v1.5.1

fix: replace large file-sized dictionaries in Git LFS with URLs to enable the usage of NewBM25Encoder with language-specific dictionaries when customers refer to the "tcvdbtext" directory in this Go package. Also support to load dictionary in /tmp/tencent/vectordatabase/data

v1.5.0

• feat: support binary vector
• feat: support to return documents' count when using Count function
• feat: support to delete documents with limit
• feat: add params terminateAfter and cutoffFrequency, when hybrid search
• feat: support to hybrid search by text in embedding collection
• feat: support to modify vector indexes

v1.4.9

doc: add comments for tcvectordb package
fix: ExistsDatabase checks the existence of a specific database, including all database types AI_DB or BASE_DB.

v1.4.8

• feat: CreateCollectionView/LoadAndSplitText interfaces add parsingType params, which can support different parsing type (AlgorithmParsing/VisionModelParsing) for files

v1.4.7

fix: remove buildtag to support windows platform compilation

v1.4.6

feat: add ModelName to support setting a new model string directly when creating an embedding coll

v1.4.5

feat: change jieba go dependency to support cross-platform compilation

v1.4.4

feat: support /index/add interface

v1.4.3

fix: compatible processing unsupported windows jieba tokenizer case

v1.4.2

fix: upgrade gojieba dependency version to v1.4.3