llama-cloud 0.0.15__py3-none-any.whl → 0.0.16__py3-none-any.whl

llama_cloud/types/configurable_transformation_names.py

@@ -24,6 +24,7 @@ class ConfigurableTransformationNames(str, enum.Enum):
     BEDROCK_EMBEDDING = "BEDROCK_EMBEDDING"
     HUGGINGFACE_API_EMBEDDING = "HUGGINGFACE_API_EMBEDDING"
     GEMINI_EMBEDDING = "GEMINI_EMBEDDING"
+    VERTEXAI_EMBEDDING = "VERTEXAI_EMBEDDING"
 
     def visit(
         self,
@@ -40,6 +41,7 @@ class ConfigurableTransformationNames(str, enum.Enum):
         bedrock_embedding: typing.Callable[[], T_Result],
         huggingface_api_embedding: typing.Callable[[], T_Result],
         gemini_embedding: typing.Callable[[], T_Result],
+        vertexai_embedding: typing.Callable[[], T_Result],
     ) -> T_Result:
         if self is ConfigurableTransformationNames.CHARACTER_SPLITTER:
             return character_splitter()
@@ -67,3 +69,5 @@ class ConfigurableTransformationNames(str, enum.Enum):
             return huggingface_api_embedding()
         if self is ConfigurableTransformationNames.GEMINI_EMBEDDING:
             return gemini_embedding()
+        if self is ConfigurableTransformationNames.VERTEXAI_EMBEDDING:
+            return vertexai_embedding()

llama_cloud/types/configured_transformation_item_component_one.py

@@ -7,6 +7,7 @@ from .bedrock_embedding import BedrockEmbedding
 from .character_splitter import CharacterSplitter
 from .code_splitter import CodeSplitter
 from .cohere_embedding import CohereEmbedding
+from .extend_vertex_text_embedding import ExtendVertexTextEmbedding
 from .gemini_embedding import GeminiEmbedding
 from .hugging_face_inference_api_embedding import HuggingFaceInferenceApiEmbedding
 from .markdown_element_node_parser import MarkdownElementNodeParser
@@ -30,4 +31,5 @@ ConfiguredTransformationItemComponentOne = typing.Union[
     BedrockEmbedding,
     HuggingFaceInferenceApiEmbedding,
     GeminiEmbedding,
+    ExtendVertexTextEmbedding,
 ]

llama_cloud/types/extend_vertex_text_embedding.py

@@ -0,0 +1,58 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import datetime as dt
+import typing
+
+from ..core.datetime_utils import serialize_datetime
+from .vertex_embedding_mode import VertexEmbeddingMode
+
+try:
+    import pydantic
+    if pydantic.__version__.startswith("1."):
+        raise ImportError
+    import pydantic.v1 as pydantic  # type: ignore
+except ImportError:
+    import pydantic  # type: ignore
+
+
+class ExtendVertexTextEmbedding(pydantic.BaseModel):
+    """
+    Base class for embeddings.
+    """
+
+    model_name: typing.Optional[str] = pydantic.Field(description="The name of the embedding model.")
+    embed_batch_size: typing.Optional[int] = pydantic.Field(description="The batch size for embedding calls.")
+    callback_manager: typing.Optional[typing.Dict[str, typing.Any]]
+    num_workers: typing.Optional[int] = pydantic.Field(
+        description="The number of workers to use for async embedding calls."
+    )
+    embed_mode: VertexEmbeddingMode = pydantic.Field(description="The embedding mode to use.")
+    additional_kwargs: typing.Optional[typing.Dict[str, typing.Any]] = pydantic.Field(
+        description="Additional kwargs for the Vertex."
+    )
+    client_email: typing.Optional[str] = pydantic.Field(
+        description="The client email to use when making Vertex API calls."
+    )
+    token_uri: typing.Optional[str] = pydantic.Field(description="The token uri to use when making Vertex API calls.")
+    private_key_id: typing.Optional[str] = pydantic.Field(
+        description="The private key id to use when making Vertex API calls."
+    )
+    private_key: typing.Optional[str] = pydantic.Field(
+        description="The private key to use when making Vertex API calls."
+    )
+    project: str = pydantic.Field(description="The default GCP project to use when making Vertex API calls.")
+    location: str = pydantic.Field(description="The default location to use when making API calls.")
+    class_name: typing.Optional[str]
+
+    def json(self, **kwargs: typing.Any) -> str:
+        kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}
+        return super().json(**kwargs_with_defaults)
+
+    def dict(self, **kwargs: typing.Any) -> typing.Dict[str, typing.Any]:
+        kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}
+        return super().dict(**kwargs_with_defaults)
+
+    class Config:
+        frozen = True
+        smart_union = True
+        json_encoders = {dt.datetime: serialize_datetime}

