llama-cloud 0.1.15__py3-none-any.whl → 0.1.17__py3-none-any.whl

llama_cloud/resources/retrievers/client.py

@@ -12,6 +12,7 @@ from ...errors.unprocessable_entity_error import UnprocessableEntityError
 from ...types.composite_retrieval_mode import CompositeRetrievalMode
 from ...types.composite_retrieval_result import CompositeRetrievalResult
 from ...types.http_validation_error import HttpValidationError
+from ...types.re_rank_config import ReRankConfig
 from ...types.retriever import Retriever
 from ...types.retriever_create import RetrieverCreate
 from ...types.retriever_pipeline import RetrieverPipeline
@@ -296,6 +297,7 @@ class RetrieversClient:
         *,
         mode: typing.Optional[CompositeRetrievalMode] = OMIT,
         rerank_top_n: typing.Optional[int] = OMIT,
+        rerank_config: typing.Optional[ReRankConfig] = OMIT,
         query: str,
     ) -> CompositeRetrievalResult:
         """
@@ -306,11 +308,13 @@ class RetrieversClient:
 
             - mode: typing.Optional[CompositeRetrievalMode]. The mode of composite retrieval.
 
-            - rerank_top_n: typing.Optional[int]. The number of nodes to retrieve after reranking over retrieved nodes from all retrieval tools.
+            - rerank_top_n: typing.Optional[int].
+
+            - rerank_config: typing.Optional[ReRankConfig]. The rerank configuration for composite retrieval.
 
             - query: str. The query to retrieve against.
         ---
-        from llama_cloud import CompositeRetrievalMode
+        from llama_cloud import CompositeRetrievalMode, ReRankConfig, ReRankerType
         from llama_cloud.client import LlamaCloud
 
         client = LlamaCloud(
@@ -319,6 +323,9 @@ class RetrieversClient:
         client.retrievers.retrieve(
             retriever_id="string",
             mode=CompositeRetrievalMode.ROUTING,
+            rerank_config=ReRankConfig(
+                type=ReRankerType.SYSTEM_DEFAULT,
+            ),
             query="string",
         )
         """
@@ -327,6 +334,8 @@ class RetrieversClient:
             _request["mode"] = mode
         if rerank_top_n is not OMIT:
             _request["rerank_top_n"] = rerank_top_n
+        if rerank_config is not OMIT:
+            _request["rerank_config"] = rerank_config
         _response = self._client_wrapper.httpx_client.request(
             "POST",
             urllib.parse.urljoin(
@@ -346,6 +355,76 @@ class RetrieversClient:
             raise ApiError(status_code=_response.status_code, body=_response.text)
         raise ApiError(status_code=_response.status_code, body=_response_json)
 
