AutoRAG 0.3.12__tar.gz → 0.3.13__tar.gz

{autorag-0.3.12 → autorag-0.3.13}/.github/workflows/sphinx.yml

@@ -29,12 +29,12 @@ jobs:
       run: |
         cp docs/source/CNAME docs/build/html/CNAME
     - name: Upload artifacts
-      uses: actions/upload-artifact@v3
+      uses: actions/upload-artifact@v4
       with:
         name: html-docs
         path: docs/build/html/
     - name: Deploy
-      uses: peaceiris/actions-gh-pages@v3
+      uses: peaceiris/actions-gh-pages@v4
       if: github.ref == 'refs/heads/main'
       with:
         github_token: ${{ secrets.GITHUB_TOKEN }}

{autorag-0.3.12 → autorag-0.3.13}/AutoRAG.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.2
 Name: AutoRAG
-Version: 0.3.12
+Version: 0.3.13
 Summary: Automatically Evaluate RAG pipelines with your own data. Find optimal structure for new RAG product.
 Author-email: Marker-Inc <vkehfdl1@gmail.com>
 License:                                  Apache License
@@ -254,6 +254,7 @@ Requires-Dist: llama-index>=0.11.0
 Requires-Dist: llama-index-core>=0.11.0
 Requires-Dist: llama-index-readers-file
 Requires-Dist: llama-index-embeddings-openai
+Requires-Dist: llama-index-embeddings-ollama
 Requires-Dist: llama-index-llms-openai>=0.2.7
 Requires-Dist: llama-index-llms-openai-like
 Requires-Dist: llama-index-retrievers-bm25
@@ -750,7 +751,7 @@ docker run --rm -it \
   -v ~/.cache/huggingface:/root/.cache/huggingface \
   -v $(pwd)/projects:/usr/src/app/projects \
   -e OPENAI_API_KEY=${OPENAI_API_KEY} \
-  autoraghq/autorag:api evaluate \
+  autoraghq/autorag:api-latest evaluate \
   --config /usr/src/app/projects/tutorial_1/config.yaml \
   --qa_data_path /usr/src/app/projects/tutorial_1/qa_test.parquet \
   --corpus_data_path /usr/src/app/projects/tutorial_1/corpus.parquet \
@@ -764,7 +765,7 @@ docker run --rm -it \
   -v ~/.cache/huggingface:/root/.cache/huggingface \
   -v $(pwd)/projects:/usr/src/app/projects \
   -e OPENAI_API_KEY=${OPENAI_API_KEY} \
-  autoraghq/autorag:api validate \
+  autoraghq/autorag:api-latest validate \
   --config /usr/src/app/projects/tutorial_1/config.yaml \
   --qa_data_path /usr/src/app/projects/tutorial_1/qa_test.parquet \
   --corpus_data_path /usr/src/app/projects/tutorial_1/corpus.parquet
@@ -778,7 +779,7 @@ docker run --rm -it \
   -v $(pwd)/projects:/usr/src/app/projects \
   -e OPENAI_API_KEY=${OPENAI_API_KEY} \
   -p 8502:8502 \
-  autoraghq/autorag:api dashboard \
+  autoraghq/autorag:api-latest dashboard \
     --trial_dir /usr/src/app/projects/tutorial_1/0
 ```
 
@@ -790,7 +791,7 @@ docker run --rm -it \
   -v $(pwd)/projects:/usr/src/app/projects \
   -e OPENAI_API_KEY=${OPENAI_API_KEY} \
   -p 8501:8501 \
-  autoraghq/autorag:api run_web --trial_path ./projects/tutorial_1/0
+  autoraghq/autorag:api-latest run_web --trial_path ./projects/tutorial_1/0
 ```
 
