Jaguar

JaguarVectorStore

Bases: BasePydanticVectorStore

Jaguar向量存储。

请参阅 http://www.jaguardb.com 请参阅 http://github.com/fserv/jaguar-sdk

示例

pip install llama-index-vector-stores-jaguar

from llama_index.vector_stores.jaguar import JaguarVectorStore
vectorstore = JaguarVectorStore(
    pod = 'vdb',
    store = 'mystore',
    vector_index = 'v',
    vector_type = 'cosine_fraction_float',
    vector_dimension = 1536,
    url='http://192.168.8.88:8080/fwww/',
)

Source code in llama_index/vector_stores/jaguar/base.py

class JaguarVectorStore(BasePydanticVectorStore):
    """Jaguar向量存储。

    请参阅 http://www.jaguardb.com
    请参阅 http://github.com/fserv/jaguar-sdk

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

        ```python
        from llama_index.vector_stores.jaguar import JaguarVectorStore
        vectorstore = JaguarVectorStore(
            pod = 'vdb',
            store = 'mystore',
            vector_index = 'v',
            vector_type = 'cosine_fraction_float',
            vector_dimension = 1536,
            url='http://192.168.8.88:8080/fwww/',
        )
        ```"""

    stores_text: bool = True

    _pod: str = PrivateAttr()
    _store: str = PrivateAttr()
    _vector_index: str = PrivateAttr()
    _vector_type: str = PrivateAttr()
    _vector_dimension: int = PrivateAttr()
    _jag: JaguarHttpClient = PrivateAttr()
    _token: str = PrivateAttr()

    def __init__(
        self,
        pod: str,
        store: str,
        vector_index: str,
        vector_type: str,
        vector_dimension: int,
        url: str,
    ):
        """JaguarVectorStore的构造函数。

Args:
    pod: str:  pod的名称（数据库）
    store: str:  pod中向量存储的名称
    vector_index: str:  存储的向量索引的名称
    vector_type: str:  向量索引的类型
    vector_dimension: int:  向量索引的维度
    url: str:  jaguar http服务器的URL端点
"""
        super().__init__()
        self._pod = pod
        self._store = store
        self._vector_index = vector_index
        self._vector_type = vector_type
        self._vector_dimension = vector_dimension
        self._jag = JaguarHttpClient(url)
        self._token = ""

    def __del__(self) -> None:
        pass

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

    @property
    def client(self) -> Any:
        """获取客户端。"""
        return self._jag

    def add(
        self,
        nodes: List[BaseNode],
        **add_kwargs: Any,
    ) -> List[str]:
        """将节点添加到索引中。

Args:
    节点: List[BaseNode]: 带有嵌入的节点列表
"""
        use_node_metadata = add_kwargs.get("use_node_metadata", False)
        ids = []
        for node in nodes:
            text = node.get_text()
            embedding = node.get_embedding()
            if use_node_metadata is True:
                metadata = node.metadata
            else:
                metadata = None
            zid = self.add_text(text, embedding, metadata, **add_kwargs)
            ids.append(zid)

        return ids

    def delete(self, ref_doc_id: str, **delete_kwargs: Any) -> None:
        """使用ref_doc_id删除节点。

Args:
    ref_doc_id（str）：要删除的文档的doc_id。
"""
        podstore = self._pod + "." + self._store
        q = "delete from " + podstore + " where zid='" + ref_doc_id + "'"
        self.run(q)

    def query(self, query: VectorStoreQuery, **kwargs: Any) -> VectorStoreQueryResult:
        """查询前k个最相似节点的索引。

Args:
    query: VectorStoreQuery对象
    kwargs: 可能包含'where'、'metadata_fields'、'args'、'fetch_k'
"""
        embedding = query.query_embedding
        k = query.similarity_top_k
        (nodes, ids, simscores) = self.similarity_search_with_score(
            embedding, k=k, form="node", **kwargs
        )
        return VectorStoreQueryResult(nodes=nodes, ids=ids, similarities=simscores)

    def load_documents(
        self, embedding: List[float], k: int, **kwargs: Any
    ) -> List[Document]:
        """查询索引以加载最相似的前k个文档。