+    def direct_retrieve(
+        self,
+        *,
+        project_id: typing.Optional[str] = None,
+        organization_id: typing.Optional[str] = None,
+        mode: typing.Optional[CompositeRetrievalMode] = OMIT,
+        rerank_top_n: typing.Optional[int] = OMIT,
+        rerank_config: typing.Optional[ReRankConfig] = OMIT,
+        query: str,
+        pipelines: typing.Optional[typing.List[RetrieverPipeline]] = OMIT,
+    ) -> CompositeRetrievalResult:
+        """
+        Retrieve data using specified pipelines without creating a persistent retriever.
+
+        Parameters:
+            - project_id: typing.Optional[str].
+
+            - organization_id: typing.Optional[str].
+
+            - mode: typing.Optional[CompositeRetrievalMode]. The mode of composite retrieval.
+
+            - rerank_top_n: typing.Optional[int].
+
+            - rerank_config: typing.Optional[ReRankConfig]. The rerank configuration for composite retrieval.
+
+            - query: str. The query to retrieve against.
+
+            - pipelines: typing.Optional[typing.List[RetrieverPipeline]]. The pipelines to use for retrieval.
+        ---
+        from llama_cloud import CompositeRetrievalMode, ReRankConfig, ReRankerType
+        from llama_cloud.client import LlamaCloud
+
+        client = LlamaCloud(
+            token="YOUR_TOKEN",
+        )
+        client.retrievers.direct_retrieve(
+            mode=CompositeRetrievalMode.ROUTING,
+            rerank_config=ReRankConfig(
+                type=ReRankerType.SYSTEM_DEFAULT,
+            ),
+            query="string",
+        )
+        """
+        _request: typing.Dict[str, typing.Any] = {"query": query}
+        if mode is not OMIT:
+            _request["mode"] = mode
+        if rerank_top_n is not OMIT:
+            _request["rerank_top_n"] = rerank_top_n
+        if rerank_config is not OMIT:
+            _request["rerank_config"] = rerank_config
+        if pipelines is not OMIT:
+            _request["pipelines"] = pipelines
+        _response = self._client_wrapper.httpx_client.request(
+            "POST",
+            urllib.parse.urljoin(f"{self._client_wrapper.get_base_url()}/", "api/v1/retrievers/retrieve"),
+            params=remove_none_from_dict({"project_id": project_id, "organization_id": organization_id}),
+            json=jsonable_encoder(_request),
+            headers=self._client_wrapper.get_headers(),
+            timeout=60,
+        )
+        if 200 <= _response.status_code < 300:
+            return pydantic.parse_obj_as(CompositeRetrievalResult, _response.json())  # type: ignore
+        if _response.status_code == 422:
+            raise UnprocessableEntityError(pydantic.parse_obj_as(HttpValidationError, _response.json()))  # type: ignore
+        try:
+            _response_json = _response.json()
+        except JSONDecodeError:
+            raise ApiError(status_code=_response.status_code, body=_response.text)
+        raise ApiError(status_code=_response.status_code, body=_response_json)
+
 
 class AsyncRetrieversClient:
     def __init__(self, *, client_wrapper: AsyncClientWrapper):
@@ -615,6 +694,7 @@ class AsyncRetrieversClient:
         *,
         mode: typing.Optional[CompositeRetrievalMode] = OMIT,
         rerank_top_n: typing.Optional[int] = OMIT,
+        rerank_config: typing.Optional[ReRankConfig] = OMIT,
         query: str,
     ) -> CompositeRetrievalResult:
         """
@@ -625,11 +705,13 @@ class AsyncRetrieversClient:
 
             - mode: typing.Optional[CompositeRetrievalMode]. The mode of composite retrieval.
 
-            - rerank_top_n: typing.Optional[int]. The number of nodes to retrieve after reranking over retrieved nodes from all retrieval tools.
+            - rerank_top_n: typing.Optional[int].
+
+            - rerank_config: typing.Optional[ReRankConfig]. The rerank configuration for composite retrieval.
 
             - query: str. The query to retrieve against.
         ---
-        from llama_cloud import CompositeRetrievalMode
+        from llama_cloud import CompositeRetrievalMode, ReRankConfig, ReRankerType
         from llama_cloud.client import AsyncLlamaCloud
 