llama_cloud/types/llama_parse_parameters.py

@@ -23,6 +23,7 @@ class LlamaParseParameters(pydantic.BaseModel):
     languages: typing.Optional[typing.List[ParserLanguages]]
     parsing_instruction: typing.Optional[str]
     disable_ocr: typing.Optional[bool]
+    disable_reconstruction: typing.Optional[bool]
     invalidate_cache: typing.Optional[bool]
     do_not_cache: typing.Optional[bool]
     fast_mode: typing.Optional[bool]
@@ -33,11 +34,12 @@ class LlamaParseParameters(pydantic.BaseModel):
     page_separator: typing.Optional[str]
     bounding_box: typing.Optional[str]
     target_pages: typing.Optional[str]
-    use_vendor_multimodal_model: typing.Optional[str]
+    use_vendor_multimodal_model: typing.Optional[bool]
     vendor_multimodal_model_name: typing.Optional[str]
     vendor_multimodal_api_key: typing.Optional[str]
     page_prefix: typing.Optional[str]
     page_suffix: typing.Optional[str]
+    webhook_url: typing.Optional[str]
     take_screenshot: typing.Optional[bool]
     s_3_input_path: typing.Optional[str] = pydantic.Field(alias="s3_input_path")
     s_3_output_path_prefix: typing.Optional[str] = pydantic.Field(alias="s3_output_path_prefix")

llama_cloud/types/llm_model_data.py

@@ -21,6 +21,7 @@ class LlmModelData(pydantic.BaseModel):
 
     name: str = pydantic.Field(description="The name of the LLM model.")
     description: str = pydantic.Field(description="The description of the LLM model.")
+    multi_modal: bool = pydantic.Field(description="Whether the model supports multi-modal image input")
 
     def json(self, **kwargs: typing.Any) -> str:
         kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}

llama_cloud/types/llm_parameters.py

@@ -4,6 +4,7 @@ import datetime as dt
 import typing
 
 from ..core.datetime_utils import serialize_datetime
+from .supported_llm_model_names import SupportedLlmModelNames
 
 try:
     import pydantic
@@ -20,7 +21,9 @@ class LlmParameters(pydantic.BaseModel):
     Comes with special serialization logic for types used commonly in platform codebase.
     """
 
-    model_name: typing.Optional[str] = pydantic.Field(description="The name of the model to use for LLM completions.")
+    model_name: typing.Optional[SupportedLlmModelNames] = pydantic.Field(
+        description="The name of the model to use for LLM completions."
+    )
     system_prompt: typing.Optional[str] = pydantic.Field(description="The system prompt to use for the completion.")
     temperature: typing.Optional[float] = pydantic.Field(description="The temperature value for the model.")
     class_name: typing.Optional[str]

llama_cloud/types/page_screenshot_metadata.py

@@ -0,0 +1,33 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import datetime as dt
+import typing
+
+from ..core.datetime_utils import serialize_datetime
+
+try:
+    import pydantic
+    if pydantic.__version__.startswith("1."):
+        raise ImportError
+    import pydantic.v1 as pydantic  # type: ignore
+except ImportError:
+    import pydantic  # type: ignore
+
+
+class PageScreenshotMetadata(pydantic.BaseModel):
+    page_index: int = pydantic.Field(description="The index of the page for which the screenshot is taken (0-indexed)")
+    file_id: str = pydantic.Field(description="The ID of the file that the page screenshot was taken from")
+    image_size: int = pydantic.Field(description="The size of the image in bytes")
+
+    def json(self, **kwargs: typing.Any) -> str:
+        kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}
+        return super().json(**kwargs_with_defaults)
+
+    def dict(self, **kwargs: typing.Any) -> typing.Dict[str, typing.Any]:
+        kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}
+        return super().dict(**kwargs_with_defaults)
+
+    class Config:
+        frozen = True
+        smart_union = True
+        json_encoders = {dt.datetime: serialize_datetime}

llama_cloud/types/page_screenshot_node_with_score.py

@@ -0,0 +1,38 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import datetime as dt
+import typing
+
+from ..core.datetime_utils import serialize_datetime
+from .page_screenshot_metadata import PageScreenshotMetadata
+
+try:
+    import pydantic
+    if pydantic.__version__.startswith("1."):
+        raise ImportError
+    import pydantic.v1 as pydantic  # type: ignore
+except ImportError:
+    import pydantic  # type: ignore
+
+
+class PageScreenshotNodeWithScore(pydantic.BaseModel):
+    """
+    Page screenshot metadata with score
+    """
+
+    node: PageScreenshotMetadata
+    score: float = pydantic.Field(description="The score of the screenshot node")
+    class_name: typing.Optional[str]
+
+    def json(self, **kwargs: typing.Any) -> str:
+        kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}
+        return super().json(**kwargs_with_defaults)
+
+    def dict(self, **kwargs: typing.Any) -> typing.Dict[str, typing.Any]:
+        kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}
+        return super().dict(**kwargs_with_defaults)
+
+    class Config:
+        frozen = True
+        smart_union = True
+        json_encoders = {dt.datetime: serialize_datetime}

