isage-rag-benchmark 0.1.0.1__cp311-none-any.whl

isage_rag_benchmark-0.1.0.1.dist-info/METADATA

@@ -0,0 +1,63 @@
+Metadata-Version: 2.4
+Name: isage-rag-benchmark
+Version: 0.1.0.1
+Summary: RAG benchmarking suite for the SAGE ecosystem
+Author-email: IntelliStream Team <shuhao_zhang@hust.edu.cn>
+License-Expression: MIT
+Project-URL: Homepage, https://github.com/intellistream/sage-rag-benchmark
+Project-URL: Documentation, https://github.com/intellistream/sage-rag-benchmark#readme
+Project-URL: Repository, https://github.com/intellistream/sage-rag-benchmark
+Project-URL: Issues, https://github.com/intellistream/sage-rag-benchmark/issues
+Keywords: sage,benchmark,rag,retrieval-augmented-generation,evaluation,intellistream
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: isage-common
+Requires-Dist: isage-kernel
+Requires-Dist: isage-middleware>=0.2.4.0
+Requires-Dist: isage-libs
+Requires-Dist: sentence-transformers<4.0.0,>=3.1.0
+Requires-Dist: pyyaml>=6.0
+Requires-Dist: chromadb>=1.0.20
+Requires-Dist: pymilvus[model]>=2.4.0
+Requires-Dist: faiss-cpu<2.0.0,>=1.7.0
+Requires-Dist: pandas>=2.0.0
+Requires-Dist: numpy<2.3.0,>=1.26.0
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0.0; extra == "dev"
+Requires-Dist: pytest-cov>=4.0.0; extra == "dev"
+Requires-Dist: black>=23.0.0; extra == "dev"
+Requires-Dist: ruff==0.14.6; extra == "dev"
+Requires-Dist: pre-commit>=3.0.0; extra == "dev"
+Requires-Dist: mypy>=1.0.0; extra == "dev"
+Requires-Dist: types-PyYAML>=6.0.0; extra == "dev"
+Dynamic: license-file
+
+# SAGE RAG Benchmark
+
+RAG benchmarking suite for the SAGE ecosystem. This repository provides
+pipelines, evaluation utilities, and configuration templates for evaluating
+retrieval-augmented generation systems.
+
+## Install
+
+```bash
+pip install isage-rag-benchmark
+```
+
+## Quick Start
+
+See the configs under `src/sage/benchmark_rag/config/` and the implementations
+in `src/sage/benchmark_rag/implementations/`.
+
+## Repository
+
+- GitHub: https://github.com/intellistream/sage-rag-benchmark