         client = AsyncLlamaCloud(
@@ -638,6 +720,9 @@ class AsyncRetrieversClient:
         await client.retrievers.retrieve(
             retriever_id="string",
             mode=CompositeRetrievalMode.ROUTING,
+            rerank_config=ReRankConfig(
+                type=ReRankerType.SYSTEM_DEFAULT,
+            ),
             query="string",
         )
         """
@@ -646,6 +731,8 @@ class AsyncRetrieversClient:
             _request["mode"] = mode
         if rerank_top_n is not OMIT:
             _request["rerank_top_n"] = rerank_top_n
+        if rerank_config is not OMIT:
+            _request["rerank_config"] = rerank_config
         _response = await self._client_wrapper.httpx_client.request(
             "POST",
             urllib.parse.urljoin(
@@ -664,3 +751,73 @@ class AsyncRetrieversClient:
         except JSONDecodeError:
             raise ApiError(status_code=_response.status_code, body=_response.text)
         raise ApiError(status_code=_response.status_code, body=_response_json)
+
+    async def direct_retrieve(
+        self,
+        *,
+        project_id: typing.Optional[str] = None,
+        organization_id: typing.Optional[str] = None,
+        mode: typing.Optional[CompositeRetrievalMode] = OMIT,
+        rerank_top_n: typing.Optional[int] = OMIT,
+        rerank_config: typing.Optional[ReRankConfig] = OMIT,
+        query: str,
+        pipelines: typing.Optional[typing.List[RetrieverPipeline]] = OMIT,
+    ) -> CompositeRetrievalResult:
+        """
+        Retrieve data using specified pipelines without creating a persistent retriever.
+
+        Parameters:
+            - project_id: typing.Optional[str].
+
+            - organization_id: typing.Optional[str].
+
+            - mode: typing.Optional[CompositeRetrievalMode]. The mode of composite retrieval.
+
+            - rerank_top_n: typing.Optional[int].
+
+            - rerank_config: typing.Optional[ReRankConfig]. The rerank configuration for composite retrieval.
+
+            - query: str. The query to retrieve against.
+
+            - pipelines: typing.Optional[typing.List[RetrieverPipeline]]. The pipelines to use for retrieval.
+        ---
+        from llama_cloud import CompositeRetrievalMode, ReRankConfig, ReRankerType
+        from llama_cloud.client import AsyncLlamaCloud
+
+        client = AsyncLlamaCloud(
+            token="YOUR_TOKEN",
+        )
+        await client.retrievers.direct_retrieve(
+            mode=CompositeRetrievalMode.ROUTING,
+            rerank_config=ReRankConfig(
+                type=ReRankerType.SYSTEM_DEFAULT,
+            ),
+            query="string",
+        )
+        """
+        _request: typing.Dict[str, typing.Any] = {"query": query}
+        if mode is not OMIT:
+            _request["mode"] = mode
+        if rerank_top_n is not OMIT:
+            _request["rerank_top_n"] = rerank_top_n
+        if rerank_config is not OMIT:
+            _request["rerank_config"] = rerank_config
+        if pipelines is not OMIT:
+            _request["pipelines"] = pipelines
+        _response = await self._client_wrapper.httpx_client.request(
+            "POST",
+            urllib.parse.urljoin(f"{self._client_wrapper.get_base_url()}/", "api/v1/retrievers/retrieve"),
+            params=remove_none_from_dict({"project_id": project_id, "organization_id": organization_id}),
+            json=jsonable_encoder(_request),
+            headers=self._client_wrapper.get_headers(),
+            timeout=60,
+        )
+        if 200 <= _response.status_code < 300:
+            return pydantic.parse_obj_as(CompositeRetrievalResult, _response.json())  # type: ignore
+        if _response.status_code == 422:
+            raise UnprocessableEntityError(pydantic.parse_obj_as(HttpValidationError, _response.json()))  # type: ignore
+        try:
+            _response_json = _response.json()
+        except JSONDecodeError:
+            raise ApiError(status_code=_response.status_code, body=_response.text)
+        raise ApiError(status_code=_response.status_code, body=_response_json)

