v1.6.2

Change Log

Deprecations ⏳

• #1110 - drop python3.9 support by @joein

Fixes ⚙️

• #1132- adjust numpy versioning by @joein
• #1133 - propagate lookup_from correctly in query_points_groups by @joein
• #1134 - fix qdrant-client import in read-only systems by @holyMolyTolli

Thanks to everyone who contributed to the current release!
@holyMolyTolli @joein

v1.16.1

Change Log

Features 🪐

• #1116 - implement grpc version of cluster_collection_update by @joein
• #1123 - expose the remaining cluster methods: remove_peer, collection_cluster_info, recover_current_peer, cluster_status by @joein

Fixes 😮

• #1125 - rename common.proto to qdrant_common.proto to avoid conflicts with other libraries by @joein
• #1124 - fix local mode with persistence backward compatibility by @joein
• #1121 - add .pyi files for grpc files to help static code analyzers by @joein

v1.16.0

Change Log

Features 🪐

• #1060 - builtin BM25 support (fastembed is not longer required to generate bm25 embeddings) by @joein
• #1071 - grpc connection pooling by @JojiiOfficial
• #1088 - queryless scores is now equal to 1 by @joein
• #1077 - pass grpc credentials to grpc channels by @ivandasch @joein
• #1085 - allow passing UUID to all methods without converting to strings by @joein
• #1069 - add TextAny filter, metadata to collections, add parametrized rrf query, ACORN (qdrant changelog) by @joein
• #1106 - update grpc generator by @joein
• #1114 - expose cluster collection update operations by @joein @generall

Deprecations ⌚

• #1093 - add, query, query_batch methods deprecated in favour of upsert and query_points by @joein
• #1103 - remove deprecated methods: search, recommend, discovery, upload_records, search_batch, recommend_batch, discovery_batch, rest, init_from in create_collection, all lock-related methods by @joein
• #1104 - remove structs used only in the removed methods (#1103) by @joein

Fixes 😮

• #1081 - fix overriding prefetch limit by @joein

Thanks to everyone who contributed to the release!

@ivandasch @JojiiOfficial @joein @generall @tbung @coszio

v1.15.1

Change Log

Fixes 🚒

• #1051 - fix grpc FieldCondition conversion by @coszio
• #1052 - fix conversion for Document, Image and InferenceObject by @joein
• #1053 - fix handling cloud inference responses by @joein

Thanks to everyone who contributed to the release!
@coszio @joein

v1.15.0

Change Log

Features 🌟

• #1031 - add base64 encoding function to convert images for inference @joein
• #1044 - added MMR (maximal margin relevance) @coszio @joein @generall

Fixes 🌵

• #1030 - fix edge case in upsert and multivector updates @joein
• #1034 - relax portalocker version by @kurtqq

Thanks to everyone who contributed to the release!
@kurtqq @coszio @generall @joein

v1.14.3

Change Log

Features 🎸

• #994 - add convenience methods like list_text_models, list_image_models, etc to fastembed mixin by @joein
• #1024 - do not inspect models when cloud inference is enabled by @joein

Fixes 🌵

• #990 - fix docstrings in models to get rid of invalid escape sequence warnings in python3.12+ by @joein
• #977 - fix set payload in batch update points, fix operation conversion, set payload key handling, vector normalization in update operations by @joein
• #993 - unpickleable objects in init_options should not raise an exception by @joein
• #994 - proper exception when trying to use local inference without installed fastembed by @joein
• #996 - fix urljoin to work correctly with prefixes by @joein

v1.14.2

Change Log

Fixes 🧄

• #966 - do not raise an exception in local mode on attempting to search for sparse vectors with IDF
• #959 - fix deduplication of results when using order_by
• #967 - fix query_batch_points behaviour in local mode with some queries when limit was not provided
• #965 - fix local inference on queries with a list of prefetches, reduce the amount of inspections on vectors

Thanks to everyone who contributed to this release!
@coszio @joein

v1.14.1

Change Log

Features 💥

• #862 #907 #908 - native inference support in all QdrantClient's methods via seamless integration with Fastembed by @joein
• #915 - support server side score boosting by @coszio
• #940 - add new recommend strategy sum_score by @joein
• #896 - delete grpcio-tools from dependencies, require just protobuf by @joein
• #903 - extend strict mode settings by @generall @agourlay
• #905 - make vectors config optional in order to create collections without dense vectors by @joein
• #914 - handle too many requests exception by @tellet-q
• #955 - offset is no longer propagated into prefetches by @joein
• #946 - type hints improvements by @hh-space-invader

Fixes ⚙️

• #895 - an attempt to delete points which do not exist do not throw an error anymore in local mode by @I8dNLo
• #898 - fix incorrect check compatibility var name in warning by @abionics
• #950 - fix geo polygon support by @edjinks @joein
• #945 - allow single conditions to be passed without a list in local mode by @joein
• #910 - fix payload schema grpc conversion by @generall

Deprecations 👴

• add and query methods are now deprecated in favour of local inference usage
• #943 - pydantic >2.0,<2.2.1 is no longer supported (supported versions are v1.10.x and >=2.21) by @joein

Thanks to everyone who contributed to the current release! @edjinks @abionics @tellet-q @I8dNLo @hh-space-invader @agourlay @coszio @generall @joein

v.1.13.0

Change Log

Features 📖

• #838 #836 #842 #845 #846 #847 - local inference support: raw data like texts and images can now be passed to all the methods like upsert, query_points, etc. instead of vectors by @joein
• #849 - python3.13 support by @joein
• #835 - propagate gpu options to fastembed models by @joein
• #852 #859 #864 - improve warning coverage by @hh-space-invader @tellet-q
• #861 - make qdrant server aware of client version by @tellet-q

Fixes 🐛

• #840 - proper cleanup of the multiprocessing code in upload_collection and upload_points methods by @hh-space-invader
• #841 - add default grpc timeout by @hh-space-invader
• #876 - fix parsing url with prefix by @rovermicrover
• #875 #872 - fix dbsf corner cases by @pavelm10

Deprecations ⏳

• #848 - drop python3.8 support by @joein
• #843 - add deprecation warnings to the methods superseded by query_points (search, recommend, discover, etc.) by @hh-space-invader
• #866 - raw async grpc methods are now unavailable from the sync client by @I8dNLo

Thanks to everyone who contributed to the current release! @tellet-q @pavelm10 @hh-space-invader @I8dNLo @rovermicrover @generall @joein

v1.12.2

Change Log

Features

• #870 updated fastembed to 0.5.0
• #870 removed support for python3.8