pyobvector 0.2.22__py3-none-any.whl → 0.2.23__py3-none-any.whl

pyobvector/__init__.py

@@ -1,10 +1,10 @@
 """A python SDK for OceanBase Vector Store, based on SQLAlchemy, compatible with Milvus API.
 
-`pyobvector` supports two modes: 
-1. `Milvus compatible mode`: You can use the `MilvusLikeClient` class to use vector storage 
+`pyobvector` supports two modes:
+1. `Milvus compatible mode`: You can use the `MilvusLikeClient` class to use vector storage
 in a way similar to the Milvus API.
-2. `SQLAlchemy hybrid mode`: You can use the vector storage function provided by the 
-`ObVecClient` class and execute the relational database statement with the SQLAlchemy library. 
+2. `SQLAlchemy hybrid mode`: You can use the vector storage function provided by the
+`ObVecClient` class and execute the relational database statement with the SQLAlchemy library.
 In this mode, you can regard `pyobvector` as an extension of SQLAlchemy.
 
 * ObVecClient           MySQL client in SQLAlchemy hybrid mode
@@ -19,7 +19,7 @@ In this mode, you can regard `pyobvector` as an extension of SQLAlchemy.
 * FtsIndex              Full Text Search Index
 * FieldSchema           Clas to define field schema in collection for MilvusLikeClient
 * CollectionSchema      Class to define collection schema for MilvusLikeClient
-* PartType              Specify partition type of table or collection 
+* PartType              Specify partition type of table or collection
                         for both ObVecClient and MilvusLikeClient
 * ObPartition           Abstract type class of all kind of Partition strategy
 * RangeListPartInfo     Specify Range/RangeColumns/List/ListColumns partition info
@@ -40,6 +40,7 @@ In this mode, you can regard `pyobvector` as an extension of SQLAlchemy.
 * FtsIndexParam         Full Text Search index parameter
 * MatchAgainst          Full Text Search clause
 """
+
 from .client import *
 from .schema import (
     ARRAY,

pyobvector/client/__init__.py

@@ -1,9 +1,9 @@
 """Multi-type Vector Store Client:
 
-1. `Milvus compatible mode`: You can use the `MilvusLikeClient` class to use vector storage 
+1. `Milvus compatible mode`: You can use the `MilvusLikeClient` class to use vector storage
 in a way similar to the Milvus API.
-2. `SQLAlchemy hybrid mode`: You can use the vector storage function provided by the 
-`ObVecClient` class and execute the relational database statement with the SQLAlchemy library. 
+2. `SQLAlchemy hybrid mode`: You can use the vector storage function provided by the
+`ObVecClient` class and execute the relational database statement with the SQLAlchemy library.
 In this mode, you can regard `pyobvector` as an extension of SQLAlchemy.
 
 * ObVecClient           MySQL client in SQLAlchemy hybrid mode
@@ -14,7 +14,7 @@ In this mode, you can regard `pyobvector` as an extension of SQLAlchemy.
 * DataType              Specify field type in collection schema for MilvusLikeClient
 * FieldSchema           Clas to define field schema in collection for MilvusLikeClient
 * CollectionSchema      Class to define collection schema for MilvusLikeClient
-* PartType              Specify partition type of table or collection 
+* PartType              Specify partition type of table or collection
                         for both ObVecClient and MilvusLikeClient
 * ObPartition           Abstract type class of all kind of Partition strategy
 * RangeListPartInfo     Specify Range/RangeColumns/List/ListColumns partition info
@@ -30,6 +30,7 @@ In this mode, you can regard `pyobvector` as an extension of SQLAlchemy.
 * FtsParser             Text Parser Type for Full Text Search
 * FtsIndexParam         Full Text Search index parameter
 """
+
 from .ob_vec_client import ObVecClient
 from .milvus_like_client import MilvusLikeClient
 from .ob_vec_json_table_client import ObVecJsonTableClient

pyobvector/client/collection_schema.py

@@ -1,4 +1,5 @@
 """FieldSchema & CollectionSchema definition module to be compatible with Milvus."""
+
 import copy
 from typing import Optional
 from sqlalchemy import Column
@@ -6,6 +7,7 @@ from .schema_type import DataType, convert_datatype_to_sqltype
 from .exceptions import *
 from .partitions import *
 
+
 class FieldSchema:
     """FieldSchema definition.
 