llama_cloud/types/__init__.py

@@ -103,15 +103,7 @@ from .embedding_model_config_update_embedding_config import (
     EmbeddingModelConfigUpdateEmbeddingConfig_OpenaiEmbedding,
     EmbeddingModelConfigUpdateEmbeddingConfig_VertexaiEmbedding,
 )
-from .eval_dataset import EvalDataset
-from .eval_dataset_job_params import EvalDatasetJobParams
-from .eval_dataset_job_record import EvalDatasetJobRecord
 from .eval_execution_params import EvalExecutionParams
-from .eval_execution_params_override import EvalExecutionParamsOverride
-from .eval_metric import EvalMetric
-from .eval_question import EvalQuestion
-from .eval_question_create import EvalQuestionCreate
-from .eval_question_result import EvalQuestionResult
 from .extract_agent import ExtractAgent
 from .extract_agent_create import ExtractAgentCreate
 from .extract_agent_create_data_schema import ExtractAgentCreateDataSchema
@@ -178,9 +170,6 @@ from .llama_parse_supported_file_extensions import LlamaParseSupportedFileExtens
 from .llm import Llm
 from .llm_model_data import LlmModelData
 from .llm_parameters import LlmParameters
-from .local_eval import LocalEval
-from .local_eval_results import LocalEvalResults
-from .local_eval_sets import LocalEvalSets
 from .managed_ingestion_status import ManagedIngestionStatus
 from .managed_ingestion_status_response import ManagedIngestionStatusResponse
 from .markdown_element_node_parser import MarkdownElementNodeParser
@@ -191,7 +180,6 @@ from .metadata_filter import MetadataFilter
 from .metadata_filter_value import MetadataFilterValue
 from .metadata_filters import MetadataFilters
 from .metadata_filters_filters_item import MetadataFiltersFiltersItem
-from .metric_result import MetricResult
 from .node_parser import NodeParser
 from .node_relationship import NodeRelationship
 from .none_chunking_config import NoneChunkingConfig
@@ -259,6 +247,7 @@ from .pipeline_file_create_custom_metadata_value import PipelineFileCreateCustom
 from .pipeline_file_custom_metadata_value import PipelineFileCustomMetadataValue
 from .pipeline_file_permission_info_value import PipelineFilePermissionInfoValue
 from .pipeline_file_resource_info_value import PipelineFileResourceInfoValue
+from .pipeline_file_status import PipelineFileStatus
 from .pipeline_transform_config import (
     PipelineTransformConfig,
     PipelineTransformConfig_Advanced,
@@ -276,9 +265,9 @@ from .progress_event_status import ProgressEventStatus
 from .project import Project
 from .project_create import ProjectCreate
 from .prompt_conf import PromptConf
-from .prompt_mixin_prompts import PromptMixinPrompts
-from .prompt_spec import PromptSpec
 from .pydantic_program_mode import PydanticProgramMode
+from .re_rank_config import ReRankConfig
+from .re_ranker_type import ReRankerType
 from .recurring_credit_grant import RecurringCreditGrant
 from .related_node_info import RelatedNodeInfo
 from .related_node_info_node_type import RelatedNodeInfoNodeType
@@ -324,10 +313,10 @@ from .text_node_with_score import TextNodeWithScore
 from .token_chunking_config import TokenChunkingConfig
 from .token_text_splitter import TokenTextSplitter
 from .transformation_category_names import TransformationCategoryNames
-from .usage import Usage
-from .usage_active_alerts_item import UsageActiveAlertsItem
 from .usage_and_plan import UsageAndPlan
 from .usage_metric_response import UsageMetricResponse
+from .usage_response import UsageResponse
+from .usage_response_active_alerts_item import UsageResponseActiveAlertsItem
 from .user_job_record import UserJobRecord
 from .user_organization import UserOrganization
 from .user_organization_create import UserOrganizationCreate
@@ -435,15 +424,7 @@ __all__ = [
     "EmbeddingModelConfigUpdateEmbeddingConfig_HuggingfaceApiEmbedding",
     "EmbeddingModelConfigUpdateEmbeddingConfig_OpenaiEmbedding",
     "EmbeddingModelConfigUpdateEmbeddingConfig_VertexaiEmbedding",
-    "EvalDataset",
-    "EvalDatasetJobParams",
-    "EvalDatasetJobRecord",
     "EvalExecutionParams",
-    "EvalExecutionParamsOverride",
-    "EvalMetric",
-    "EvalQuestion",
-    "EvalQuestionCreate",
-    "EvalQuestionResult",
     "ExtractAgent",
     "ExtractAgentCreate",
     "ExtractAgentCreateDataSchema",
@@ -508,9 +489,6 @@ __all__ = [
     "Llm",
     "LlmModelData",
     "LlmParameters",
-    "LocalEval",
-    "LocalEvalResults",
-    "LocalEvalSets",
     "ManagedIngestionStatus",
     "ManagedIngestionStatusResponse",
     "MarkdownElementNodeParser",
@@ -521,7 +499,6 @@ __all__ = [
     "MetadataFilterValue",
     "MetadataFilters",
     "MetadataFiltersFiltersItem",
-    "MetricResult",
     "NodeParser",
     "NodeRelationship",
     "NoneChunkingConfig",
@@ -585,6 +562,7 @@ __all__ = [
     "PipelineFileCustomMetadataValue",
     "PipelineFilePermissionInfoValue",
     "PipelineFileResourceInfoValue",
+    "PipelineFileStatus",
     "PipelineTransformConfig",
     "PipelineTransformConfig_Advanced",
     "PipelineTransformConfig_Auto",
@@ -600,9 +578,9 @@ __all__ = [
     "Project",
     "ProjectCreate",
     "PromptConf",
-    "PromptMixinPrompts",
-    "PromptSpec",
     "PydanticProgramMode",
+    "ReRankConfig",
+    "ReRankerType",
     "RecurringCreditGrant",
     "RelatedNodeInfo",
     "RelatedNodeInfoNodeType",
@@ -646,10 +624,10 @@ __all__ = [
     "TokenChunkingConfig",
     "TokenTextSplitter",
     "TransformationCategoryNames",
-    "Usage",
-    "UsageActiveAlertsItem",
     "UsageAndPlan",
     "UsageMetricResponse",
+    "UsageResponse",
+    "UsageResponseActiveAlertsItem",
     "UserJobRecord",
     "UserOrganization",
     "UserOrganizationCreate",