 #### Key Points :

{autorag-0.3.12 → autorag-0.3.13}/AutoRAG.egg-info/SOURCES.txt

@@ -1,6 +1,7 @@
 .dockerignore
 .gitignore
 .pre-commit-config.yaml
+CNAME
 CODE_OF_CONDUCT.md
 CONTRIBUTING.md
 Dockerfile.base
@@ -89,6 +90,8 @@ autorag/deploy/api.py
 autorag/deploy/base.py
 autorag/deploy/gradio.py
 autorag/deploy/swagger.yml
+autorag/embedding/__init__.py
+autorag/embedding/base.py
 autorag/evaluation/__init__.py
 autorag/evaluation/generation.py
 autorag/evaluation/retrieval.py
@@ -112,6 +115,7 @@ autorag/nodes/generator/llama_index_llm.py
 autorag/nodes/generator/openai_llm.py
 autorag/nodes/generator/run.py
 autorag/nodes/generator/vllm.py
+autorag/nodes/generator/vllm_api.py
 autorag/nodes/passageaugmenter/__init__.py
 autorag/nodes/passageaugmenter/base.py
 autorag/nodes/passageaugmenter/pass_passage_augmenter.py
@@ -192,6 +196,7 @@ autorag/vectordb/milvus.py
 autorag/vectordb/pinecone.py
 autorag/vectordb/qdrant.py
 autorag/vectordb/weaviate.py
+docs/CNAME
 docs/Makefile
 docs/make.bat
 docs/requirements.txt
@@ -325,6 +330,7 @@ docs/source/nodes/generator/generator.md
 docs/source/nodes/generator/llama_index_llm.md
 docs/source/nodes/generator/openai_llm.md
 docs/source/nodes/generator/vllm.md
+docs/source/nodes/generator/vllm_api.md
 docs/source/nodes/passage_augmenter/passage_augmenter.md
 docs/source/nodes/passage_augmenter/prev_next_augmenter.md
 docs/source/nodes/passage_compressor/longllmlingua.md
@@ -461,6 +467,7 @@ tests/autorag/data/qa/generation_gt/test_openai_gen_gt.py
 tests/autorag/data/qa/query/base_test_query_gen.py
 tests/autorag/data/qa/query/test_llama_gen_query.py
 tests/autorag/data/qa/query/test_openai_gen_query.py
+tests/autorag/embedding/test_base.py
 tests/autorag/evaluate/test_evaluate_util.py
 tests/autorag/evaluate/test_generation_evaluate.py
 tests/autorag/evaluate/test_retrieval_contents_evaluate.py

{autorag-0.3.12 → autorag-0.3.13}/AutoRAG.egg-info/requires.txt

@@ -33,6 +33,7 @@ llama-index>=0.11.0
 llama-index-core>=0.11.0
 llama-index-readers-file
 llama-index-embeddings-openai
+llama-index-embeddings-ollama
 llama-index-llms-openai>=0.2.7
 llama-index-llms-openai-like
 llama-index-retrievers-bm25

{autorag-0.3.12 → autorag-0.3.13}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.2
 Name: AutoRAG
-Version: 0.3.12
+Version: 0.3.13
 Summary: Automatically Evaluate RAG pipelines with your own data. Find optimal structure for new RAG product.
 Author-email: Marker-Inc <vkehfdl1@gmail.com>
 License:                                  Apache License
@@ -254,6 +254,7 @@ Requires-Dist: llama-index>=0.11.0
 Requires-Dist: llama-index-core>=0.11.0
 Requires-Dist: llama-index-readers-file
 Requires-Dist: llama-index-embeddings-openai
+Requires-Dist: llama-index-embeddings-ollama
 Requires-Dist: llama-index-llms-openai>=0.2.7
 Requires-Dist: llama-index-llms-openai-like
 Requires-Dist: llama-index-retrievers-bm25
@@ -750,7 +751,7 @@ docker run --rm -it \
   -v ~/.cache/huggingface:/root/.cache/huggingface \
   -v $(pwd)/projects:/usr/src/app/projects \
   -e OPENAI_API_KEY=${OPENAI_API_KEY} \
-  autoraghq/autorag:api evaluate \
+  autoraghq/autorag:api-latest evaluate \
   --config /usr/src/app/projects/tutorial_1/config.yaml \
   --qa_data_path /usr/src/app/projects/tutorial_1/qa_test.parquet \
   --corpus_data_path /usr/src/app/projects/tutorial_1/corpus.parquet \
@@ -764,7 +765,7 @@ docker run --rm -it \
   -v ~/.cache/huggingface:/root/.cache/huggingface \
   -v $(pwd)/projects:/usr/src/app/projects \
   -e OPENAI_API_KEY=${OPENAI_API_KEY} \
-  autoraghq/autorag:api validate \
+  autoraghq/autorag:api-latest validate \
   --config /usr/src/app/projects/tutorial_1/config.yaml \
   --qa_data_path /usr/src/app/projects/tutorial_1/qa_test.parquet \
   --corpus_data_path /usr/src/app/projects/tutorial_1/corpus.parquet
@@ -778,7 +779,7 @@ docker run --rm -it \
   -v $(pwd)/projects:/usr/src/app/projects \
   -e OPENAI_API_KEY=${OPENAI_API_KEY} \
   -p 8502:8502 \
-  autoraghq/autorag:api dashboard \
+  autoraghq/autorag:api-latest dashboard \
     --trial_dir /usr/src/app/projects/tutorial_1/0
 ```
 
@@ -790,7 +791,7 @@ docker run --rm -it \
   -v $(pwd)/projects:/usr/src/app/projects \
   -e OPENAI_API_KEY=${OPENAI_API_KEY} \
   -p 8501:8501 \
-  autoraghq/autorag:api run_web --trial_path ./projects/tutorial_1/0
+  autoraghq/autorag:api-latest run_web --trial_path ./projects/tutorial_1/0
 ```
 