Args:
    embedding: 一组浮点数
    k: topK 数量
    kwargs: 可能包含'where'、'metadata_fields'、'args'、'fetch_k'
"""
        return cast(
            List[Document],
            self.similarity_search_with_score(embedding, k=k, form="doc", **kwargs),
        )

    def create(
        self,
        metadata_fields: str,
        text_size: int,
    ) -> None:
        """在后端数据库上创建向量存储。

Args:
    metadata_fields (str): 额外的元数据列和类型
Returns:
    如果成功则为True；如果不成功则为False
"""
        podstore = self._pod + "." + self._store

        """
        v:text column is required.
        """
        q = "create store "
        q += podstore
        q += f" ({self._vector_index} vector({self._vector_dimension},"
        q += f" '{self._vector_type}'),"
        q += f"  v:text char({text_size}),"
        q += metadata_fields + ")"
        self.run(q)

    def add_text(
        self,
        text: str,
        embedding: List[float],
        metadata: Optional[dict] = None,
        **kwargs: Any,
    ) -> str:
        """将文本通过嵌入添加到向量存储中。

Args:
  texts: 要添加到jaguar向量存储中的文本字符串。
  embedding: 文本的嵌入向量，浮点数列表
  metadata: {'file_path': '../data/paul_graham/paul_graham_essay.txt',
                  'file_name': 'paul_graham_essay.txt',
                  'file_type': 'text/plain',
                  'file_size': 75042,
                  'creation_date': '2023-12-24',
                  'last_modified_date': '2023-12-24',
                  'last_accessed_date': '2023-12-28'}
  kwargs: vector_index=向量索引的名称
          file_column=文件列的名称
          metadata={...}

Returns:
    将文本添加到向量存储中的ID
"""
        text = text.replace("'", "\\'")
        vcol = self._vector_index
        filecol = kwargs.get("file_column", "")
        text_tag = kwargs.get("text_tag", "")

        if text_tag != "":
            text = text_tag + " " + text

        podstorevcol = self._pod + "." + self._store + "." + vcol
        q = "textcol " + podstorevcol
        js = self.run(q)
        if js == "":
            return ""
        textcol = js["data"]

        zid = ""
        if metadata is None:
            ### no metadata and no files to upload
            str_vec = [str(x) for x in embedding]
            values_comma = ",".join(str_vec)
            podstore = self._pod + "." + self._store
            q = "insert into " + podstore + " ("
            q += vcol + "," + textcol + ") values ('" + values_comma
            q += "','" + text + "')"
            js = self.run(q, False)
            zid = js["zid"]
        else:
            str_vec = [str(x) for x in embedding]
            nvec, vvec, filepath = self._parseMeta(metadata, filecol)
            if filecol != "":
                rc = self._jag.postFile(self._token, filepath, 1)
                if not rc:
                    return ""
            names_comma = ",".join(nvec)
            names_comma += "," + vcol
            ## col1,col2,col3,vecl

            if vvec is not None and len(vvec) > 0:
                values_comma = "'" + "','".join(vvec) + "'"
            else:
                values_comma = "'" + "','".join(vvec) + "'"

            ### 'va1','val2','val3'
            values_comma += ",'" + ",".join(str_vec) + "'"
            ### 'v1,v2,v3'
            podstore = self._pod + "." + self._store
            q = "insert into " + podstore + " ("
            q += names_comma + "," + textcol + ") values (" + values_comma
            q += ",'" + text + "')"
            if filecol != "":
                js = self.run(q, True)
            else:
                js = self.run(q, False)
            zid = js["zid"]

        return zid

    def similarity_search_with_score(
        self,
        embedding: Optional[List[float]],
        k: int = 3,
        form: str = "node",
        **kwargs: Any,
    ) -> Union[Tuple[List[TextNode], List[str], List[float]], List[Document]]:
        """返回与查询嵌入最相似的节点，以及其ID和分数。