@@ -18,6 +20,7 @@ class FieldSchema:
     nullable (bool) : whether the field can be null
     type_params (dict) : different parameters for different data type
     """
+
     def __init__(
         self,
         name: str,
@@ -117,12 +120,13 @@ class FieldSchema:
 
 class CollectionSchema:
     """CollectionSchema definition.
-    
+
     Attributes:
     fields (List[FieldSchema]) : a list of FieldSchema
     description (string) : collection description (not used in OceanBase)
     partitions (ObPartition) : partition strategy of this collection
     """
+
     def __init__(
         self,
         fields: Optional[list[FieldSchema]] = None,

pyobvector/client/enum.py

@@ -1,7 +1,7 @@
 """Common module for int type enumerate."""
+
 from enum import Enum
 
 
 class IntEnum(int, Enum):
     """Int type enumerate definition."""
-    

pyobvector/client/exceptions.py

@@ -1,9 +1,11 @@
 """Exception for MilvusLikeClient."""
+
 from .enum import IntEnum
 
 
 class ErrorCode(IntEnum):
     """Error codes for MilvusLikeClient."""
+
     SUCCESS = 0
     UNEXPECTED_ERROR = 1
     INVALID_ARGUMENT = 2
@@ -14,6 +16,7 @@ class ErrorCode(IntEnum):
 
 class ObException(Exception):
     """Base class for MilvusLikeClient exception."""
+
     def __init__(
         self,
         code: int = ErrorCode.UNEXPECTED_ERROR,
@@ -75,6 +78,7 @@ class ClusterVersionException(ObException):
 
 class ExceptionsMessage:
     """Exception Messages definition."""
+
     PartitionExprNotExists = "Partition expression string does not exist."
     PartitionMultiField = "Multi-Partition Field is not supported."
     PartitionLevelMoreThanTwo = "Partition Level should less than or equal to 2."
@@ -93,12 +97,8 @@ class ExceptionsMessage:
     PartitionListColNameListMissing = (
         "Column name list is necessary when partition type is ListColumns"
     )
-    PartitionHashNameListAndPartCntMissing = (
-        "One of hash_part_name_list and part_count must be set when partition type is Hash"
-    )
-    PartitionKeyNameListAndPartCntMissing = (
-        "One of key_part_name_list and part_count must be set when partition type is Key"
-    )
+    PartitionHashNameListAndPartCntMissing = "One of hash_part_name_list and part_count must be set when partition type is Hash"
+    PartitionKeyNameListAndPartCntMissing = "One of key_part_name_list and part_count must be set when partition type is Key"
     PrimaryFieldType = "Param primary_field must be int or str type."
     VectorFieldMissingDimParam = "Param 'dim' must be set for vector field."
     VarcharFieldMissingLengthParam = "Param 'max_length' must be set for varchar field."
@@ -108,7 +108,9 @@ class ExceptionsMessage:
     )
     CollectionNotExists = "Collection does not exist."
     MetricTypeParamTypeInvalid = "MetricType param type should be string."
-    MetricTypeValueInvalid = "MetricType should be 'l2'/'ip'/'neg_ip'/'cosine' in ann search."
+    MetricTypeValueInvalid = (
+        "MetricType should be 'l2'/'ip'/'neg_ip'/'cosine' in ann search."
+    )
     UsingInIDsWhenMultiPrimaryKey = "Using 'ids' when table has multi primary key."
     ClusterVersionIsLow = (
         "OceanBase %s feature is not supported because cluster version is below %s."

pyobvector/client/fts_index_param.py

@@ -1,9 +1,12 @@
 """A module to specify fts index parameters"""
+
 from enum import Enum
 from typing import Optional, Union
 
+
 class FtsParser(Enum):
     """Built-in full-text search parser types supported by OceanBase"""
+
     IK = 0
     NGRAM = 1
     NGRAM2 = 2  # NGRAM2 parser (supported from V4.3.5 BP2+)
@@ -13,13 +16,14 @@ class FtsParser(Enum):
 
 class FtsIndexParam:
     """Full-text search index parameter.
-    
+
     Args:
         index_name: Index name
         field_names: List of field names to create full-text index on
         parser_type: Parser type, can be FtsParser enum or string (for custom parsers)
                     If None, uses default Space parser
     """
+
     def __init__(
         self,
         index_name: str,
@@ -34,11 +38,11 @@ class FtsIndexParam:
         """Convert parser type to string format for SQL."""
         if self.parser_type is None:
             return None  # Default Space parser, no need to specify
-        
+
         if isinstance(self.parser_type, str):
             # Custom parser name (e.g., "thai_ftparser")
             return self.parser_type.lower()
-        
+
         if isinstance(self.parser_type, FtsParser):
             if self.parser_type == FtsParser.IK:
                 return "ik"
@@ -52,7 +56,7 @@ class FtsIndexParam:
                 return "jieba"
             # Raise exception for unrecognized FtsParser enum values
             raise ValueError(f"Unrecognized FtsParser enum value: {self.parser_type}")
-        
+
         return None
 