 #### Key Points :

{autorag-0.3.12 → autorag-0.3.13}/README.md

@@ -441,7 +441,7 @@ docker run --rm -it \
   -v ~/.cache/huggingface:/root/.cache/huggingface \
   -v $(pwd)/projects:/usr/src/app/projects \
   -e OPENAI_API_KEY=${OPENAI_API_KEY} \
-  autoraghq/autorag:api evaluate \
+  autoraghq/autorag:api-latest evaluate \
   --config /usr/src/app/projects/tutorial_1/config.yaml \
   --qa_data_path /usr/src/app/projects/tutorial_1/qa_test.parquet \
   --corpus_data_path /usr/src/app/projects/tutorial_1/corpus.parquet \
@@ -455,7 +455,7 @@ docker run --rm -it \
   -v ~/.cache/huggingface:/root/.cache/huggingface \
   -v $(pwd)/projects:/usr/src/app/projects \
   -e OPENAI_API_KEY=${OPENAI_API_KEY} \
-  autoraghq/autorag:api validate \
+  autoraghq/autorag:api-latest validate \
   --config /usr/src/app/projects/tutorial_1/config.yaml \
   --qa_data_path /usr/src/app/projects/tutorial_1/qa_test.parquet \
   --corpus_data_path /usr/src/app/projects/tutorial_1/corpus.parquet
@@ -469,7 +469,7 @@ docker run --rm -it \
   -v $(pwd)/projects:/usr/src/app/projects \
   -e OPENAI_API_KEY=${OPENAI_API_KEY} \
   -p 8502:8502 \
-  autoraghq/autorag:api dashboard \
+  autoraghq/autorag:api-latest dashboard \
     --trial_dir /usr/src/app/projects/tutorial_1/0
 ```
 
@@ -481,7 +481,7 @@ docker run --rm -it \
   -v $(pwd)/projects:/usr/src/app/projects \
   -e OPENAI_API_KEY=${OPENAI_API_KEY} \
   -p 8501:8501 \
-  autoraghq/autorag:api run_web --trial_path ./projects/tutorial_1/0
+  autoraghq/autorag:api-latest run_web --trial_path ./projects/tutorial_1/0
 ```
 
 #### Key Points :

autorag-0.3.13/autorag/VERSION

@@ -0,0 +1 @@
+0.3.13

{autorag-0.3.12 → autorag-0.3.13}/autorag/__init__.py

@@ -41,13 +41,6 @@ class LazyInit:
 		return getattr(self._instance, name)
 
 
-class MockEmbeddingRandom(MockEmbedding):
-	"""Mock embedding with random vectors."""
-
-	def _get_vector(self) -> List[float]:
-		return [random() for _ in range(self.embed_dim)]
-
-
 rich_format = "[%(filename)s:%(lineno)s] >> %(message)s"
 logging.basicConfig(
 	level="INFO", format=rich_format, handlers=[RichHandler(rich_tracebacks=True)]
@@ -62,46 +55,6 @@ def handle_exception(exc_type, exc_value, exc_traceback):
 