Args:
    embedding：要查找的文本嵌入。
    k：要返回的节点数。默认为3。
    form：如果是“node”，则返回Tuple[List[TextNode], List[str], List[float]]
          如果是“doc”，则返回List[Document]
    kwargs：可能包括where、metadata_fields、args、fetch_k
Returns:
    元组（节点列表，ID列表，相似度分数列表）
"""
        where = kwargs.get("where", None)
        metadata_fields = kwargs.get("metadata_fields", None)

        args = kwargs.get("args", None)
        fetch_k = kwargs.get("fetch_k", -1)

        vcol = self._vector_index
        vtype = self._vector_type
        if embedding is None:
            return ([], [], [])
        str_embeddings = [str(f) for f in embedding]
        qv_comma = ",".join(str_embeddings)
        podstore = self._pod + "." + self._store
        q = (
            "select similarity("
            + vcol
            + ",'"
            + qv_comma
            + "','topk="
            + str(k)
            + ",fetch_k="
            + str(fetch_k)
            + ",type="
            + vtype
        )
        q += ",with_score=yes,with_text=yes"
        if args is not None:
            q += "," + args

        if metadata_fields is not None:
            x = "&".join(metadata_fields)
            q += ",metadata=" + x

        q += "') from " + podstore

        if where is not None:
            q += " where " + where

        jarr = self.run(q)

        if jarr is None:
            return ([], [], [])

        nodes = []
        ids = []
        simscores = []
        docs = []
        for js in jarr:
            score = js["score"]
            text = js["text"]
            zid = js["zid"]

            md = {}
            md["zid"] = zid
            if metadata_fields is not None:
                for m in metadata_fields:
                    mv = js[m]
                    md[m] = mv

            if form == "node":
                node = TextNode(
                    id_=zid,
                    text=text,
                    metadata=md,
                )
                nodes.append(node)
                ids.append(zid)
                simscores.append(float(score))
            else:
                doc = Document(
                    id_=zid,
                    text=text,
                    metadata=md,
                )
                docs.append(doc)

        if form == "node":
            return (nodes, ids, simscores)
        else:
            return docs

    def is_anomalous(
        self,
        node: BaseNode,
        **kwargs: Any,
    ) -> bool:
        """检测给定文本是否在数据集中是异常的。

Args:
    query：要检测是否是异常的文本
Returns:
    True 或 False
"""
        vcol = self._vector_index
        vtype = self._vector_type
        str_embeddings = [str(f) for f in node.get_embedding()]
        qv_comma = ",".join(str_embeddings)
        podstore = self._pod + "." + self._store
        q = "select anomalous(" + vcol + ", '" + qv_comma + "', 'type=" + vtype + "')"
        q += " from " + podstore

        js = self.run(q)
        if isinstance(js, list) and len(js) == 0:
            return False
        jd = json.loads(js[0])
        if jd["anomalous"] == "YES":
            return True
        return False

    def run(self, query: str, withFile: bool = False) -> dict:
        """在jaguardb中运行任何查询语句。

Args:
    query (str): 要发送到jaguardb的查询语句
Returns:
    无效令牌时返回None，或者
    json结果字符串
"""
        if self._token == "":
            logger.error(f"E0005 error run({query})")
            return {}

        resp = self._jag.post(query, self._token, withFile)
        txt = resp.text
        try:
            return json.loads(txt)
        except Exception:
            return {}

    def count(self) -> int:
        """统计jaguardb中商店的记录数。

Args:无参数
Returns:(int) 商店中记录的数量
"""
        podstore = self._pod + "." + self._store
        q = "select count() from " + podstore
        js = self.run(q)
        if isinstance(js, list) and len(js) == 0:
            return 0
        jd = json.loads(js[0])
        return int(jd["data"])

    def clear(self) -> None:
        """删除jaguardb中的所有记录。

Args:无参数
Returns:无
"""
        podstore = self._pod + "." + self._store
        q = "truncate store " + podstore
        self.run(q)

    def drop(self) -> None:
        """删除或移除jaguardb中的存储。

Args:无
Returns:无
"""
        podstore = self._pod + "." + self._store
        q = "drop store " + podstore
        self.run(q)

    def prt(self, msg: str) -> None:
        nows = datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")
        with open("/tmp/debugjaguar.log", "a") as file:
            print(f"{nows} msg={msg}", file=file, flush=True)

    def login(
        self,
        jaguar_api_key: Optional[str] = "",
    ) -> bool:
        """使用 jaguar_api_key 登录到 jaguar 服务器，或者让 self._jag 找到一个密钥。