llama_cloud/types/base_plan.py

@@ -32,6 +32,9 @@ class BasePlan(pydantic.BaseModel):
     starting_on: typing.Optional[dt.datetime]
     ending_before: typing.Optional[dt.datetime]
     current_billing_period: typing.Optional[BillingPeriod]
+    is_payment_failed: typing.Optional[bool] = pydantic.Field(
+        description="Whether the organization has a failed payment that requires support contact"
+    )
 
     def json(self, **kwargs: typing.Any) -> str:
         kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}

llama_cloud/types/base_plan_name.py

@@ -15,6 +15,9 @@ class BasePlanName(str, enum.Enum):
     PRO = "pro"
     ENTERPRISE_CONTRACT = "enterprise_contract"
     ENTERPRISE_POC = "enterprise_poc"
+    FREE_V_1 = "free_v1"
+    STARTER_V_1 = "starter_v1"
+    PRO_V_1 = "pro_v1"
 
     def visit(
         self,
@@ -26,6 +29,9 @@ class BasePlanName(str, enum.Enum):
         pro: typing.Callable[[], T_Result],
         enterprise_contract: typing.Callable[[], T_Result],
         enterprise_poc: typing.Callable[[], T_Result],
+        free_v_1: typing.Callable[[], T_Result],
+        starter_v_1: typing.Callable[[], T_Result],
+        pro_v_1: typing.Callable[[], T_Result],
     ) -> T_Result:
         if self is BasePlanName.FREE:
             return free()
@@ -43,3 +49,9 @@ class BasePlanName(str, enum.Enum):
             return enterprise_contract()
         if self is BasePlanName.ENTERPRISE_POC:
             return enterprise_poc()
+        if self is BasePlanName.FREE_V_1:
+            return free_v_1()
+        if self is BasePlanName.STARTER_V_1:
+            return starter_v_1()
+        if self is BasePlanName.PRO_V_1:
+            return pro_v_1()

llama_cloud/types/cloud_confluence_data_source.py

@@ -26,6 +26,7 @@ class CloudConfluenceDataSource(pydantic.BaseModel):
     page_ids: typing.Optional[str]
     cql: typing.Optional[str]
     label: typing.Optional[str]
+    index_restricted_pages: typing.Optional[bool] = pydantic.Field(description="Whether to index restricted pages.")
     class_name: typing.Optional[str]
 
     def json(self, **kwargs: typing.Any) -> str:

llama_cloud/types/extract_config.py

@@ -23,9 +23,6 @@ class ExtractConfig(pydantic.BaseModel):
 
     extraction_target: typing.Optional[ExtractTarget] = pydantic.Field(description="The extraction target specified.")
     extraction_mode: typing.Optional[ExtractMode] = pydantic.Field(description="The extraction mode specified.")
-    handle_missing: typing.Optional[bool] = pydantic.Field(
-        description="Whether to handle missing fields in the schema."
-    )
     system_prompt: typing.Optional[str]
 
     def json(self, **kwargs: typing.Any) -> str:

llama_cloud/types/extract_mode.py

@@ -8,10 +8,22 @@ T_Result = typing.TypeVar("T_Result")
 
 class ExtractMode(str, enum.Enum):
     FAST = "FAST"
+    BALANCED = "BALANCED"
+    MULTIMODAL = "MULTIMODAL"
     ACCURATE = "ACCURATE"
 
-    def visit(self, fast: typing.Callable[[], T_Result], accurate: typing.Callable[[], T_Result]) -> T_Result:
+    def visit(
+        self,
+        fast: typing.Callable[[], T_Result],
+        balanced: typing.Callable[[], T_Result],
+        multimodal: typing.Callable[[], T_Result],
+        accurate: typing.Callable[[], T_Result],
+    ) -> T_Result:
         if self is ExtractMode.FAST:
             return fast()
+        if self is ExtractMode.BALANCED:
+            return balanced()
+        if self is ExtractMode.MULTIMODAL:
+            return multimodal()
         if self is ExtractMode.ACCURATE:
             return accurate()

llama_cloud/types/extract_run.py

@@ -39,6 +39,7 @@ class ExtractRun(pydantic.BaseModel):
     job_id: typing.Optional[str]
     data: typing.Optional[ExtractRunData] = pydantic.Field(description="The data extracted from the file")
     extraction_metadata: typing.Optional[typing.Dict[str, typing.Optional[ExtractRunExtractionMetadataValue]]]
+    from_ui: bool = pydantic.Field(description="Whether this extraction run was triggered from the UI")
 
     def json(self, **kwargs: typing.Any) -> str:
         kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}

llama_cloud/types/llama_extract_settings.py

@@ -39,6 +39,7 @@ class LlamaExtractSettings(pydantic.BaseModel):
     extraction_agent_config: typing.Optional[typing.Dict[str, StructParseConf]] = pydantic.Field(
         description="The configuration for the extraction agent."
     )
+    use_multimodal_extraction: typing.Optional[bool]
     llama_parse_params: typing.Optional[LlamaParseParameters] = pydantic.Field(
         description="LlamaParse related settings."
     )

llama_cloud/types/llama_parse_parameters.py

@@ -56,6 +56,7 @@ class LlamaParseParameters(pydantic.BaseModel):
     page_prefix: typing.Optional[str]
     page_suffix: typing.Optional[str]
     webhook_url: typing.Optional[str]
+    preset: typing.Optional[str]
     take_screenshot: typing.Optional[bool]
     is_formatting_instruction: typing.Optional[bool]
     premium_mode: typing.Optional[bool]

llama_cloud/types/parsing_mode.py

@@ -15,7 +15,10 @@ class ParsingMode(str, enum.Enum):
     PARSE_PAGE_WITH_LLM = "parse_page_with_llm"
     PARSE_PAGE_WITH_LVM = "parse_page_with_lvm"
     PARSE_PAGE_WITH_AGENT = "parse_page_with_agent"
+    PARSE_PAGE_WITH_LAYOUT_AGENT = "parse_page_with_layout_agent"
     PARSE_DOCUMENT_WITH_LLM = "parse_document_with_llm"
+    PARSE_DOCUMENT_WITH_LVM = "parse_document_with_lvm"
+    PARSE_DOCUMENT_WITH_AGENT = "parse_document_with_agent"
 