 sys.excepthook = handle_exception
 
-embedding_models = {
-	# llama index
-	"openai": LazyInit(
-		OpenAIEmbedding
-	),  # default model is OpenAIEmbeddingModelType.TEXT_EMBED_ADA_002
-	"openai_embed_3_large": LazyInit(
-		OpenAIEmbedding, model_name=OpenAIEmbeddingModelType.TEXT_EMBED_3_LARGE
-	),
-	"openai_embed_3_small": LazyInit(
-		OpenAIEmbedding, model_name=OpenAIEmbeddingModelType.TEXT_EMBED_3_SMALL
-	),
-	"mock": LazyInit(MockEmbeddingRandom, embed_dim=768),
-	# langchain
-	"openai_langchain": LazyInit(OpenAIEmbeddings),
-}
-
-try:
-	# you can use your own model in this way.
-	from llama_index.embeddings.huggingface import HuggingFaceEmbedding
-
-	embedding_models["huggingface_baai_bge_small"] = LazyInit(
-		HuggingFaceEmbedding, model_name="BAAI/bge-small-en-v1.5"
-	)
-	embedding_models["huggingface_cointegrated_rubert_tiny2"] = LazyInit(
-		HuggingFaceEmbedding, model_name="cointegrated/rubert-tiny2"
-	)
-	embedding_models["huggingface_all_mpnet_base_v2"] = LazyInit(
-		HuggingFaceEmbedding,
-		model_name="sentence-transformers/all-mpnet-base-v2",
-		max_length=512,
-	)
-	embedding_models["huggingface_bge_m3"] = LazyInit(
-		HuggingFaceEmbedding, model_name="BAAI/bge-m3"
-	)
-except ImportError:
-	logger.info(
-		"You are using API version of AutoRAG."
-		"To use local version, run pip install 'AutoRAG[gpu]'"
-	)
-
 
 class AutoRAGBedrock(Bedrock):
 	async def acomplete(

{autorag-0.3.12 → autorag-0.3.13}/autorag/data/chunk/base.py

@@ -4,7 +4,7 @@ from typing import Tuple, List, Dict, Any
 
 import pandas as pd
 
-from autorag import embedding_models
+from autorag.embedding.base import EmbeddingModel
 from autorag.data import chunk_modules, sentence_splitter_modules
 from autorag.utils import result_to_dataframe
 
@@ -80,7 +80,7 @@ def __get_chunk_instance(module_type: str, chunk_method: str, **kwargs):
 		if _embed_model_str == "openai":
 			if _module_type == "langchain_chunk":
 				_embed_model_str = "openai_langchain"
-		return embedding_models[_embed_model_str]()
+		return EmbeddingModel.load(_embed_model_str)()
 
 	# Add embed_model to kwargs
 	embedding_available_methods = ["semantic_llama_index", "semantic_langchain"]

{autorag-0.3.12 → autorag-0.3.13}/autorag/data/parse/run.py

@@ -107,7 +107,16 @@ def run_parser(
 				module_params,
 			)
 		)
-	list(map(lambda x: x[0].to_parquet(x[1], index=False), zip(results, filepaths)))
+
+	_files = {}
+	for result, filepath in zip(results, filepaths):
+		_files[filepath].append(result) if filepath in _files.keys() else _files.update(
+			{filepath: [result]}
+		)
+	# Save files with a specific file type as Parquet files.
+	for filepath, value in _files.items():
+		pd.concat(value).to_parquet(filepath, index=False)
+
 	filenames = list(map(lambda x: os.path.basename(x), filepaths))
 
 	summary_df = pd.DataFrame(
@@ -121,8 +130,9 @@ def run_parser(
 	summary_df.to_csv(os.path.join(project_dir, "summary.csv"), index=False)
 
 	# concat all parquet files here if not all_files.
+	_filepaths = list(_files.keys())
 	if not all_files:
-		dataframes = [pd.read_parquet(file) for file in filepaths]
+		dataframes = [pd.read_parquet(file) for file in _filepaths]
 		combined_df = pd.concat(dataframes, ignore_index=True)
 		combined_df.to_parquet(
 			os.path.join(project_dir, "parsed_result.parquet"), index=False

{autorag-0.3.12 → autorag-0.3.13}/autorag/deploy/api.py

@@ -30,6 +30,7 @@ class QueryRequest(BaseModel):
 class RetrievedPassage(BaseModel):
 	content: str
 	doc_id: str
+	score: float
 	filepath: Optional[str] = None
 	file_page: Optional[int] = None
 	start_idx: Optional[int] = None
@@ -41,14 +42,8 @@ class RunResponse(BaseModel):
 	retrieved_passage: List[RetrievedPassage]
 
 
-class Passage(BaseModel):
-	doc_id: str
-	content: str
-	score: float
-
-
 class RetrievalResponse(BaseModel):
-	passages: List[Passage]
+	passages: List[RetrievedPassage]
 
 
 class StreamResponse(BaseModel):
@@ -153,18 +148,9 @@ class ApiRunner(BaseRunner):
 				previous_result = pd.concat([drop_previous_result, new_result], axis=1)
 