Args:
    可选的 jaguar_api_key (str): 用户到 jaguardb 服务器的 API 密钥
Returns:
    如果成功则返回 True；如果不成功则返回 False
"""
        if jaguar_api_key == "":
            jaguar_api_key = self._jag.getApiKey()
        self._jaguar_api_key = jaguar_api_key
        self._token = self._jag.login(jaguar_api_key)
        if self._token == "":
            logger.error("E0001 error init(): invalid jaguar_api_key")
            return False
        return True

    def logout(self) -> None:
        """登出以清理资源。

Args:无参数
Returns:无
"""
        self._jag.logout(self._token)

    def _parseMeta(self, nvmap: dict, filecol: str) -> Tuple[List[str], List[str], str]:
        filepath = ""
        if filecol == "":
            nvec = list(nvmap.keys())
            vvec = list(nvmap.values())
        else:
            nvec = []
            vvec = []
            if filecol in nvmap:
                nvec.append(filecol)
                vvec.append(nvmap[filecol])
                filepath = nvmap[filecol]

            for k, v in nvmap.items():
                if k != filecol:
                    nvec.append(k)
                    vvec.append(v)

        return nvec, vvec, filepath

client property

client: Any

获取客户端。

add

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

将节点添加到索引中。

Parameters:

Name	Type	Description	Default
节点		List[BaseNode]: 带有嵌入的节点列表	required

Source code in llama_index/vector_stores/jaguar/base.py

    def add(
        self,
        nodes: List[BaseNode],
        **add_kwargs: Any,
    ) -> List[str]:
        """将节点添加到索引中。

Args:
    节点: List[BaseNode]: 带有嵌入的节点列表
"""
        use_node_metadata = add_kwargs.get("use_node_metadata", False)
        ids = []
        for node in nodes:
            text = node.get_text()
            embedding = node.get_embedding()
            if use_node_metadata is True:
                metadata = node.metadata
            else:
                metadata = None
            zid = self.add_text(text, embedding, metadata, **add_kwargs)
            ids.append(zid)

        return ids

delete

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

使用ref_doc_id删除节点。

Source code in llama_index/vector_stores/jaguar/base.py

    def delete(self, ref_doc_id: str, **delete_kwargs: Any) -> None:
        """使用ref_doc_id删除节点。

Args:
    ref_doc_id（str）：要删除的文档的doc_id。
"""
        podstore = self._pod + "." + self._store
        q = "delete from " + podstore + " where zid='" + ref_doc_id + "'"
        self.run(q)

query

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

查询前k个最相似节点的索引。

Parameters:

Name	Type	Description	Default
query	VectorStoreQuery	VectorStoreQuery对象	required
kwargs	Any	可能包含'where'、'metadata_fields'、'args'、'fetch_k'	{}

Source code in llama_index/vector_stores/jaguar/base.py

    def query(self, query: VectorStoreQuery, **kwargs: Any) -> VectorStoreQueryResult:
        """查询前k个最相似节点的索引。

Args:
    query: VectorStoreQuery对象
    kwargs: 可能包含'where'、'metadata_fields'、'args'、'fetch_k'
"""
        embedding = query.query_embedding
        k = query.similarity_top_k
        (nodes, ids, simscores) = self.similarity_search_with_score(
            embedding, k=k, form="node", **kwargs
        )
        return VectorStoreQueryResult(nodes=nodes, ids=ids, similarities=simscores)

load_documents

load_documents(
    embedding: List[float], k: int, **kwargs: Any
) -> List[Document]

查询索引以加载最相似的前k个文档。

Parameters:

Name	Type	Description	Default
embedding	List[float]	一组浮点数	required
k	int	topK 数量	required
kwargs	Any	可能包含'where'、'metadata_fields'、'args'、'fetch_k'	{}