isage_rag_benchmark-0.1.0.1.dist-info/RECORD

@@ -0,0 +1,48 @@
+isage_rag_benchmark-0.1.0.1.dist-info/licenses/LICENSE,sha256=vBNVIGkYYZY0B8f0Ui1ITYwRu7WNtSwyxvIAVGYS6jU,1075
+sage/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+sage/benchmark_rag/__init__.py,sha256=7p8QRU5ShDaYMjqiFOVr_1aAlMULm6z8Mle39sX5c2Y,561
+sage/benchmark_rag/_version.py,sha256=qd_nvt_nQyHmhAy0bKVhzaD5BSHwZtyFAFIfgbH2EuY,147
+sage/benchmark_rag/config/config_bm25s.yaml,sha256=oSAKyk4YoU_C65c6EDo8vVZxFuFOkvsVBaedgLeW7m0,905
+sage/benchmark_rag/config/config_dense_milvus.yaml,sha256=OjB6r06n1nvpeh1_vh-cGQ3fOddh10L4Xqaodsh1Org,1716
+sage/benchmark_rag/config/config_hf.yaml,sha256=TGS8_05qRiElPpBy0A5rLlZvKsCbJuqo1bVtW8e8fa0,1008
+sage/benchmark_rag/config/config_mixed.yaml,sha256=wbSquG4SQYxM1teuXvPx559wXtCA8imZ7BxbNnVgIbw,954
+sage/benchmark_rag/config/config_monitoring_demo.yaml,sha256=iYE5Kh8UiIGeVq0jFq-1cPwQLLf-vqDbBLLZ2y4nshY,1622
+sage/benchmark_rag/config/config_multiplex.yaml,sha256=R2uyAw5Jxzge_ohEjfi_18i25vtd3qODChNijYqiv-M,1605
+sage/benchmark_rag/config/config_qa_chroma.yaml,sha256=WzOvBNYMJ-CAkV1yTkJFdDYJ_Uv_e-fc8my4nq3fl6I,1168
+sage/benchmark_rag/config/config_ray.yaml,sha256=1J-ZGBAEyKXb_WERAAQuS9LTWmaxL1ViugGRSgbQZxc,999
+sage/benchmark_rag/config/config_refiner.yaml,sha256=Y-8UDsEqfptZPchdpZRgk-F-uPgj7hF_5PBuuXo_GlU,1976
+sage/benchmark_rag/config/config_rerank.yaml,sha256=Rp65KPCjQYaA53VMTe9DglAoXuHYy-27XNfJQI2De9o,1253
+sage/benchmark_rag/config/config_selfrag.yaml,sha256=kZlwdMsttTiuDQ7CQq-hF4Z4pJhHGvF0K6h4OL7uCcQ,657
+sage/benchmark_rag/config/config_source.yaml,sha256=oP19eedcvFfB654iFgFfI6pweQOigSq6vuy5YkOZz1g,554
+sage/benchmark_rag/config/config_source_local.yaml,sha256=gz2BwBQX4akdodhuS5finNXnSYgxC6VdgdqMp191ZUc,378
+sage/benchmark_rag/config/config_sparse_milvus.yaml,sha256=TvAM4CqYqTPDJ7_ZknKx8kNBLwM0R7BYz9-7LzpC0FA,1308
+sage/benchmark_rag/evaluation/__init__.py,sha256=dfDLqVIsgMV7WCZKVjMzMFfZUAC6vUZ4miOCZZvvWUA,296
+sage/benchmark_rag/evaluation/benchmark_runner.py,sha256=8groX0mdvTnnld8a_ZrMkrTqOzHBuoidYvVag9svNQc,10840
+sage/benchmark_rag/evaluation/evaluate_results.py,sha256=jma6SGq8DSNoasceGz2o55xyraTXasKOYptD7YFxcxA,13294
+sage/benchmark_rag/evaluation/config/benchmark_config.yaml,sha256=4tOcF6Af0t9FaAFVH7Qxyfxc2Xm8JCJzYpBceUyPa5E,1151
+sage/benchmark_rag/implementations/__init__.py,sha256=XFgI3OxoEvimYZ1lAA7xTAIHWmsUIP8ujGQc0JEyolE,812
+sage/benchmark_rag/implementations/pipelines/__init__.py,sha256=2jv6g8Cd_N6g7uk-qTMcEllhdfrFfVCz4VUGmK55TDc,630
+sage/benchmark_rag/implementations/pipelines/qa_bm25_retrieval.py,sha256=s1QrQE4gyagHVY1Yt3S1idp2hhSxi7qwNzFG0Q0Wx9I,2207