 			# Simulate processing the query
-			retrieved_contents = previous_result["retrieved_contents"].tolist()[0]
-			retrieved_ids = previous_result["retrieved_ids"].tolist()[0]
-			retrieve_scores = previous_result["retrieve_scores"].tolist()[0]
-
-			retrieval_response = RetrievalResponse(
-				passages=[
-					Passage(doc_id=doc_id, content=content, score=score)
-					for doc_id, content, score in zip(
-						retrieved_ids, retrieved_contents, retrieve_scores
-					)
-				]
-			)
+			retrieved_passages = self.extract_retrieve_passage(previous_result)
+
+			retrieval_response = RetrievalResponse(passages=retrieved_passages)
 			return jsonify(retrieval_response.model_dump()), 200
 
 		@self.app.route("/v1/stream", methods=["POST"])
@@ -264,6 +250,7 @@ class ApiRunner(BaseRunner):
 	def extract_retrieve_passage(self, df: pd.DataFrame) -> List[RetrievedPassage]:
 		retrieved_ids: List[str] = df["retrieved_ids"].tolist()[0]
 		contents = fetch_contents(self.corpus_df, [retrieved_ids])[0]
+		scores = df["retrieve_scores"].tolist()[0]
 		if "path" in self.corpus_df.columns:
 			paths = fetch_contents(self.corpus_df, [retrieved_ids], column_name="path")[
 				0
@@ -282,9 +269,15 @@ class ApiRunner(BaseRunner):
 		start_end_indices = to_list(start_end_indices)
 		return list(
 			map(
-				lambda content, doc_id, path, metadata, start_end_idx: RetrievedPassage(
+				lambda content,
+				doc_id,
+				score,
+				path,
+				metadata,
+				start_end_idx: RetrievedPassage(
 					content=content,
 					doc_id=doc_id,
+					score=score,
 					filepath=path,
 					file_page=metadata.get("page", None),
 					start_idx=start_end_idx[0] if start_end_idx else None,
@@ -292,6 +285,7 @@ class ApiRunner(BaseRunner):
 				),
 				contents,
 				retrieved_ids,
+				scores,
 				paths,
 				metadatas,
 				start_end_indices,