Source code in llama_index/vector_stores/jaguar/base.py

    def load_documents(
        self, embedding: List[float], k: int, **kwargs: Any
    ) -> List[Document]:
        """查询索引以加载最相似的前k个文档。

Args:
    embedding: 一组浮点数
    k: topK 数量
    kwargs: 可能包含'where'、'metadata_fields'、'args'、'fetch_k'
"""
        return cast(
            List[Document],
            self.similarity_search_with_score(embedding, k=k, form="doc", **kwargs),
        )

create

create(metadata_fields: str, text_size: int) -> None

在后端数据库上创建向量存储。

Parameters:

Name	Type	Description	Default
metadata_fields	str	额外的元数据列和类型	required

Returns: 如果成功则为True；如果不成功则为False

Source code in llama_index/vector_stores/jaguar/base.py

    def create(
        self,
        metadata_fields: str,
        text_size: int,
    ) -> None:
        """在后端数据库上创建向量存储。

Args:
    metadata_fields (str): 额外的元数据列和类型
Returns:
    如果成功则为True；如果不成功则为False
"""
        podstore = self._pod + "." + self._store

        """
        v:text column is required.
        """
        q = "create store "
        q += podstore
        q += f" ({self._vector_index} vector({self._vector_dimension},"
        q += f" '{self._vector_type}'),"
        q += f"  v:text char({text_size}),"
        q += metadata_fields + ")"
        self.run(q)

add_text

add_text(
    text: str,
    embedding: List[float],
    metadata: Optional[dict] = None,
    **kwargs: Any
) -> str

将文本通过嵌入添加到向量存储中。

Parameters:

Name	Type	Description	Default
texts		要添加到jaguar向量存储中的文本字符串。	required
embedding	List[float]	文本的嵌入向量，浮点数列表	required
metadata	Optional[dict]	{'file_path': '../data/paul_graham/paul_graham_essay.txt', 'file_name': 'paul_graham_essay.txt', 'file_type': 'text/plain', 'file_size': 75042, 'creation_date': '2023-12-24', 'last_modified_date': '2023-12-24', 'last_accessed_date': '2023-12-28'}	None
kwargs	Any	vector_index=向量索引的名称 file_column=文件列的名称 metadata={...}	{}

Returns:

Type	Description
str	将文本添加到向量存储中的ID

Source code in llama_index/vector_stores/jaguar/base.py

    def add_text(
        self,
        text: str,
        embedding: List[float],
        metadata: Optional[dict] = None,
        **kwargs: Any,
    ) -> str:
        """将文本通过嵌入添加到向量存储中。

Args:
  texts: 要添加到jaguar向量存储中的文本字符串。
  embedding: 文本的嵌入向量，浮点数列表
  metadata: {'file_path': '../data/paul_graham/paul_graham_essay.txt',
                  'file_name': 'paul_graham_essay.txt',
                  'file_type': 'text/plain',
                  'file_size': 75042,
                  'creation_date': '2023-12-24',
                  'last_modified_date': '2023-12-24',
                  'last_accessed_date': '2023-12-28'}
  kwargs: vector_index=向量索引的名称
          file_column=文件列的名称
          metadata={...}

Returns:
    将文本添加到向量存储中的ID
"""
        text = text.replace("'", "\\'")
        vcol = self._vector_index
        filecol = kwargs.get("file_column", "")
        text_tag = kwargs.get("text_tag", "")

        if text_tag != "":
            text = text_tag + " " + text

        podstorevcol = self._pod + "." + self._store + "." + vcol
        q = "textcol " + podstorevcol
        js = self.run(q)
        if js == "":
            return ""
        textcol = js["data"]

        zid = ""
        if metadata is None:
            ### no metadata and no files to upload
            str_vec = [str(x) for x in embedding]
            values_comma = ",".join(str_vec)
            podstore = self._pod + "." + self._store
            q = "insert into " + podstore + " ("
            q += vcol + "," + textcol + ") values ('" + values_comma
            q += "','" + text + "')"
            js = self.run(q, False)
            zid = js["zid"]
        else:
            str_vec = [str(x) for x in embedding]
            nvec, vvec, filepath = self._parseMeta(metadata, filecol)
            if filecol != "":
                rc = self._jag.postFile(self._token, filepath, 1)
                if not rc:
                    return ""
            names_comma = ",".join(nvec)
            names_comma += "," + vcol
            ## col1,col2,col3,vecl

            if vvec is not None and len(vvec) > 0:
                values_comma = "'" + "','".join(vvec) + "'"
            else:
                values_comma = "'" + "','".join(vvec) + "'"