+sage/benchmark_rag/implementations/pipelines/qa_dense_retrieval.py,sha256=jZEtC_gyOcpf6oauNYsdm5EmGtwf9hplXujXf2eWvdE,2064
+sage/benchmark_rag/implementations/pipelines/qa_dense_retrieval_chroma.py,sha256=scFrU93cU82xzzZPtKh53w7mKyqmnaLOclTxU0rypEo,2506
+sage/benchmark_rag/implementations/pipelines/qa_dense_retrieval_milvus.py,sha256=Y8G2ForbqZz7kUXPHhRj5VLNT_56zfgSdoVXb-n5Flw,2784
+sage/benchmark_rag/implementations/pipelines/qa_dense_retrieval_mixed.py,sha256=7DfoeCiUV6c1282rdWdfzRCG_x-S1mJSXCtjo-2Nzfw,2259
+sage/benchmark_rag/implementations/pipelines/qa_dense_retrieval_ray.py,sha256=3CQd_Cr1QmmqpQl0goUCsPmvXd_YHNh5lQ5uh1oK67A,6582
+sage/benchmark_rag/implementations/pipelines/qa_hf_model.py,sha256=XcGLVUkQoAdOIMjM8mdttTNKePyA7EGEjtnQtz0AY6A,1790
+sage/benchmark_rag/implementations/pipelines/qa_monitoring_demo.py,sha256=Go6GBcW-dgsEWVmruwb76tsuS3vcYeTFfSxOv4vs5Uk,5861
+sage/benchmark_rag/implementations/pipelines/qa_multimodal_fusion.py,sha256=UhzszEQteBTNTsiC03guIgNst93VBOgu9k2NrrKHWCg,10855
+sage/benchmark_rag/implementations/pipelines/qa_multiplex.py,sha256=cTvn1k1gHbcaTgmKZnhahmaPGy0FgjyFD_GzMyz2jFE,3241
+sage/benchmark_rag/implementations/pipelines/qa_refiner.py,sha256=68TpwwvcCQ8mCMFEzHhWI0xigBzXMhVWOSv0xk51ALw,3126
+sage/benchmark_rag/implementations/pipelines/qa_rerank.py,sha256=ueDnEFRXzRJFKWZJbk7PLWXIfwUqG8ix2q3PlVozilY,2225
+sage/benchmark_rag/implementations/pipelines/qa_sparse_retrieval_milvus.py,sha256=u2iD1CIedv-uuuOdCG4VZ6H_2ZT6TS3mRvMFRYYYc58,2648
+sage/benchmark_rag/implementations/pipelines/selfrag.py,sha256=DeIOCBqRYz7u7rmPUDrZd10dppDJWmubIQsZrTL7LfU,6914
+sage/benchmark_rag/implementations/tools/__init__.py,sha256=Wm5XCdGogWYRamK6qwja1VWTMLkLthbvsT8yZetV3zQ,528
+sage/benchmark_rag/implementations/tools/build_chroma_index.py,sha256=sEDlRoRcxzDCUId4UMQZFG0Pbw04-D3xNn8_12cS-1Y,9139
+sage/benchmark_rag/implementations/tools/build_milvus_dense_index.py,sha256=zQYxIoZLiS47owcqIj0wiLBYmgiRQgaJx_7hf2BIkTE,3010
+sage/benchmark_rag/implementations/tools/build_milvus_index.py,sha256=HLsG7CGWqSZkzppiawqqnHP0bFw9XVoH_9sM0E39X_o,2153
+sage/benchmark_rag/implementations/tools/build_milvus_sparse_index.py,sha256=Z4ucYYOCFx_CYlM7eM2-Q-php178Naminrgd6TXjJ-k,2994
+sage/benchmark_rag/implementations/tools/loaders/document_loaders.py,sha256=LfXWZ_H1bxiBOlx2Vk9g63TbNSVW1N2nGS6F5tOBFCM,1196
+isage_rag_benchmark-0.1.0.1.dist-info/METADATA,sha256=N1kRfa8YLjkJZ3nyajpk6G9gEIHv5CNcRJTIikIFy4s,2339
+isage_rag_benchmark-0.1.0.1.dist-info/WHEEL,sha256=yk-B4c9kYsinhQ_MzhPAVcDm9mhkAVmdo0rg0jgFCmo,94
+isage_rag_benchmark-0.1.0.1.dist-info/top_level.txt,sha256=hibFyzQHiLOMK68qL1OWsNKaXOmSXqZjeLTBem6Yy7I,5
+isage_rag_benchmark-0.1.0.1.dist-info/RECORD,,