autorag-0.3.13/autorag/embedding/base.py

@@ -0,0 +1,131 @@
+import logging
+import sys
+
+from random import random
+from typing import List, Union, Dict
+
+from llama_index.core.embeddings.mock_embed_model import MockEmbedding
+from llama_index.embeddings.openai import OpenAIEmbedding
+from llama_index.embeddings.openai import OpenAIEmbeddingModelType
+from llama_index.embeddings.ollama import OllamaEmbedding
+from langchain_openai.embeddings import OpenAIEmbeddings
+
+from autorag import LazyInit
+
+logger = logging.getLogger("AutoRAG")
+
+
+class MockEmbeddingRandom(MockEmbedding):
+	"""Mock embedding with random vectors."""
+
+	def _get_vector(self) -> List[float]:
+		return [random() for _ in range(self.embed_dim)]
+
+
+embedding_models = {
+	# llama index
+	"openai": LazyInit(
+		OpenAIEmbedding
+	),  # default model is OpenAIEmbeddingModelType.TEXT_EMBED_ADA_002
+	"openai_embed_3_large": LazyInit(
+		OpenAIEmbedding, model_name=OpenAIEmbeddingModelType.TEXT_EMBED_3_LARGE
+	),
+	"openai_embed_3_small": LazyInit(
+		OpenAIEmbedding, model_name=OpenAIEmbeddingModelType.TEXT_EMBED_3_SMALL
+	),
+	"mock": LazyInit(MockEmbeddingRandom, embed_dim=768),
+	# langchain
+	"openai_langchain": LazyInit(OpenAIEmbeddings),
+	"ollama": LazyInit(OllamaEmbedding),
+}
+
+try:
+	# you can use your own model in this way.
+	from llama_index.embeddings.huggingface import HuggingFaceEmbedding
+
+	embedding_models["huggingface_baai_bge_small"] = LazyInit(
+		HuggingFaceEmbedding, model_name="BAAI/bge-small-en-v1.5"
+	)
+	embedding_models["huggingface_cointegrated_rubert_tiny2"] = LazyInit(
+		HuggingFaceEmbedding, model_name="cointegrated/rubert-tiny2"
+	)
+	embedding_models["huggingface_all_mpnet_base_v2"] = LazyInit(
+		HuggingFaceEmbedding,
+		model_name="sentence-transformers/all-mpnet-base-v2",
+		max_length=512,
+	)
+	embedding_models["huggingface_bge_m3"] = LazyInit(
+		HuggingFaceEmbedding, model_name="BAAI/bge-m3"
+	)
+	embedding_models["huggingface_multilingual_e5_large"] = LazyInit(
+		HuggingFaceEmbedding, model_name="intfloat/multilingual-e5-large-instruct"
+	)
+except ImportError:
+	logger.info(
+		"You are using API version of AutoRAG."
+		"To use local version, run pip install 'AutoRAG[gpu]'"
+	)
+
+
+class EmbeddingModel:
+	@staticmethod
+	def load(config: Union[str, Dict, List[Dict]]):
+		if isinstance(config, str):
+			return EmbeddingModel.load_from_str(config)
+		elif isinstance(config, dict):
+			return EmbeddingModel.load_from_dict(config)
+		elif isinstance(config, list):
+			return EmbeddingModel.load_from_list(config)
+		else:
+			raise ValueError("Invalid type of config")
+
+	@staticmethod
+	def load_from_str(name: str):
+		try:
+			return embedding_models[name]
+		except KeyError:
+			raise ValueError(f"Embedding model '{name}' is not supported")
+
+	@staticmethod
+	def load_from_list(option: List[dict]):
+		if len(option) != 1:
+			raise ValueError("Only one embedding model is supported")
+		return EmbeddingModel.load_from_dict(option[0])
+
+	@staticmethod
+	def load_from_dict(option: dict):
+		def _check_keys(target: dict):
+			if "type" not in target or "model_name" not in target:
+				raise ValueError("Both 'type' and 'model_name' must be provided")
+			if target["type"] not in ["openai", "huggingface", "mock", "ollama"]:
+				raise ValueError(
+					f"Embedding model type '{target['type']}' is not supported"
+				)
+
+		def _get_huggingface_class():
+			module = sys.modules.get("llama_index.embeddings.huggingface")
+			if not module:
+				logger.info(
+					"You are using API version of AutoRAG. "
+					"To use local version, run `pip install 'AutoRAG[gpu]'`."
+				)
+				return None
+			return getattr(module, "HuggingFaceEmbedding", None)
+
+		_check_keys(option)
+
+		model_options = option
+		model_type = model_options.pop("type")
+
+		embedding_map = {
+			"openai": OpenAIEmbedding,
+			"mock": MockEmbeddingRandom,
+			"huggingface": _get_huggingface_class(),
+			"ollama": OllamaEmbedding,
+		}
+
+		embedding_class = embedding_map.get(model_type)
+		if not embedding_class:
+			raise ValueError(f"Embedding model type '{model_type}' is not supported")
+
+		return LazyInit(embedding_class, **model_options)

{autorag-0.3.12 → autorag-0.3.13}/autorag/evaluation/metric/generation.py

@@ -14,7 +14,7 @@ from rouge_score import tokenizers
 from rouge_score.rouge_scorer import RougeScorer
 from sacrebleu.metrics.bleu import BLEU
 
-from autorag import embedding_models
+from autorag.embedding.base import embedding_models
 from autorag.evaluation.metric.deepeval_prompt import FaithfulnessTemplate
 from autorag.evaluation.metric.util import (
 	autorag_metric_loop,

{autorag-0.3.12 → autorag-0.3.13}/autorag/evaluation/util.py

@@ -1,7 +1,7 @@
 from copy import deepcopy
 from typing import Union, List, Dict, Tuple, Any
 
-from autorag import embedding_models
+from autorag.embedding.base import EmbeddingModel
 
 
 def cast_metrics(
@@ -38,6 +38,6 @@ def cast_metrics(
 
 def cast_embedding_model(key, value):
 	if key == "embedding_model":
-		return key, embedding_models[value]()
+		return key, EmbeddingModel.load(value)()
 	else:
 		return key, value

{autorag-0.3.12 → autorag-0.3.13}/autorag/nodes/generator/__init__.py

@@ -1,3 +1,4 @@
 from .llama_index_llm import LlamaIndexLLM
 from .openai_llm import OpenAILLM
 from .vllm import Vllm
+from .vllm_api import VllmAPI

{autorag-0.3.12 → autorag-0.3.13}/autorag/nodes/generator/vllm.py

@@ -41,7 +41,8 @@ class Vllm(BaseGenerator):
 
 			if torch.cuda.is_available():
 				from vllm.distributed.parallel_state import (
-					destroy_model_parallel, destroy_distributed_environment
+					destroy_model_parallel,
+					destroy_distributed_environment,
 				)
 
 				destroy_model_parallel()