     def __iter__(self):

pyobvector/client/hybrid_search.py

@@ -1,4 +1,5 @@
 """OceanBase Hybrid Search Client."""
+
 import json
 import logging
 from typing import Any
@@ -27,7 +28,7 @@ class HybridSearch(Client):
         super().__init__(uri, user, password, db_name, **kwargs)
 
         min_required_version = ObVersion.from_db_version_nums(4, 4, 1, 0)
-        
+
         if self.ob_version < min_required_version:
             # For versions < 4.4.1.0, check if it's SeekDB
             if self._is_seekdb():
@@ -35,7 +36,8 @@ class HybridSearch(Client):
                 return
             raise ClusterVersionException(
                 code=ErrorCode.NOT_SUPPORTED,
-                message=ExceptionsMessage.ClusterVersionIsLow % ("Hybrid Search", "4.4.1.0"),
+                message=ExceptionsMessage.ClusterVersionIsLow
+                % ("Hybrid Search", "4.4.1.0"),
             )
 
     def search(
@@ -60,7 +62,9 @@ class HybridSearch(Client):
 
         with self.engine.connect() as conn:
             with conn.begin():
-                res = conn.execute(sql, {"index": index, "body_str": body_str}).fetchone()
+                res = conn.execute(
+                    sql, {"index": index, "body_str": body_str}
+                ).fetchone()
                 if res[0] is None:
                     return []
                 return json.loads(res[0])
@@ -85,7 +89,9 @@ class HybridSearch(Client):
 
         with self.engine.connect() as conn:
             with conn.begin():
-                res = conn.execute(sql, {"index": index, "body_str": body_str}).fetchone()
+                res = conn.execute(
+                    sql, {"index": index, "body_str": body_str}
+                ).fetchone()
                 if res[0] is None:
                     return ""
                 return res[0]

pyobvector/client/index_param.py

@@ -1,9 +1,12 @@
 """A module to specify vector index parameters for MilvusLikeClient"""
+
 from enum import Enum
 from typing import Union
 
+
 class VecIndexType(Enum):
     """Vector index algorithm type"""
+
     HNSW = 0
     HNSW_SQ = 1
     IVFFLAT = 2
@@ -11,21 +14,23 @@ class VecIndexType(Enum):
     IVFPQ = 4
     DAAT = 5
 
+
 class IndexParam:
     """Vector index parameters.
-    
+
     Attributes:
     index_name (string) : vector index name
     field_name (string) : vector index built on which field
     index_type (VecIndexType) :
         vector index algorithms (Only HNSW supported)
-    kwargs : 
+    kwargs :
         vector index parameters for different algorithms
     """
+
     HNSW_DEFAULT_M = 16
     HNSW_DEFAULT_EF_CONSTRUCTION = 200
     HNSW_DEFAULT_EF_SEARCH = 40
-    OCEANBASE_DEFAULT_ALGO_LIB = 'vsag'
+    OCEANBASE_DEFAULT_ALGO_LIB = "vsag"
     HNSW_ALGO_NAME = "hnsw"
     HNSW_SQ_ALGO_NAME = "hnsw_sq"
     IVFFLAT_ALGO_NAME = "ivf_flat"
@@ -34,7 +39,11 @@ class IndexParam:
     DAAT_ALGO_NAME = "daat"
 
     def __init__(
-        self, index_name: str, field_name: str, index_type: Union[VecIndexType, str], **kwargs
+        self,
+        index_name: str,
+        field_name: str,
+        index_type: Union[VecIndexType, str],
+        **kwargs,
     ):
         self.index_name = index_name
         self.field_name = field_name
@@ -44,21 +53,22 @@ class IndexParam:
 