isage_rag_benchmark-0.1.0.1.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (80.10.2)
+Root-Is-Purelib: true
+Tag: cp311-none-any
+

isage_rag_benchmark-0.1.0.1.dist-info/licenses/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 IntelliStream Team
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

isage_rag_benchmark-0.1.0.1.dist-info/top_level.txt

@@ -0,0 +1 @@
+sage

sage/benchmark_rag/__init__.py

@@ -0,0 +1,16 @@
+"""
+RAG Benchmarking Tools
+======================
+
+Tools for benchmarking and evaluating RAG (Retrieval-Augmented Generation) systems.
+
+This module provides:
+- **implementations/**: Various RAG implementation approaches (dense, sparse, hybrid, multimodal)
+- **evaluation/**: Pipeline experiments for batch processing and evaluation
+- **config/**: Configuration files for different RAG setups
+
+Test datasets and queries are located in the shared `data/qa/` directory.
+Evaluation metrics include: Accuracy, F1, Exact Match, and more.
+"""
+
+__all__: list[str] = []

sage/benchmark_rag/_version.py

@@ -0,0 +1,4 @@
+"""Version information for sage-rag-benchmark."""
+__version__ = "0.1.0.0"
+__author__ = "IntelliStream Team"
+__email__ = "shuhao_zhang@hust.edu.cn"

sage/benchmark_rag/config/config_bm25s.yaml

@@ -0,0 +1,51 @@
+pipeline:
+  name: "sage-api-operator-test"
+  description: "Test pipeline for Sage API Operator"
+  version: "1.0.0"
+  type: "remote"
+
+source:
+  data_path: "packages/sage-benchmark/src/sage/data/qa/sample/question.txt"
+
+retriever:
+  kv_collection:
+    bm25s: true
+    topk: 3
+  ltm:
+    topk: 3
+
+reranker:
+  model_name: "BAAI/bge-reranker-v2-m3"
+  top_k: 3
+
+promptor:
+
+refiner:
+  method: "openai"
+  model_name: "qwen-turbo-0919"
+  base_url: "http://127.0.0.1:8889/v1"
+  api_key: ""
+  seed: 42
+
+generator:
+  local:
+    method: "hf"
+    model_name: "meta-llama/Llama-2-13b-chat-hf"
+    seed: 42
+
+  vllm:
+    api_key: ""
+    method: "openai"
+    model_name: "meta-llama/Llama-2-13b-chat-hf"
+    base_url: "http://sage3:8000/v1"
+    seed: 42
+
+  remote:
+    api_key: ""
+    method: "openai"
+    model_name: "qwen-turbo-0919"
+    base_url: "http://127.0.0.1:8889/v1"
+    seed: 42
+
+sink:
+  file_path: "output.txt"