            ### 'va1','val2','val3'
            values_comma += ",'" + ",".join(str_vec) + "'"
            ### 'v1,v2,v3'
            podstore = self._pod + "." + self._store
            q = "insert into " + podstore + " ("
            q += names_comma + "," + textcol + ") values (" + values_comma
            q += ",'" + text + "')"
            if filecol != "":
                js = self.run(q, True)
            else:
                js = self.run(q, False)
            zid = js["zid"]

        return zid

similarity_search_with_score

similarity_search_with_score(
    embedding: Optional[List[float]],
    k: int = 3,
    form: str = "node",
    **kwargs: Any
) -> Union[
    Tuple[List[TextNode], List[str], List[float]],
    List[Document],
]

返回与查询嵌入最相似的节点，以及其ID和分数。

Returns: 元组（节点列表，ID列表，相似度分数列表）

Source code in llama_index/vector_stores/jaguar/base.py

    def similarity_search_with_score(
        self,
        embedding: Optional[List[float]],
        k: int = 3,
        form: str = "node",
        **kwargs: Any,
    ) -> Union[Tuple[List[TextNode], List[str], List[float]], List[Document]]:
        """返回与查询嵌入最相似的节点，以及其ID和分数。

Args:
    embedding：要查找的文本嵌入。
    k：要返回的节点数。默认为3。
    form：如果是“node”，则返回Tuple[List[TextNode], List[str], List[float]]
          如果是“doc”，则返回List[Document]
    kwargs：可能包括where、metadata_fields、args、fetch_k
Returns:
    元组（节点列表，ID列表，相似度分数列表）
"""
        where = kwargs.get("where", None)
        metadata_fields = kwargs.get("metadata_fields", None)

        args = kwargs.get("args", None)
        fetch_k = kwargs.get("fetch_k", -1)

        vcol = self._vector_index
        vtype = self._vector_type
        if embedding is None:
            return ([], [], [])
        str_embeddings = [str(f) for f in embedding]
        qv_comma = ",".join(str_embeddings)
        podstore = self._pod + "." + self._store
        q = (
            "select similarity("
            + vcol
            + ",'"
            + qv_comma
            + "','topk="
            + str(k)
            + ",fetch_k="
            + str(fetch_k)
            + ",type="
            + vtype
        )
        q += ",with_score=yes,with_text=yes"
        if args is not None:
            q += "," + args

        if metadata_fields is not None:
            x = "&".join(metadata_fields)
            q += ",metadata=" + x

        q += "') from " + podstore

        if where is not None:
            q += " where " + where

        jarr = self.run(q)

        if jarr is None:
            return ([], [], [])

        nodes = []
        ids = []
        simscores = []
        docs = []
        for js in jarr:
            score = js["score"]
            text = js["text"]
            zid = js["zid"]

            md = {}
            md["zid"] = zid
            if metadata_fields is not None:
                for m in metadata_fields:
                    mv = js[m]
                    md[m] = mv

            if form == "node":
                node = TextNode(
                    id_=zid,
                    text=text,
                    metadata=md,
                )
                nodes.append(node)
                ids.append(zid)
                simscores.append(float(score))
            else:
                doc = Document(
                    id_=zid,
                    text=text,
                    metadata=md,
                )
                docs.append(doc)

        if form == "node":
            return (nodes, ids, simscores)
        else:
            return docs

is_anomalous

is_anomalous(node: BaseNode, **kwargs: Any) -> bool

检测给定文本是否在数据集中是异常的。

Returns: True 或 False

Source code in llama_index/vector_stores/jaguar/base.py

    def is_anomalous(
        self,
        node: BaseNode,
        **kwargs: Any,
    ) -> bool:
        """检测给定文本是否在数据集中是异常的。

