diff --git a/.release-please-manifest.json b/.release-please-manifest.json index 26b1ce2..2b2b4fa 100644 --- a/.release-please-manifest.json +++ b/.release-please-manifest.json @@ -1,3 +1,3 @@ { - ".": "0.50.0" + ".": "0.51.0" } \ No newline at end of file diff --git a/.stats.yml b/.stats.yml index a5e2a67..2da06a1 100644 --- a/.stats.yml +++ b/.stats.yml @@ -1,4 +1,4 @@ configured_endpoints: 55 -openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/mixedbread%2Fmixedbread-ebd391dad1252eb00dd69ac50455b93bcdcd2cf0177d678e160e47f1d017287f.yml -openapi_spec_hash: 3bfd5f9eb34711238caef851aa81f5c0 +openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/mixedbread/mixedbread-046e94aa51d48644b3b896d989979ba724c63c7646569e4c607ba12653f08053.yml +openapi_spec_hash: 778f0ae64db73f363cf3543227e9c5e9 config_hash: 594a43c9cb8089f079bb9c5442646791 diff --git a/CHANGELOG.md b/CHANGELOG.md index 219fa54..546c8bf 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,5 +1,27 @@ # Changelog +## 0.51.0 (2026-05-11) + +Full Changelog: [v0.50.0...v0.51.0](https://github.com/mixedbread-ai/mixedbread-python/compare/v0.50.0...v0.51.0) + +### Features + +* **api:** api update ([6fe066e](https://github.com/mixedbread-ai/mixedbread-python/commit/6fe066e17443211a47a21f677344bf8c349e6fde)) +* **api:** api update ([eac4b8f](https://github.com/mixedbread-ai/mixedbread-python/commit/eac4b8f34275c9377765cfe62f677d5ca36cf061)) +* **api:** api update ([66aa7dc](https://github.com/mixedbread-ai/mixedbread-python/commit/66aa7dc394357766b869a3d067e4c130cdf482e8)) +* support setting headers via env ([38de0ee](https://github.com/mixedbread-ai/mixedbread-python/commit/38de0ee2781d9e27d6aecafe89f5a5b5baaed75c)) + + +### Bug Fixes + +* **client:** add missing f-string prefix in file type error message ([fe74792](https://github.com/mixedbread-ai/mixedbread-python/commit/fe74792cd81bb58bd62e58c086d6d1ee6c233c25)) +* use correct field name format for multipart file arrays ([361fcd6](https://github.com/mixedbread-ai/mixedbread-python/commit/361fcd66f8422020375b076d4c70a83c2b82241d)) + + +### Chores + +* **internal:** reformat pyproject.toml ([8802c53](https://github.com/mixedbread-ai/mixedbread-python/commit/8802c53023895e9248f03b92477c9a17013ec7b7)) + ## 0.50.0 (2026-04-23) Full Changelog: [v0.49.0...v0.50.0](https://github.com/mixedbread-ai/mixedbread-python/compare/v0.49.0...v0.50.0) diff --git a/pyproject.toml b/pyproject.toml index 1dc4f15..51d0020 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -1,6 +1,6 @@ [project] name = "mixedbread" -version = "0.50.0" +version = "0.51.0" description = "The official Python library for the Mixedbread API" dynamic = ["readme"] license = "Apache-2.0" @@ -168,7 +168,7 @@ show_error_codes = true # # We also exclude our `tests` as mypy doesn't always infer # types correctly and Pyright will still catch any type errors. -exclude = ['src/mixedbread/_files.py', '_dev/.*.py', 'tests/.*'] +exclude = ["src/mixedbread/_files.py", "_dev/.*.py", "tests/.*"] strict_equality = true implicit_reexport = true diff --git a/src/mixedbread/_client.py b/src/mixedbread/_client.py index 7391953..c8d8e32 100644 --- a/src/mixedbread/_client.py +++ b/src/mixedbread/_client.py @@ -27,6 +27,7 @@ ) from ._utils import ( is_given, + is_mapping_t, maybe_transform, get_async_library, async_maybe_transform, @@ -150,6 +151,15 @@ def __init__( except KeyError as exc: raise ValueError(f"Unknown environment: {environment}") from exc + custom_headers_env = os.environ.get("MIXEDBREAD_CUSTOM_HEADERS") + if custom_headers_env is not None: + parsed: dict[str, str] = {} + for line in custom_headers_env.split("\n"): + colon = line.find(":") + if colon >= 0: + parsed[line[:colon].strip()] = line[colon + 1 :].strip() + default_headers = {**parsed, **(default_headers if is_mapping_t(default_headers) else {})} + super().