     def visit(
         self,
@@ -23,7 +26,10 @@ class ParsingMode(str, enum.Enum):
         parse_page_with_llm: typing.Callable[[], T_Result],
         parse_page_with_lvm: typing.Callable[[], T_Result],
         parse_page_with_agent: typing.Callable[[], T_Result],
+        parse_page_with_layout_agent: typing.Callable[[], T_Result],
         parse_document_with_llm: typing.Callable[[], T_Result],
+        parse_document_with_lvm: typing.Callable[[], T_Result],
+        parse_document_with_agent: typing.Callable[[], T_Result],
     ) -> T_Result:
         if self is ParsingMode.PARSE_PAGE_WITHOUT_LLM:
             return parse_page_without_llm()
@@ -33,5 +39,11 @@ class ParsingMode(str, enum.Enum):
             return parse_page_with_lvm()
         if self is ParsingMode.PARSE_PAGE_WITH_AGENT:
             return parse_page_with_agent()
+        if self is ParsingMode.PARSE_PAGE_WITH_LAYOUT_AGENT:
+            return parse_page_with_layout_agent()
         if self is ParsingMode.PARSE_DOCUMENT_WITH_LLM:
             return parse_document_with_llm()
+        if self is ParsingMode.PARSE_DOCUMENT_WITH_LVM:
+            return parse_document_with_lvm()
+        if self is ParsingMode.PARSE_DOCUMENT_WITH_AGENT:
+            return parse_document_with_agent()

llama_cloud/types/pipeline_file.py

@@ -8,6 +8,7 @@ from .pipeline_file_config_hash_value import PipelineFileConfigHashValue
 from .pipeline_file_custom_metadata_value import PipelineFileCustomMetadataValue
 from .pipeline_file_permission_info_value import PipelineFilePermissionInfoValue
 from .pipeline_file_resource_info_value import PipelineFileResourceInfoValue
+from .pipeline_file_status import PipelineFileStatus
 
 try:
     import pydantic
@@ -40,7 +41,7 @@ class PipelineFile(pydantic.BaseModel):
     custom_metadata: typing.Optional[typing.Dict[str, typing.Optional[PipelineFileCustomMetadataValue]]]
     config_hash: typing.Optional[typing.Dict[str, typing.Optional[PipelineFileConfigHashValue]]]
     indexed_page_count: typing.Optional[int]
-    status: typing.Optional[str]
+    status: typing.Optional[PipelineFileStatus]
 
     def json(self, **kwargs: typing.Any) -> str:
         kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}

llama_cloud/types/pipeline_file_status.py

@@ -0,0 +1,33 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import enum
+import typing
+
+T_Result = typing.TypeVar("T_Result")
+
+
+class PipelineFileStatus(str, enum.Enum):
+    NOT_STARTED = "NOT_STARTED"
+    IN_PROGRESS = "IN_PROGRESS"
+    SUCCESS = "SUCCESS"
+    ERROR = "ERROR"
+    CANCELLED = "CANCELLED"
+
+    def visit(
+        self,
+        not_started: typing.Callable[[], T_Result],
+        in_progress: typing.Callable[[], T_Result],
+        success: typing.Callable[[], T_Result],
+        error: typing.Callable[[], T_Result],
+        cancelled: typing.Callable[[], T_Result],
+    ) -> T_Result:
+        if self is PipelineFileStatus.NOT_STARTED:
+            return not_started()
+        if self is PipelineFileStatus.IN_PROGRESS:
+            return in_progress()
+        if self is PipelineFileStatus.SUCCESS:
+            return success()
+        if self is PipelineFileStatus.ERROR:
+            return error()
+        if self is PipelineFileStatus.CANCELLED:
+            return cancelled()

llama_cloud/types/plan_limits.py

@@ -18,6 +18,7 @@ class PlanLimits(pydantic.BaseModel):
     allow_pay_as_you_go: bool = pydantic.Field(description="Whether usage is allowed after credit grants are exhausted")
     subscription_cost_usd: int
     max_monthly_invoice_total_usd: typing.Optional[int]
+    spending_soft_alerts_usd_cents: typing.Optional[typing.List[int]]
     max_concurrent_parse_jobs_premium: typing.Optional[int]
     max_concurrent_parse_jobs_other: typing.Optional[int]
     max_extraction_agents: typing.Optional[int]

