Skip to content

Bagel

BagelVectorStore #

Bases: BasePydanticVectorStore

Vector store for Bagel.

示例

pip install llama-index-vector-stores-bagel

from llama_index.core import VectorStoreIndex, StorageContext
from llama_index.vector_stores.bagel import BagelVectorStore

import bagel
from bagel import Settings

server_settings = Settings(
    bagel_api_impl="rest", bagel_server_host="api.bageldb.ai"
)

client = bagel.Client(server_settings)

collection = client.get_or_create_cluster("testing_embeddings")
vector_store = BagelVectorStore(collection=collection)
Source code in llama_index/vector_stores/bagel/base.py 
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
class BagelVectorStore(BasePydanticVectorStore):
    """Vector store for Bagel.

    示例:
        `pip install llama-index-vector-stores-bagel`

        ```python
        from llama_index.core import VectorStoreIndex, StorageContext
        from llama_index.vector_stores.bagel import BagelVectorStore

        import bagel
        from bagel import Settings

        server_settings = Settings(
            bagel_api_impl="rest", bagel_server_host="api.bageldb.ai"
        )

        client = bagel.Client(server_settings)

        collection = client.get_or_create_cluster("testing_embeddings")
        vector_store = BagelVectorStore(collection=collection)
        ```"""

    # support for Bagel specific parameters
    stores_text: bool = True
    flat_metadata: bool = True

    _collection: Any = PrivateAttr()

    def __init__(self, collection: Any, **kwargs: Any) -> None:
        """初始化BagelVectorStore。

Args:
    collection: Bagel集合。
    **kwargs: 额外的参数。
"""
        super().__init__()

        try:
            from bagel.api.Cluster import Cluster
        except ImportError:
            raise ImportError("Bagel is not installed. Please install bagel.")

        if not isinstance(collection, Cluster):
            raise ValueError("Collection must be a bagel Cluster.")

        self._collection = collection

    @classmethod
    def class_name(cls) -> str:
        return "BagelVectorStore"

    def add(self, nodes: List[BaseNode], **add_kwargs: Any) -> List[str]:
        """将带有嵌入的节点列表添加到向量存储中。

Args:
    nodes：带有嵌入的节点列表。
    kwargs：额外的参数。

Returns:
    文档id的列表。
"""
        if not self._collection:
            raise ValueError("collection not set")

        ids = []
        embeddings = []
        metadatas = []
        documents = []

        for node in nodes:
            ids.append(node.node_id)
            embeddings.append(node.get_embedding())
            metadatas.append(
                node_to_metadata_dict(
                    node,
                    remove_text=True,
                    flat_metadata=self.flat_metadata,
                )
            )
            documents.append(node.get_content(metadata_mode=MetadataMode.NONE) or "")

        self._collection.add(
            ids=ids, embeddings=embeddings, metadatas=metadatas, documents=documents
        )

        return ids

    def delete(self, ref_doc_id: str, **kwargs: Any) -> None:
        """从向量存储中删除文档。

Args:
    ref_doc_id：参考文档的id。
    kwargs：额外的参数。
"""
        if not self._collection:
            raise ValueError("collection not set")

        results = self._collection.get(where={"doc_id": ref_doc_id})
        if results and "ids" in results:
            self._collection.delete(ids=results["ids"])

    @property
    def client(self) -> Any:
        """
        获取Bagel集群。
        """
        return self._collection

    def query(self, query: VectorStoreQuery, **kwargs: Any) -> VectorStoreQueryResult:
        """查询向量存储。

Args:
    query: 要运行的查询。
    kwargs: 附加参数。

Returns:
    查询结果。
"""
        if not self._collection:
            raise ValueError("collection not set")

        if query.filters is not None:
            if "where" in kwargs:
                raise ValueError("Cannot specify both filters and where")
            where = _to_bagel_filter(query.filters)
        else:
            where = kwargs.get("where", {})

        results = self._collection.find(
            query_embeddings=query.query_embedding,
            where=where,
            n_results=query.similarity_top_k,
            **kwargs,
        )

        logger.debug(f"query results: {results}")

        nodes = []
        similarities = []
        ids = []

        for node_id, text, metadata, distance in zip(
            results["ids"][0],
            results["documents"][0],
            results["metadatas"][0],
            results["distances"][0],
        ):
            try:
                node = metadata_dict_to_node(metadata)
                node.set_content(text)
            except Exception:
                # NOTE: deprecated legacy logic for backward compatibility
                metadata, node_info, relationships = legacy_metadata_dict_to_node(
                    metadata
                )