     def is_index_type_hnsw_serial(self):
         return self.index_type in [
-            IndexParam.HNSW_ALGO_NAME, IndexParam.HNSW_SQ_ALGO_NAME
+            IndexParam.HNSW_ALGO_NAME,
+            IndexParam.HNSW_SQ_ALGO_NAME,
         ]
-    
+
     def is_index_type_ivf_serial(self):
         return self.index_type in [
             IndexParam.IVFFLAT_ALGO_NAME,
             IndexParam.IVFSQ_ALGO_NAME,
             IndexParam.IVFPQ_ALGO_NAME,
         ]
-    
+
     def is_index_type_product_quantization(self):
         return self.index_type in [
             IndexParam.IVFPQ_ALGO_NAME,
         ]
-    
+
     def is_index_type_sparse_vector(self):
         return self.index_type in [
             IndexParam.DAAT_ALGO_NAME,
@@ -97,48 +107,52 @@ class IndexParam:
         ob_params = {}
         # handle lib
         if self.is_index_type_hnsw_serial():
-            ob_params['lib'] = 'vsag'
+            ob_params["lib"] = "vsag"
         else:
-            ob_params['lib'] = 'OB'
+            ob_params["lib"] = "OB"
         # handle metric_type
-        ob_params['distance'] = "l2"
-        if 'metric_type' in self.kwargs:
-            ob_params['distance'] = self.kwargs['metric_type']
+        ob_params["distance"] = "l2"
+        if "metric_type" in self.kwargs:
+            ob_params["distance"] = self.kwargs["metric_type"]
         # handle param
         if self.is_index_type_ivf_serial():
-            if (self.is_index_type_product_quantization() and
-                'params' not in self.kwargs):
-                raise ValueError('params must be configured for IVF index type')
-            
-            if 'params' not in self.kwargs:
+            if (
+                self.is_index_type_product_quantization()
+                and "params" not in self.kwargs
+            ):
+                raise ValueError("params must be configured for IVF index type")
+
+            if "params" not in self.kwargs:
                 params = {}
             else:
-                params = self.kwargs['params']
-            
+                params = self.kwargs["params"]
+
             if self.is_index_type_product_quantization():
-                if 'm' not in params:
-                    raise ValueError('m must be configured for IVFSQ or IVFPQ')
-                ob_params['m'] = params['m']
-            if 'nlist' in params:
-                ob_params['nlist'] = params['nlist']
-            if 'samples_per_nlist' in params:
-                ob_params['samples_per_nlist'] = params['samples_per_nlist']
+                if "m" not in params:
+                    raise ValueError("m must be configured for IVFSQ or IVFPQ")
+                ob_params["m"] = params["m"]
+            if "nlist" in params:
+                ob_params["nlist"] = params["nlist"]
+            if "samples_per_nlist" in params:
+                ob_params["samples_per_nlist"] = params["samples_per_nlist"]
 
         if self.is_index_type_hnsw_serial():
-            if 'params' in self.kwargs:
-                params = self.kwargs['params']
-                if 'M' in params:
-                    ob_params['m'] = params['M']
-                if 'efConstruction' in params:
-                    ob_params['ef_construction'] = params['efConstruction']
-                if 'efSearch' in params:
-                    ob_params['ef_search'] = params['efSearch']
-        
+            if "params" in self.kwargs:
+                params = self.kwargs["params"]
+                if "M" in params:
+                    ob_params["m"] = params["M"]
+                if "efConstruction" in params:
+                    ob_params["ef_construction"] = params["efConstruction"]
+                if "efSearch" in params:
+                    ob_params["ef_search"] = params["efSearch"]
+
         if self.is_index_type_sparse_vector():
-            if ob_params['distance'] != 'inner_product':
-                raise ValueError("Metric type should be 'inner_product' for sparse vector index.")
-            if 'sparse_index_type' in self.kwargs:
-                ob_params['type'] = self.kwargs['sparse_index_type']
+            if ob_params["distance"] != "inner_product":
+                raise ValueError(
+                    "Metric type should be 'inner_product' for sparse vector index."
+                )
+            if "sparse_index_type" in self.kwargs:
+                ob_params["type"] = self.kwargs["sparse_index_type"]
         return ob_params
 
     def param_str(self):
@@ -172,6 +186,7 @@ class IndexParam:
 
 class IndexParams:
     """Vector index parameters for MilvusLikeClient"""
+
     def __init__(self):
         self._indexes = {}
 
@@ -179,7 +194,7 @@ class IndexParams:
         self, field_name: str, index_type: VecIndexType, index_name: str, **kwargs
     ):
         """Add `IndexParam` to `IndexParams`
-        
+
         Args:
             field_name (string): vector index built on which field
             index_type (VecIndexType): vector index algorithms (Only HNSW supported)