__init__( version=__version__, base_url=base_url, @@ -547,6 +557,15 @@ def __init__( except KeyError as exc: raise ValueError(f"Unknown environment: {environment}") from exc + custom_headers_env = os.environ.get("MIXEDBREAD_CUSTOM_HEADERS") + if custom_headers_env is not None: + parsed: dict[str, str] = {} + for line in custom_headers_env.split("\n"): + colon = line.find(":") + if colon >= 0: + parsed[line[:colon].strip()] = line[colon + 1 :].strip() + default_headers = {**parsed, **(default_headers if is_mapping_t(default_headers) else {})} + super().__init__( version=__version__, base_url=base_url, diff --git a/src/mixedbread/_files.py b/src/mixedbread/_files.py index 4df9de4..f557c29 100644 --- a/src/mixedbread/_files.py +++ b/src/mixedbread/_files.py @@ -99,7 +99,7 @@ async def async_to_httpx_files(files: RequestFiles | None) -> HttpxRequestFiles elif is_sequence_t(files): files = [(key, await _async_transform_file(file)) for key, file in files] else: - raise TypeError("Unexpected file type input {type(files)}, expected mapping or sequence") + raise TypeError(f"Unexpected file type input {type(files)}, expected mapping or sequence") return files diff --git a/src/mixedbread/_qs.py b/src/mixedbread/_qs.py index de8c99b..4127c19 100644 --- a/src/mixedbread/_qs.py +++ b/src/mixedbread/_qs.py @@ -2,17 +2,13 @@ from typing import Any, List, Tuple, Union, Mapping, TypeVar from urllib.parse import parse_qs, urlencode -from typing_extensions import Literal, get_args +from typing_extensions import get_args -from ._types import NotGiven, not_given +from ._types import NotGiven, ArrayFormat, NestedFormat, not_given from ._utils import flatten _T = TypeVar("_T") - -ArrayFormat = Literal["comma", "repeat", "indices", "brackets"] -NestedFormat = Literal["dots", "brackets"] - PrimitiveData = Union[str, int, float, bool, None] # this should be Data = Union[PrimitiveData, "List[Data]", "Tuple[Data]", "Mapping[str, Data]"] # https://github.com/microsoft/pyright/issues/3555 diff --git a/src/mixedbread/_types.py b/src/mixedbread/_types.py index fa2086c..39dc5bd 100644 --- a/src/mixedbread/_types.py +++ b/src/mixedbread/_types.py @@ -47,6 +47,9 @@ ModelT = TypeVar("ModelT", bound=pydantic.BaseModel) _T = TypeVar("_T") +ArrayFormat = Literal["comma", "repeat", "indices", "brackets"] +NestedFormat = Literal["dots", "brackets"] + # Approximates httpx internal ProxiesTypes and RequestFiles types # while adding support for `PathLike` instances diff --git a/src/mixedbread/_utils/_utils.py b/src/mixedbread/_utils/_utils.py index 771859f..199cd23 100644 --- a/src/mixedbread/_utils/_utils.py +++ b/src/mixedbread/_utils/_utils.py @@ -17,11 +17,11 @@ ) from pathlib import Path from datetime import date, datetime -from typing_extensions import TypeGuard +from typing_extensions import TypeGuard, get_args import sniffio -from .._types import Omit, NotGiven, FileTypes, HeadersLike +from .._types import Omit, NotGiven, FileTypes, ArrayFormat, HeadersLike _T = TypeVar("_T") _TupleT = TypeVar("_TupleT", bound=Tuple[object, ...]) @@ -40,25 +40,45 @@ def extract_files( query: Mapping[str, object], *, paths: Sequence[Sequence[str]], + array_format: ArrayFormat = "brackets", ) -> list[tuple[str, FileTypes]]: """Recursively extract files from the given dictionary based on specified paths. A path may look like this ['foo', 'files', '', 'data']. + ``array_format`` controls how ```` segments contribute to the emitted + field name. Supported values: ``"brackets"`` (``foo[]``), ``"repeat"`` and + ``"comma"`` (``foo``), ``"indices"`` (``foo[0]``, ``foo[1]``). + Note: this mutates the given dictionary. """ files: list[tuple[str, FileTypes]] = [] for path in paths: - files.extend(_extract_items(query, path, index=0, flattened_key=None)) + files.extend(_extract_items(query, path, index=0, flattened_key=None, array_format=array_format)) return files +def _array_suffix(array_format: ArrayFormat, array_index: int) -> str: + if array_format == "brackets": + return "[]" + if array_format == "indices": + return f"[{array_index}]" + if array_format == "repeat" or array_format == "comma": + # Both repeat the bare field name for each file part; there is no + # meaningful way to comma-join binary parts. + return "" + raise NotImplementedError( + f"Unknown array_format value: {array_format}, choose from {', '.join(get_args(ArrayFormat))}" + ) + + def _extract_items( obj: object, path: Sequence[str], *, index: int, flattened_key: str | None, + array_format: ArrayFormat, ) -> list[tuple[str, FileTypes]]: try: key = path[index] @@ -75,9 +95,11 @@ def _extract_items( if is_list(obj): files: list[tuple[str, FileTypes]] = [] - for entry in obj: - assert_is_file_content(entry, key=flattened_key + "[]" if flattened_key else "") - files.append((flattened_key + "[]", cast(FileTypes, entry))) + for array_index, entry in enumerate(obj): + suffix = _array_suffix(array_format, array_index) + emitted_key = (flattened_key + suffix) if flattened_key else suffix + assert_is_file_content(entry, key=emitted_key) + files.append((emitted_key, cast(FileTypes, entry))) return files assert_is_file_content(obj, key=flattened_key) @@ -106,6 +128,7 @@ def _extract_items( path, index=index, flattened_key=flattened_key, + array_format=array_format, ) elif is_list(obj): if key != "": @@ -117,9 +140,12 @@ def _extract_items( item, path, index=index, - flattened_key=flattened_key + "[]" if flattened_key is not None else "[]", + flattened_key=( + (flattened_key if flattened_key is not None else "") + _array_suffix(array_format, array_index) + ), + array_format=array_format, ) - for item in obj + for array_index, item in enumerate(obj) ] ) diff --git a/src/mixedbread/_version.py b/src/mixedbread/_version.py index 84dd557..b259f3b 100644 --- a/src/mixedbread/_version.py +++ b/src/mixedbread/_version.py @@ -1,4 +1,4 @@ # File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details. __title__ = "mixedbread" -__version__ = "0.50.0" # x-release-please-version +__version__ = "0.51.0" # x-release-please-version diff --git a/src/mixedbread/resources/stores/stores.py b/src/mixedbread/resources/stores/stores.py index 20271dc..c28665f 100644 --- a/src/mixedbread/resources/stores/stores.py +++ b/src/mixedbread/resources/stores/stores.py @@ -368,6 +368,9 @@ def metadata_facets( file_ids: Union[Iterable[object], SequenceNotStr[str], None] | Omit = omit, search_options: StoreChunkSearchOptionsParam | Omit = omit, facets: Optional[SequenceNotStr[str]] | Omit = omit, + max_fields: int | Omit = omit, + max_values_per_field: int | Omit = omit, + max_files: int | Omit = omit, # Use the following arguments if you need to pass additional parameters to the API that aren't available via kwargs. # The extra values given here take precedence over values defined on the client or passed to this method. extra_headers: Headers | None = None, @@ -393,6 +396,12 @@ def metadata_facets( facets: Optional list of facets to return. Use dot for nested fields. + max_fields: Maximum number of distinct metadata fields (keys) to return. + + max_values_per_field: Maximum number of distinct values returned per field, ranked by count. + + max_files: Maximum number of store files scanned to compute facets. + extra_headers: Send extra headers extra_query: Add additional query parameters to the request @@ -412,6 +421,9 @@ def metadata_facets( "file_ids": file_ids, "search_options": search_options, "facets": facets, + "max_fields": max_fields, + "max_values_per_field": max_values_per_field, + "max_files": max_files, }, store_metadata_facets_params.StoreMetadataFacetsParams, ), @@ -897,6 +909,9 @@ async def metadata_facets( file_ids: Union[Iterable[object], SequenceNotStr[str], None] | Omit = omit, search_options: StoreChunkSearchOptionsParam | Omit = omit, facets: Optional[SequenceNotStr[str]] | Omit = omit, + max_fields: int | Omit = omit, + max_values_per_field: int | Omit = omit, + max_files: int | Omit = omit, # Use the following arguments if you need to pass additional parameters to the API that aren't available via kwargs. # The extra values given here take precedence over values defined on the client or passed to this method. extra_headers: Headers | None = None, @@ -922,6 +937,12 @@ async def metadata_facets( facets: Optional list of facets to return. Use dot for nested fields. + max_fields: Maximum number of distinct metadata fields (keys) to return. + + max_values_per_field: Maximum