sage/benchmark_rag/config/config_dense_milvus.yaml

@@ -0,0 +1,61 @@
+# Milvus 稠密向量检索配置（适配 MilvusDenseRetriever）
+
+source:
+  data_path: "./packages/sage-benchmark/src/sage/data/qa/queries.jsonl"
+  platform: "local"
+
+retriever:
+  preload_knowledge_file:
+  - "./packages/sage-benchmark/src/sage/data/qa/qa_knowledge_chromaDB.txt"
+  - "./packages/sage-benchmark/src/sage/data/qa/qa_knowledge_rag.md"
+
+  # 通用参数
+  dimension: 384            # 向量维度，应与 embedding 模型一致
+  top_k: 2                 # 返回文档数量
+
+  # 嵌入模型配置（用于对文档和查询编码）
+  embedding:
+    method: "hf"
+    model: "sentence-transformers/all-MiniLM-L6-v2"
+
+  # Milvus 后端（稠密检索）
+  milvus_dense:
+    # 本地 Milvus Lite（推荐用于快速试用）
+    persistence_path: "./data/milvus_qa_dense.db"
+
+    # 远程 Milvus（如需远程，请注释上面的 persistence_path，改为如下配置）
+    # host: "127.0.0.1"
+    # port: 19530
+    # force_http: true
+
+    collection_name: "qa_dense_collection"
+    dim: 384
+    metric_type: "COSINE"   # 只允许: IP / COSINE / L2
+    search_type: "dense"     # 稠密检索
+
+    # 可选项
+    dense_insert_batch_size: 128
+    # 知识文件（可选）：提供后将自动按段落读取并入库
+    # knowledge_file: "./packages/sage-benchmark/src/sage/data/qa/qa_knowledge_base.txt"
+
+promptor:
+  template: |
+    基于以下检索到的相关文档，回答用户问题：
+
+    相关文档：
+    {retrieved_documents}
+
+    用户问题：{query}
+
+    请提供准确、有用的回答：
+
+generator:
+  vllm:
+    api_key: ""
+    method: "openai"
+    model_name: "meta-llama/Llama-2-7b-chat-hf"
+    base_url: "http://sage3:8000/v1"
+    seed: 42
+
+sink:
+  enable_log: true

sage/benchmark_rag/config/config_hf.yaml

@@ -0,0 +1,43 @@
+source:
+  data_path: "./packages/sage-benchmark/src/sage/data/qa/queries.jsonl"  # 修正相对路径
+  platform: "local"
+
+retriever:
+  # ChromaDB 专用配置
+  dimension: 384    # 修改为与 HuggingFace all-MiniLM-L6-v2 匹配
+  top_k: 2
+
+  chroma:
+    persistence_path: "./data/chroma_qa_database"
+    collection_name: "qa_knowledge_base"
+    use_embedding_query: true
+
+    # 知识库已预加载，注释掉自动加载
+    # knowledge_file: "../../data/qa_knowledge_base.txt"
+
+    # ChromaDB 元数据配置 (简化格式)
+    metadata:
+      hnsw:space: "cosine"           # 距离度量
+
+  # 嵌入模型配置
+  embedding:
+    method: "hf"
+    model: "sentence-transformers/all-MiniLM-L6-v2"
+
+promptor:
+  template: |
+    基于以下检索到的相关文档，回答用户问题：
+
+    相关文档：
+    {external_corpus}
+
+    请提供准确、有用的回答。
+
+generator:
+  local:
+    method: "hf"
+    model_name: "mistralai/Mistral-7B-Instruct-v0.1"
+    seed: 42
+
+sink:
+  enable_log: true

sage/benchmark_rag/config/config_mixed.yaml

@@ -0,0 +1,53 @@
+pipeline:
+  name: "sage-api-operator-operator_test"
+  description: "Test pipeline for Sage API Operator"
+  version: "1.0.0"
+  type: "remote"
+
+source:
+  # data_path: "sample/one_question.txt"
+  data_path: "packages/sage-benchmark/src/sage/data/qa/sample/question.txt"
+
+retriever:
+  platform: "remote"
+  ltm:
+    topk: 3
+
+reranker:
+  model_name: "BAAI/bge-reranker-v2-m3"
+  top_k: 3
+
+refiner:
+  method: "openai"
+  model_name: "qwen-turbo-0919"
+  base_url: "http://127.0.0.1:8889/v1"
+  api_key: ""
+  seed: 42
+
+generator:
+  local:
+    method: "hf"
+    model_name: "meta-llama/Llama-2-13b-chat-hf"
+    seed: 42
+
+  vllm:
+    api_key: ""
+    method: "openai"
+    model_name: "meta-llama/Llama-2-13b-chat-hf"
+    base_url: "http://sage3:8000/v1"
+    seed: 42
+
+  remote:
+    api_key: ""
+    method: "openai"
+    model_name: "qwen-turbo-0919"
+    base_url: "http://127.0.0.1:8889/v1"
+    seed: 42
+
+writer:
+
+promptor:
+  platform: "local"
+
+sink:
+  platform: "local"