                node = TextNode(
                    text=text,
                    id_=node_id,
                    metadata=metadata,
                    start_char_idx=node_info.get("start", None),
                    end_char_idx=node_info.get("end", None),
                    relationships=relationships,
                )

            nodes.append(node)
            similarities.append(1.0 - math.exp(-distance))
            ids.append(node_id)

            logger.debug(f"node: {node}")
            logger.debug(f"similarity: {1.0 - math.exp(-distance)}")
            logger.debug(f"id: {node_id}")

        return VectorStoreQueryResult(nodes=nodes, similarities=similarities, ids=ids)

client property #

client: Any

获取Bagel集群。

add #

add(nodes: List[BaseNode], **add_kwargs: Any) -> List[str]

将带有嵌入的节点列表添加到向量存储中。

Returns:

Type Description
List[str]

文档id的列表。
Source code in llama_index/vector_stores/bagel/base.py 
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
    def add(self, nodes: List[BaseNode], **add_kwargs: Any) -> List[str]:
        """将带有嵌入的节点列表添加到向量存储中。

Args:
    nodes：带有嵌入的节点列表。
    kwargs：额外的参数。

Returns:
    文档id的列表。
"""
        if not self._collection:
            raise ValueError("collection not set")

        ids = []
        embeddings = []
        metadatas = []
        documents = []

        for node in nodes:
            ids.append(node.node_id)
            embeddings.append(node.get_embedding())
            metadatas.append(
                node_to_metadata_dict(
                    node,
                    remove_text=True,
                    flat_metadata=self.flat_metadata,
                )
            )
            documents.append(node.get_content(metadata_mode=MetadataMode.NONE) or "")

        self._collection.add(
            ids=ids, embeddings=embeddings, metadatas=metadatas, documents=documents
        )

        return ids

delete #

delete(ref_doc_id: str, **kwargs: Any) -> None

从向量存储中删除文档。

Source code in llama_index/vector_stores/bagel/base.py 
120
121
122
123
124
125
126
127
128
129
130
131
132
    def delete(self, ref_doc_id: str, **kwargs: Any) -> None:
        """从向量存储中删除文档。

Args:
    ref_doc_id：参考文档的id。
    kwargs：额外的参数。
"""
        if not self._collection:
            raise ValueError("collection not set")

        results = self._collection.get(where={"doc_id": ref_doc_id})
        if results and "ids" in results:
            self._collection.delete(ids=results["ids"])

query #

query(
    query: VectorStoreQuery, **kwargs: Any
) -> VectorStoreQueryResult

查询向量存储。

Parameters:

Name Type Description Default
query VectorStoreQuery

要运行的查询。

 required
kwargs Any

附加参数。

 {}

Returns:

Type Description
VectorStoreQueryResult

查询结果。
Source code in llama_index/vector_stores/bagel/base.py 
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
    def query(self, query: VectorStoreQuery, **kwargs: Any) -> VectorStoreQueryResult:
        """查询向量存储。

Args:
    query: 要运行的查询。
    kwargs: 附加参数。

Returns:
    查询结果。
"""
        if not self._collection:
            raise ValueError("collection not set")

        if query.filters is not None:
            if "where" in kwargs:
                raise ValueError("Cannot specify both filters and where")
            where = _to_bagel_filter(query.filters)
        else:
            where = kwargs.get("where", {})

        results = self._collection.find(
            query_embeddings=query.query_embedding,
            where=where,
            n_results=query.similarity_top_k,
            **kwargs,
        )

        logger.debug(f"query results: {results}")

        nodes = []
        similarities = []
        ids = []

        for node_id, text, metadata, distance in zip(
            results["ids"][0],
            results["documents"][0],
            results["metadatas"][0],
            results["distances"][0],
        ):
            try:
                node = metadata_dict_to_node(metadata)
                node.set_content(text)
            except Exception:
                # NOTE: deprecated legacy logic for backward compatibility
                metadata, node_info, relationships = legacy_metadata_dict_to_node(
                    metadata
                )

                node = TextNode(
                    text=text,
                    id_=node_id,
                    metadata=metadata,
                    start_char_idx=node_info.get("start", None),
                    end_char_idx=node_info.get("end", None),
                    relationships=relationships,
                )

            nodes.append(node)
            similarities.append(1.0 - math.exp(-distance))
            ids.append(node_id)

            logger.debug(f"node: {node}")
            logger.debug(f"similarity: {1.0 - math.exp(-distance)}")
            logger.debug(f"id: {node_id}")

        return VectorStoreQueryResult(nodes=nodes, similarities=similarities, ids=ids)