number of distinct values returned per field, ranked by count. + + max_files: Maximum number of store files scanned to compute facets. + extra_headers: Send extra headers extra_query: Add additional query parameters to the request @@ -941,6 +962,9 @@ async def metadata_facets( "file_ids": file_ids, "search_options": search_options, "facets": facets, + "max_fields": max_fields, + "max_values_per_field": max_values_per_field, + "max_files": max_files, }, store_metadata_facets_params.StoreMetadataFacetsParams, ), diff --git a/src/mixedbread/types/agentic_search_config_param.py b/src/mixedbread/types/agentic_search_config_param.py index 8a01d8e..3ffe184 100644 --- a/src/mixedbread/types/agentic_search_config_param.py +++ b/src/mixedbread/types/agentic_search_config_param.py @@ -3,7 +3,7 @@ from __future__ import annotations from typing import Optional -from typing_extensions import TypedDict +from typing_extensions import Literal, TypedDict __all__ = ["AgenticSearchConfigParam"] @@ -17,6 +17,16 @@ class AgenticSearchConfigParam(TypedDict, total=False): queries_per_round: int """Maximum queries per round""" + strict_top_k: bool + """Whether the final retrieved chunk list must provide exactly top_k ranked chunks""" + + media_content: Literal["auto", "never", "always"] + """Controls when retrieved image content is provided to the agent. + + `auto` sends images only when no OCR text or summary is available, `never` + disables image content, and `always` sends image content when available. + """ + instructions: Optional[str] """ Additional custom instructions (followed only when not in conflict with existing diff --git a/src/mixedbread/types/scored_audio_url_input_chunk.py b/src/mixedbread/types/scored_audio_url_input_chunk.py index e12b2d9..d5d29d8 100644 --- a/src/mixedbread/types/scored_audio_url_input_chunk.py +++ b/src/mixedbread/types/scored_audio_url_input_chunk.py @@ -68,6 +68,9 @@ class ScoredAudioURLInputChunk(BaseModel): transcription: Optional[str] = None """speech recognition (sr) text of the audio""" + summary: Optional[str] = None + """summary of the audio""" + audio_url: Optional[AudioURL] = None """Model for audio URL validation.""" diff --git a/src/mixedbread/types/store_metadata_facets_params.py b/src/mixedbread/types/store_metadata_facets_params.py index 64aa124..3e098ad 100644 --- a/src/mixedbread/types/store_metadata_facets_params.py +++ b/src/mixedbread/types/store_metadata_facets_params.py @@ -34,6 +34,15 @@ class StoreMetadataFacetsParams(TypedDict, total=False): facets: Optional[SequenceNotStr[str]] """Optional list of facets to return. Use dot for nested fields.""" + max_fields: int + """Maximum number of distinct metadata fields (keys) to return.""" + + max_values_per_field: int + """Maximum number of distinct values returned per field, ranked by count.""" + + max_files: int + """Maximum number of store files scanned to compute facets.""" + FiltersUnionMember2: TypeAlias = Union["SearchFilter", SearchFilterCondition] diff --git a/src/mixedbread/types/store_search_response.py b/src/mixedbread/types/store_search_response.py index 6ab757c..ad42424 100644 --- a/src/mixedbread/types/store_search_response.py +++ b/src/mixedbread/types/store_search_response.py @@ -1,8 +1,11 @@ # File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details. -from typing import List, Union, Optional +import builtins +from typing import TYPE_CHECKING, Dict, List, Union, Optional from typing_extensions import Literal, Annotated, TypeAlias +from pydantic import Field as FieldInfo + from .._utils import PropertyInfo from .._models import BaseModel from .scored_text_input_chunk import ScoredTextInputChunk @@ -24,3 +27,15 @@ class StoreSearchResponse(BaseModel): data: List[Data] """The list of scored store file chunks""" + + if TYPE_CHECKING: + # Some versions of Pydantic <2.8.0 have a bug and don’t allow assigning a + # value to this field, so for compatibility we avoid doing it at runtime. + __pydantic_extra__: Dict[str, builtins.object] = FieldInfo(init=False) # pyright: ignore[reportIncompatibleVariableOverride] + + # Stub