sage/benchmark_rag/config/config_monitoring_demo.yaml

@@ -0,0 +1,59 @@
+# Performance Monitoring Demo 配置文件
+# 简化的 RAG Pipeline: Retrieval -> Prompt -> Generation
+# 启用性能监控功能,收集 TPS/延迟/资源使用等指标
+
+
+source:
+  data_path: "packages/sage-benchmark/src/sage/data/qa/queries.jsonl"
+  platform: "local"
+
+retriever:
+  # ChromaDB 向量检索配置
+  dimension: 384    # 使用 HuggingFace all-MiniLM-L6-v2 embedding 模型
+  top_k: 3          # 检索 top-3 相关文档
+
+  chroma:
+    persistence_path: "./data/chroma_qa_database"
+    collection_name: "qa_knowledge_base"
+    use_embedding_query: true
+
+    # ChromaDB 元数据配置
+    metadata:
+      hnsw:space: "cosine"           # 余弦相似度
+
+  # 嵌入模型配置
+  embedding:
+    method: "hf"
+    model: "sentence-transformers/all-MiniLM-L6-v2"
+
+promptor:
+  template: |
+    基于以下检索到的相关文档，回答用户问题：
+
+    相关文档：
+    {retrieved_documents}
+
+    用户问题：{query}
+
+    请提供准确、有用的回答：
+
+generator:
+  vllm:
+    api_key: ""
+    method: "openai"
+    model_name: "meta-llama/Llama-2-7b-chat-hf"
+    base_url: "http://sage3:8000/v1"
+    seed: 42
+    max_tokens: 256    # 限制生成长度以加快测试
+    temperature: 0.7
+
+sink:
+  enable_log: true
+
+# 监控配置 (可选,如果需要自定义监控参数)
+monitoring:
+  enabled: true                    # 启用监控
+  metrics_window_size: 1000        # 滑动窗口大小
+  resource_monitoring: true        # 启用 CPU/内存监控
+  sample_interval: 1.0             # 资源采样间隔(秒)
+  auto_report: false               # 不自动打印报告(由主程序控制)

sage/benchmark_rag/config/config_multiplex.yaml

@@ -0,0 +1,79 @@
+pipeline:
+  name: "sage-api-operator-operator_test"
+  description: "Test pipeline for Sage API Operator"
+  version: "1.0.0"
+
+source:
+  # data_path: "sample/one_question.txt"
+  data_path: "packages/sage-benchmark/src/sage/data/qa/sample/question.txt"
+  platform: "local"
+
+retriever:
+  # ChromaDB 专用配置
+  dimension: 384    # 修改为与 HuggingFace all-MiniLM-L6-v2 匹配
+  top_k: 2
+
+  chroma:
+    persistence_path: "./data/chroma_qa_database"
+    collection_name: "qa_knowledge_base"
+    use_embedding_query: true
+
+    # 知识库已预加载，注释掉自动加载
+    # knowledge_file: "../../data/qa_knowledge_base.txt"
+
+    # ChromaDB 元数据配置 (简化格式)
+    metadata:
+      hnsw:space: "cosine"           # 距离度量
+
+  # 嵌入模型配置
+  embedding:
+    method: "hf"
+    model: "sentence-transformers/all-MiniLM-L6-v2"
+
+reranker:
+  platform: "local"
+  model_name: "BAAI/bge-reranker-v2-m3"
+  top_k: 3
+
+refiner:
+  platform: "local"
+  method: "openai"
+  model_name: "qwen-turbo-0919"
+  base_url: "http://127.0.0.1:8889/v1"
+  api_key: ""
+  seed: 42
+
+generator:
+  local:
+    method: "hf"
+    model_name: "meta-llama/Llama-2-13b-chat-hf"
+    seed: 42
+
+  vllm:
+    api_key: ""
+    method: "openai"
+    model_name: "meta-llama/Llama-2-13b-chat-hf"
+    base_url: "http://sage3:8000/v1"
+    seed: 42
+
+  remote:
+    api_key: ""
+    method: "openai"
+    model_name: "qwen-turbo-0919"
+    base_url: "http://127.0.0.1:8889/v1"
+    seed: 42
+
+writer:
+
+promptor:
+  platform: "local"
+
+sink_true:
+  file_path: "output_true.txt"
+
+
+sink_false:
+  file_path: "output_false.txt"
+
+
+sink_terminal:

sage/benchmark_rag/config/config_qa_chroma.yaml

@@ -0,0 +1,51 @@
+# ChromaDB 专用 QA 配置文件
+# 适用于 qa_openai.py 的简化配置
+
+
+source:
+  data_path: "./packages/sage-benchmark/src/sage/data/qa/queries.jsonl"  # 修正相对路径
+  platform: "local"
+
+retriever:
+  # ChromaDB 专用配置
+  dimension: 384    # 修改为与 HuggingFace all-MiniLM-L6-v2 匹配
+  top_k: 2
+
+  chroma:
+    persistence_path: "./data/chroma_qa_database"
+    collection_name: "qa_knowledge_base"
+    use_embedding_query: true
+
+    # 知识库已预加载，注释掉自动加载
+    # knowledge_file: "../../data/qa_knowledge_base.txt"
+
+    # ChromaDB 元数据配置 (简化格式)
+    metadata:
+      hnsw:space: "cosine"           # 距离度量
+
+  # 嵌入模型配置
+  embedding:
+    method: "hf"
+    model: "sentence-transformers/all-MiniLM-L6-v2"
+
+promptor:
+  template: |
+    基于以下检索到的相关文档，回答用户问题：
+
+    相关文档：
+    {retrieved_documents}
+
+    用户问题：{query}
+
+    请提供准确、有用的回答：
+
+generator:
+  vllm:
+    api_key: ""
+    method: "openai"
+    model_name: "meta-llama/Llama-2-7b-chat-hf"
+    base_url: "http://sage3:8000/v1"
+    seed: 42
+
+sink:
+  enable_log: true

sage/benchmark_rag/config/config_ray.yaml

@@ -0,0 +1,57 @@
+pipeline:
+  name: "sage-api-operator-operator_test"
+  description: "Test pipeline for Sage API Operator"
+  version: "1.0.0"
+
+
+
+source:
+  # data_path: "sample/one_question.txt"
+  data_path: "packages/sage-benchmark/src/sage/data/qa/sample/question.txt"
+  platform: "remote"
+
+retriever:
+  platform: "remote"
+  ltm:
+    topk: 3
+
+reranker:
+  platform: "remote"
+  model_name: "BAAI/bge-reranker-v2-m3"
+  top_k: 3
+
+refiner:
+  platform: "remote"
+  method: "openai"
+  model_name: "qwen-turbo"
+  base_url: "http://127.0.0.1:8889/v1"
+  api_key: ""
+  seed: 42
+
+generator:
+  local:
+    method: "hf"
+    model_name: "meta-llama/Llama-2-13b-chat-hf"
+    seed: 42
+
+  vllm:
+    api_key: ""
+    method: "openai"
+    model_name: "meta-llama/Llama-2-13b-chat-hf"
+    base_url: "http://sage3:8000/v1"
+    seed: 42
+
+  remote:
+    api_key: ""
+    method: "openai"
+    model_name: "qwen-turbo-0919"
+    base_url: "http://127.0.0.1:8889/v1"
+    seed: 42
+
+writer:
+
+promptor:
+  platform: "remote"
+
+sink:
+  platform: "remote"