llama_cloud/types/preset_composite_retrieval_params.py

@@ -5,6 +5,7 @@ import typing
 
 from ..core.datetime_utils import serialize_datetime
 from .composite_retrieval_mode import CompositeRetrievalMode
+from .re_rank_config import ReRankConfig
 
 try:
     import pydantic
@@ -17,8 +18,9 @@ except ImportError:
 
 class PresetCompositeRetrievalParams(pydantic.BaseModel):
     mode: typing.Optional[CompositeRetrievalMode] = pydantic.Field(description="The mode of composite retrieval.")
-    rerank_top_n: typing.Optional[int] = pydantic.Field(
-        description="The number of nodes to retrieve after reranking over retrieved nodes from all retrieval tools."
+    rerank_top_n: typing.Optional[int]
+    rerank_config: typing.Optional[ReRankConfig] = pydantic.Field(
+        description="The rerank configuration for composite retrieval."
     )
 
     def json(self, **kwargs: typing.Any) -> str:

llama_cloud/types/prompt_conf.py

@@ -18,6 +18,7 @@ class PromptConf(pydantic.BaseModel):
     system_prompt: typing.Optional[str] = pydantic.Field(description="The system prompt to use for the extraction.")
     extraction_prompt: typing.Optional[str] = pydantic.Field(description="The prompt to use for the extraction.")
     error_handling_prompt: typing.Optional[str] = pydantic.Field(description="The prompt to use for error handling.")
+    reasoning_prompt: typing.Optional[str] = pydantic.Field(description="The prompt to use for reasoning.")
 
     def json(self, **kwargs: typing.Any) -> str:
         kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}

llama_cloud/types/{eval_question_create.py → re_rank_config.py}

@@ -4,6 +4,7 @@ import datetime as dt
 import typing
 
 from ..core.datetime_utils import serialize_datetime
+from .re_ranker_type import ReRankerType
 
 try:
     import pydantic
@@ -14,8 +15,11 @@ except ImportError:
     import pydantic  # type: ignore
 
 
-class EvalQuestionCreate(pydantic.BaseModel):
-    content: str = pydantic.Field(description="The content of the question.")
+class ReRankConfig(pydantic.BaseModel):
+    top_n: typing.Optional[int] = pydantic.Field(
+        description="The number of nodes to retrieve after reranking over retrieved nodes from all retrieval tools."
+    )
+    type: typing.Optional[ReRankerType] = pydantic.Field(description="The type of reranker to use.")
 
     def json(self, **kwargs: typing.Any) -> str:
         kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}

llama_cloud/types/re_ranker_type.py

@@ -0,0 +1,41 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import enum
+import typing
+
+T_Result = typing.TypeVar("T_Result")
+
+
+class ReRankerType(str, enum.Enum):
+    """
+    Enum for the reranker type.
+    """
+
+    SYSTEM_DEFAULT = "system_default"
+    LLM = "llm"
+    COHERE = "cohere"
+    BEDROCK = "bedrock"
+    SCORE = "score"
+    DISABLED = "disabled"
+
+    def visit(
+        self,
+        system_default: typing.Callable[[], T_Result],
+        llm: typing.Callable[[], T_Result],
+        cohere: typing.Callable[[], T_Result],
+        bedrock: typing.Callable[[], T_Result],
+        score: typing.Callable[[], T_Result],
+        disabled: typing.Callable[[], T_Result],
+    ) -> T_Result:
+        if self is ReRankerType.SYSTEM_DEFAULT:
+            return system_default()
+        if self is ReRankerType.LLM:
+            return llm()
+        if self is ReRankerType.COHERE:
+            return cohere()
+        if self is ReRankerType.BEDROCK:
+            return bedrock()
+        if self is ReRankerType.SCORE:
+            return score()
+        if self is ReRankerType.DISABLED:
+            return disabled()