to indicate that arbitrary properties are accepted. + # To access properties that are not valid identifiers you can use `getattr`, e.g. + # `getattr(obj, '$type')` + def __getattr__(self, attr: str) -> builtins.object: ... + else: + __pydantic_extra__: Dict[str, builtins.object] diff --git a/src/mixedbread/types/stores/audio_url_input_chunk.py b/src/mixedbread/types/stores/audio_url_input_chunk.py index 79cdd3c..830a577 100644 --- a/src/mixedbread/types/stores/audio_url_input_chunk.py +++ b/src/mixedbread/types/stores/audio_url_input_chunk.py @@ -50,6 +50,9 @@ class AudioURLInputChunk(BaseModel): transcription: Optional[str] = None """speech recognition (sr) text of the audio""" + summary: Optional[str] = None + """summary of the audio""" + audio_url: Optional[AudioURL] = None """Model for audio URL validation.""" diff --git a/tests/api_resources/test_stores.py b/tests/api_resources/test_stores.py index 3998817..cb69698 100644 --- a/tests/api_resources/test_stores.py +++ b/tests/api_resources/test_stores.py @@ -263,6 +263,9 @@ def test_method_metadata_facets_with_all_params(self, client: Mixedbread) -> Non "apply_search_rules": True, }, facets=["string"], + max_fields=1, + max_values_per_field=1, + max_files=1, ) assert_matches_type(StoreMetadataFacetsResponse, store, path=["response"]) @@ -652,6 +655,9 @@ async def test_method_metadata_facets_with_all_params(self, async_client: AsyncM "apply_search_rules": True, }, facets=["string"], + max_fields=1, + max_values_per_field=1, + max_files=1, ) assert_matches_type(StoreMetadataFacetsResponse, store, path=["response"]) diff --git a/tests/test_extract_files.py b/tests/test_extract_files.py index 4a25251..06c03a7 100644 --- a/tests/test_extract_files.py +++ b/tests/test_extract_files.py @@ -4,7 +4,7 @@ import pytest -from mixedbread._types import FileTypes +from mixedbread._types import FileTypes, ArrayFormat from mixedbread._utils import extract_files @@ -37,10 +37,7 @@ def test_multiple_files() -> None: def test_top_level_file_array() -> None: query = {"files": [b"file one", b"file two"], "title": "hello"} - assert extract_files(query, paths=[["files", ""]]) == [ - ("files[]", b"file one"), - ("files[]", b"file two"), - ] + assert extract_files(query, paths=[["files", ""]]) == [("files[]", b"file one"), ("files[]", b"file two")] assert query == {"title": "hello"} @@ -71,3 +68,24 @@ def test_ignores_incorrect_paths( expected: list[tuple[str, FileTypes]], ) -> None: assert extract_files(query, paths=paths) == expected + + +@pytest.mark.parametrize( + "array_format,expected_top_level,expected_nested", + [ + ("brackets", [("files[]", b"a"), ("files[]", b"b")], [("items[][file]", b"a"), ("items[][file]", b"b")]), + ("repeat", [("files", b"a"), ("files", b"b")], [("items[file]", b"a"), ("items[file]", b"b")]), + ("comma", [("files", b"a"), ("files", b"b")], [("items[file]", b"a"), ("items[file]", b"b")]), + ("indices", [("files[0]", b"a"), ("files[1]", b"b")], [("items[0][file]", b"a"), ("items[1][file]", b"b")]), + ], +) +def test_array_format_controls_file_field_names( + array_format: ArrayFormat, + expected_top_level: list[tuple[str, FileTypes]], + expected_nested: list[tuple[str, FileTypes]], +) -> None: + top_level = {"files": [b"a", b"b"]} + assert extract_files(top_level, paths=[["files", ""]], array_format=array_format) == expected_top_level + + nested = {"items": [{"file": b"a"}, {"file": b"b"}]} + assert extract_files(nested, paths=[["items", "", "file"]], array_format=array_format) == expected_nested diff --git a/tests/test_files.py b/tests/test_files.py index 8e51c40..71c9c28 100644 --- a/tests/test_files.py +++ b/tests/test_files.py @@ -131,7 +131,7 @@ def test_extract_files_does_not_mutate_original_nested_array_path(self) -> None: copied = deepcopy_with_paths(original, [["items", "", "file"]]) extracted = extract_files(copied, paths=[["items", "", "file"]]) - assert extracted == [("items[][file]", file1), ("items[][file]", file2)] + assert [entry for _, entry in extracted] == [file1, file2] assert original == { "items": [ {"file": file1, "extra": 1},