initialize_milvus_connection(self) -> None: + """初始化 Milvus 连接""" + try: + if not connections.has_connection("default"): + connections.connect("default", uri=self.db_path) + logger.debug(f"已连接到 Milvus Lite,路径: {self.db_path}") + else: + logger.debug("已存在 Milvus 连接,跳过重复连接") + except Exception as e: + logger.error(f"连接 Milvus 失败: {str(e)}") + raise RuntimeError(f"连接 Milvus 失败: {str(e)}") + + def cleanup_milvus_connection(self) -> None: + """清理 Milvus 连接""" + try: + if connections.has_connection("default"): + connections.disconnect("default") + logger.debug("已断开 Milvus 连接") + except Exception as e: + logger.warning(f"断开 Milvus 连接失败: {str(e)}") + + def delete_collection(self, db_type: str) -> Dict: + """ + 删除指定 db_type 的 Milvus 集合。 + 返回删除结果,包含集合名称和状态。 + """ + try: + # 参数验证 + if not db_type: + raise ValueError("db_type 不能为空") + if "_" in db_type: + raise ValueError("db_type 不能包含下划线") + if len(db_type) > 100: + raise ValueError("db_type 的长度应小于 100") + + self.ensure_milvus_directory() + self.initialize_milvus_connection() + + # 集合名称 + collection_name = f"ragdb_{db_type}" + logger.debug(f"集合名称: {collection_name}") + + # 检查集合是否存在 + if not utility.has_collection(collection_name): + logger.debug(f"集合 {collection_name} 不存在") + return { + "status": "success", + "collection_name": collection_name, + "message": f"集合 {collection_name} 不存在,无需删除" + } + + # 删除集合 + try: + utility.drop_collection(collection_name) + logger.debug(f"成功删除集合: {collection_name}") + return { + "status": "success", + "collection_name": collection_name, + "message": f"集合 {collection_name} 删除成功" + } + except Exception as e: + logger.error(f"删除集合 {collection_name} 失败: {str(e)}") + raise RuntimeError(f"删除集合失败: {str(e)}") + + except Exception as e: + logger.error(f"删除集合失败: {str(e)}") + return { + "status": "error", + "collection_name": collection_name, + "message": str(e) + } + finally: + self.cleanup_milvus_connection() + +collection_delete_register('Milvus', MilvusCollectionDelete) \ No newline at end of file diff --git a/test/collections/conf/config.json b/test/collections/conf/config.json new file mode 100644 index 0000000..afe36ec --- /dev/null +++ b/test/collections/conf/config.json @@ -0,0 +1,50 @@ +{ + "filesroot":"$[workdir]$/files", + "logger":{ + "name":"llmengine", + "levelname":"info", + "logfile":"$[workdir]$/logs/llmengine.log" + }, + "website":{ + "paths":[ + ["$[workdir]$/wwwroot",""] + ], + "client_max_size":10000, + "host":"0.0.0.0", + "port":9992, + "coding":"utf-8", + "indexes":[ + "index.html", + "index.ui" + ], + "startswiths":[ + { + "leading":"/idfile", + "registerfunction":"idfile" + },{ + "leading": "/v1/collections", + "registerfunction": "collections" + },{ + "leading": "/docs", + "registerfunction": "docs" + } + ], + "processors":[ + [".tmpl","tmpl"], + [".app","app"], + [".ui","bui"], + [".dspy","dspy"], + [".md","md"] + ], + "rsakey_oops":{ + "privatekey":"$[workdir]$/conf/rsa_private_key.pem", + "publickey":"$[workdir]$/conf/rsa_public_key.pem" + }, + "session_max_time":3000, + "session_issue_time":2500, + "session_redis_notuse":{ + "url":"redis://127.0.0.1:6379" + } + } +} + diff --git a/test/collections/logs/llmengine.log b/test/collections/logs/llmengine.log new file mode 100644 index 0000000..e69de29 diff --git a/test/collections/start.sh b/test/collections/start.sh new file mode 100755 index 0000000..d5ef918 --- /dev/null +++ b/test/collections/start.sh @@ -0,0 +1,3 @@ +#!/bin/bash +export CONFIG_PATH=/share/wangmeihua/rag/conf/milvusconfig.yaml +CUDA_VISIBLE_DEVICES=7 /share/vllm-0.8.5/bin/python -m llmengine.collection -p 8889 /share/wangmeihua/rag/conf/Milvus/milvus.db diff --git a/test/deletecollections/conf/config.json b/test/deletecollections/conf/config.json new file mode 100644 index 0000000..32ad3aa --- /dev/null +++ b/test/deletecollections/conf/config.json @@ -0,0 +1,50 @@ +{ + "filesroot":"$[workdir]$/files", + "logger":{ + "name":"llmengine", + "levelname":"info", + "logfile":"$[workdir]$/logs/llmengine.log" + }, + "website":{ + "paths":[ + ["$[workdir]$/wwwroot",""] + ], + "client_max_size":10000, + "host":"0.0.0.0", + "port":8888, + "coding":"utf-8", + "indexes":[ + "index.html", + "index.ui" + ], + "startswiths":[ + { + "leading":"/idfile", + "registerfunction":"idfile" + },{ + "leading": "/v1/collections/delete", + "registerfunction": "collections/delete" + },{ + "leading": "/docs", + "registerfunction": "docs" + } + ], + "processors":[ + [".tmpl","tmpl"], + [".app","app"], + [".ui","bui"], + [".dspy","dspy"], + [".md","md"] + ], + "rsakey_oops":{ + "privatekey":"$[workdir]$/conf/rsa_private_key.pem", + "publickey":"$[workdir]$/conf/rsa_public_key.pem" + }, + "session_max_time":3000, + "session_issue_time":2500, + "session_redis_notuse":{ + "url":"redis://127.0.0.1:6379" + } + } +} + diff --git a/test/deletecollections/logs/llmengine.log b/test/deletecollections/logs/llmengine.log new file mode 100644 index 0000000..e69de29 diff --git a/test/deletecollections/start.sh b/test/deletecollections/start.sh new file mode 100755 index 0000000..89dcbc0 --- /dev/null +++ b/test/deletecollections/start.sh @@ -0,0 +1,3 @@ +#!/bin/bash +export CONFIG_PATH=/share/wangmeihua/rag/conf/milvusconfig.yaml +CUDA_VISIBLE_DEVICES=7 /share/vllm-0.8.5/bin/python -m llmengine.collection_delete -p 8888 /share/wangmeihua/rag/conf/Milvus/milvus.db