Args:
    query：要检测是否是异常的文本
Returns:
    True 或 False
"""
        vcol = self._vector_index
        vtype = self._vector_type
        str_embeddings = [str(f) for f in node.get_embedding()]
        qv_comma = ",".join(str_embeddings)
        podstore = self._pod + "." + self._store
        q = "select anomalous(" + vcol + ", '" + qv_comma + "', 'type=" + vtype + "')"
        q += " from " + podstore

        js = self.run(q)
        if isinstance(js, list) and len(js) == 0:
            return False
        jd = json.loads(js[0])
        if jd["anomalous"] == "YES":
            return True
        return False

run

run(query: str, withFile: bool = False) -> dict

在jaguardb中运行任何查询语句。

Parameters:

Name	Type	Description	Default
query	str	要发送到jaguardb的查询语句	required

Returns: 无效令牌时返回None，或者 json结果字符串

Source code in llama_index/vector_stores/jaguar/base.py

    def run(self, query: str, withFile: bool = False) -> dict:
        """在jaguardb中运行任何查询语句。

Args:
    query (str): 要发送到jaguardb的查询语句
Returns:
    无效令牌时返回None，或者
    json结果字符串
"""
        if self._token == "":
            logger.error(f"E0005 error run({query})")
            return {}

        resp = self._jag.post(query, self._token, withFile)
        txt = resp.text
        try:
            return json.loads(txt)
        except Exception:
            return {}

count

count() -> int

统计jaguardb中商店的记录数。

Args:无参数 Returns:(int) 商店中记录的数量

Source code in llama_index/vector_stores/jaguar/base.py

    def count(self) -> int:
        """统计jaguardb中商店的记录数。

Args:无参数
Returns:(int) 商店中记录的数量
"""
        podstore = self._pod + "." + self._store
        q = "select count() from " + podstore
        js = self.run(q)
        if isinstance(js, list) and len(js) == 0:
            return 0
        jd = json.loads(js[0])
        return int(jd["data"])

clear

clear() -> None

删除jaguardb中的所有记录。

Args:无参数 Returns:无

Source code in llama_index/vector_stores/jaguar/base.py

    def clear(self) -> None:
        """删除jaguardb中的所有记录。

Args:无参数
Returns:无
"""
        podstore = self._pod + "." + self._store
        q = "truncate store " + podstore
        self.run(q)

drop

drop() -> None

删除或移除jaguardb中的存储。

Args:无 Returns:无

Source code in llama_index/vector_stores/jaguar/base.py

    def drop(self) -> None:
        """删除或移除jaguardb中的存储。

Args:无
Returns:无
"""
        podstore = self._pod + "." + self._store
        q = "drop store " + podstore
        self.run(q)

login

login(jaguar_api_key: Optional[str] = '') -> bool

使用 jaguar_api_key 登录到 jaguar 服务器，或者让 self._jag 找到一个密钥。

Parameters:

Name	Type	Description	Default
可选的	jaguar_api_key (str	用户到 jaguardb 服务器的 API 密钥	required

Returns: 如果成功则返回 True；如果不成功则返回 False

Source code in llama_index/vector_stores/jaguar/base.py

    def login(
        self,
        jaguar_api_key: Optional[str] = "",
    ) -> bool:
        """使用 jaguar_api_key 登录到 jaguar 服务器，或者让 self._jag 找到一个密钥。

Args:
    可选的 jaguar_api_key (str): 用户到 jaguardb 服务器的 API 密钥
Returns:
    如果成功则返回 True；如果不成功则返回 False
"""
        if jaguar_api_key == "":
            jaguar_api_key = self._jag.getApiKey()
        self._jaguar_api_key = jaguar_api_key
        self._token = self._jag.login(jaguar_api_key)
        if self._token == "":
            logger.error("E0001 error init(): invalid jaguar_api_key")
            return False
        return True

logout

logout() -> None

登出以清理资源。

Args:无参数 Returns:无

Source code in llama_index/vector_stores/jaguar/base.py

    def logout(self) -> None:
        """登出以清理资源。

Args:无参数
Returns:无
"""
        self._jag.logout(self._token)