public class CreateDocumentCollectionRequest
extends com.aliyun.tea.TeaModel
| 限定符和类型 | 字段和说明 |
|---|---|
String |
collection
This parameter is required.
|
String |
DBInstanceId
This parameter is required.
|
String |
embeddingModel
The vectorization algorithm.
|
Integer |
externalStorage
Specifies whether to use the memory mapping technology to create HNSW indexes.
|
String |
fullTextRetrievalFields
example:
title,page
|
Integer |
hnswM
The maximum number of neighbors for the Hierarchical Navigable Small World (HNSW) algorithm.
|
String |
managerAccount
This parameter is required.
|
String |
managerAccountPassword
This parameter is required.
|
String |
metadata
The metadata of the vector data, which is a JSON string in the MAP format.
|
String |
metrics
example:
cosine
|
String |
namespace
example:
mynamespace
|
Long |
ownerId |
String |
parser
example:
zh_cn
|
Integer |
pqEnable
Specifies whether to enable the product quantization (PQ) feature for index acceleration.
|
String |
regionId
This parameter is required.
|
| 构造器和说明 |
|---|
CreateDocumentCollectionRequest() |
@NameInMap(value="Collection") public String collection
This parameter is required.
example:document
@NameInMap(value="DBInstanceId") public String DBInstanceId
This parameter is required.
example:gp-xxxxxxxxx
@NameInMap(value="EmbeddingModel") public String embeddingModel
The vectorization algorithm.
Valid values:
text-embedding-v1: the algorithm that produces 1536-dimensional vectors.
text-embedding-v2: the algorithm that produces 1536-dimensional vectors.
text2vec: the algorithm that produces 1024-dimensional vectors.
m3e-base: the algorithm that produces 768-dimensional vectors.
m3e-small: the algorithm that produces 512-dimensional vectors.
multimodal-embedding-one-peace-v1: the image vectorization algorithm that produces 1536-dimensional vectors.
clip-vit-b-32: the image vectorization algorithm that uses the Contrastive Language-Image Pre-Training (CLIP) ViT-B/32 model and produces 512-dimensional vectors.
clip-vit-b-16: the image vectorization algorithm that uses the CLIP ViT-B/16 model and produces 512-dimensional vectors.
clip-vit-l-14: the image vectorization algorithm that uses the CLIP ViT-L/14 model and produces 768-dimensional vectors.
clip-vit-l-14-336px: the image vectorization algorithm that uses the CLIP ViT-L/14@336px model and produces 768-dimensional vectors.
clip-rn50: the image vectorization algorithm that uses the CLIP RN50 model and produces 1024-dimensional vectors.
clip-rn101: the image vectorization algorithm that uses the CLIP RN101 model and produces 512-dimensional vectors.
clip-rn50x4: the image vectorization algorithm that uses the CLIP RN50x4 model and produces 640-dimensional vectors.
clip-rn50x16: the image vectorization algorithm that uses the CLIP RN50x16 model and produces 768-dimensional vectors.
clip-rn50x64: the image vectorization algorithm that uses the CLIP RN50x64 model and produces 1024-dimensional vectors.
text-embedding-v1
@NameInMap(value="ExternalStorage") public Integer externalStorage
Specifies whether to use the memory mapping technology to create HNSW indexes. Valid values: 0 and 1. Default value: 0. We recommend that you set the value to 1 in scenarios that require upload speed but not data deletion.
0: uses segmented paging storage to create indexes. This method uses the shared buffer of PostgreSQL for caching and supports the delete and update operations.
1: uses the memory mapping technology to create indexes. This method does not support the delete or update operation.
0
@NameInMap(value="FullTextRetrievalFields") public String fullTextRetrievalFields
title,page
@NameInMap(value="HnswM") public Integer hnswM
The maximum number of neighbors for the Hierarchical Navigable Small World (HNSW) algorithm. Valid values: 1 to 1000. In most cases, this parameter is automatically configured based on the value of the Dimension parameter. You do not need to configure this parameter.
We recommend that you configure this parameter based on the value of the Dimension parameter.
If you set Dimension to a value less than or equal to 384, set the value of HnswM to 16.
If you set Dimension to a value greater than 384 and less than or equal to 768, set the value of HnswM to 32.
If you set Dimension to a value greater than 768 and less than or equal to 1024, set the value of HnswM to 64.
If you set Dimension to a value greater than 1024, set the value of HnswM to 128.
64
@NameInMap(value="ManagerAccount") public String managerAccount
This parameter is required.
example:testaccount
@NameInMap(value="ManagerAccountPassword") public String managerAccountPassword
This parameter is required.
example:testpassword
@NameInMap(value="Metadata") public String metadata
The metadata of the vector data, which is a JSON string in the MAP format. The key specifies the field name, and the value specifies the data type.
- For information about the supported data types, see Data types.
- The money data type is not supported.
example:Warning: Reserved fields such as id, vector, doc_name, content, loader_metadata, source, and to_tsvector cannot be used.
{"title":"text","page":"int"}
@NameInMap(value="Metrics") public String metrics
cosine
@NameInMap(value="Namespace") public String namespace
mynamespace
@NameInMap(value="OwnerId") public Long ownerId
@NameInMap(value="Parser") public String parser
zh_cn
@NameInMap(value="PqEnable") public Integer pqEnable
Specifies whether to enable the product quantization (PQ) feature for index acceleration. We recommend that you enable this feature for more than 500,000 rows of data. Valid values:
0
@NameInMap(value="RegionId") public String regionId
This parameter is required.
example:cn-hangzhou
public static CreateDocumentCollectionRequest build(Map<String,?> map) throws Exception
Exceptionpublic CreateDocumentCollectionRequest setCollection(String collection)
public String getCollection()
public CreateDocumentCollectionRequest setDBInstanceId(String DBInstanceId)
public String getDBInstanceId()
public CreateDocumentCollectionRequest setEmbeddingModel(String embeddingModel)
public String getEmbeddingModel()
public CreateDocumentCollectionRequest setExternalStorage(Integer externalStorage)
public Integer getExternalStorage()
public CreateDocumentCollectionRequest setFullTextRetrievalFields(String fullTextRetrievalFields)
public String getFullTextRetrievalFields()
public CreateDocumentCollectionRequest setHnswM(Integer hnswM)
public Integer getHnswM()
public CreateDocumentCollectionRequest setManagerAccount(String managerAccount)
public String getManagerAccount()
public CreateDocumentCollectionRequest setManagerAccountPassword(String managerAccountPassword)
public String getManagerAccountPassword()
public CreateDocumentCollectionRequest setMetadata(String metadata)
public String getMetadata()
public CreateDocumentCollectionRequest setMetrics(String metrics)
public String getMetrics()
public CreateDocumentCollectionRequest setNamespace(String namespace)
public String getNamespace()
public CreateDocumentCollectionRequest setOwnerId(Long ownerId)
public Long getOwnerId()
public CreateDocumentCollectionRequest setParser(String parser)
public String getParser()
public CreateDocumentCollectionRequest setPqEnable(Integer pqEnable)
public Integer getPqEnable()
public CreateDocumentCollectionRequest setRegionId(String regionId)
public String getRegionId()
Copyright © 2024. All rights reserved.