llama_cloud/types/pipeline.py

@@ -8,6 +8,7 @@ from .configured_transformation_item import ConfiguredTransformationItem
 from .data_sink import DataSink
 from .eval_execution_params import EvalExecutionParams
 from .llama_parse_parameters import LlamaParseParameters
+from .pipeline_configuration_hashes import PipelineConfigurationHashes
 from .pipeline_embedding_config import PipelineEmbeddingConfig
 from .pipeline_transform_config import PipelineTransformConfig
 from .pipeline_type import PipelineType
@@ -44,6 +45,9 @@ class Pipeline(pydantic.BaseModel):
     configured_transformations: typing.Optional[typing.List[ConfiguredTransformationItem]] = pydantic.Field(
         description="Deprecated don't use it, List of configured transformations."
     )
+    config_hash: typing.Optional[PipelineConfigurationHashes] = pydantic.Field(
+        description="Hashes for the configuration of the pipeline."
+    )
     transform_config: typing.Optional[PipelineTransformConfig] = pydantic.Field(
         description="Configuration for the transformation."
     )

llama_cloud/types/pipeline_configuration_hashes.py

@@ -0,0 +1,37 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import datetime as dt
+import typing
+
+from ..core.datetime_utils import serialize_datetime
+
+try:
+    import pydantic
+    if pydantic.__version__.startswith("1."):
+        raise ImportError
+    import pydantic.v1 as pydantic  # type: ignore
+except ImportError:
+    import pydantic  # type: ignore
+
+
+class PipelineConfigurationHashes(pydantic.BaseModel):
+    """
+    Hashes for the configuration of a pipeline.
+    """
+
+    embedding_config_hash: typing.Optional[str] = pydantic.Field(description="Hash of the embedding config.")
+    parsing_config_hash: typing.Optional[str] = pydantic.Field(description="Hash of the llama parse parameters.")
+    transform_config_hash: typing.Optional[str] = pydantic.Field(description="Hash of the transform config.")
+
+    def json(self, **kwargs: typing.Any) -> str:
+        kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}
+        return super().json(**kwargs_with_defaults)
+
+    def dict(self, **kwargs: typing.Any) -> typing.Dict[str, typing.Any]:
+        kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}
+        return super().dict(**kwargs_with_defaults)
+
+    class Config:
+        frozen = True
+        smart_union = True
+        json_encoders = {dt.datetime: serialize_datetime}

llama_cloud/types/pipeline_create_embedding_config.py

@@ -12,6 +12,7 @@ from .cohere_embedding_config import CohereEmbeddingConfig
 from .gemini_embedding_config import GeminiEmbeddingConfig
 from .hugging_face_inference_api_embedding_config import HuggingFaceInferenceApiEmbeddingConfig
 from .open_ai_embedding_config import OpenAiEmbeddingConfig
+from .vertex_ai_embedding_config import VertexAiEmbeddingConfig
 
 
 class PipelineCreateEmbeddingConfig_OpenaiEmbedding(OpenAiEmbeddingConfig):
@@ -68,6 +69,15 @@ class PipelineCreateEmbeddingConfig_CohereEmbedding(CohereEmbeddingConfig):
         allow_population_by_field_name = True
 
 
+class PipelineCreateEmbeddingConfig_VertexaiEmbedding(VertexAiEmbeddingConfig):
+    type: typing_extensions.Literal["VERTEXAI_EMBEDDING"]
+
+    class Config:
+        frozen = True
+        smart_union = True
+        allow_population_by_field_name = True
+
+
 PipelineCreateEmbeddingConfig = typing.Union[
     PipelineCreateEmbeddingConfig_OpenaiEmbedding,
     PipelineCreateEmbeddingConfig_AzureEmbedding,
@@ -75,4 +85,5 @@ PipelineCreateEmbeddingConfig = typing.Union[
     PipelineCreateEmbeddingConfig_BedrockEmbedding,
     PipelineCreateEmbeddingConfig_GeminiEmbedding,
     PipelineCreateEmbeddingConfig_CohereEmbedding,
+    PipelineCreateEmbeddingConfig_VertexaiEmbedding,
 ]

llama_cloud/types/pipeline_embedding_config.py

@@ -12,6 +12,7 @@ from .cohere_embedding_config import CohereEmbeddingConfig
 from .gemini_embedding_config import GeminiEmbeddingConfig
 from .hugging_face_inference_api_embedding_config import HuggingFaceInferenceApiEmbeddingConfig
 from .open_ai_embedding_config import OpenAiEmbeddingConfig
+from .vertex_ai_embedding_config import VertexAiEmbeddingConfig
 
 
 class PipelineEmbeddingConfig_OpenaiEmbedding(OpenAiEmbeddingConfig):
@@ -68,6 +69,15 @@ class PipelineEmbeddingConfig_CohereEmbedding(CohereEmbeddingConfig):
         allow_population_by_field_name = True
 
 
+class PipelineEmbeddingConfig_VertexaiEmbedding(VertexAiEmbeddingConfig):
+    type: typing_extensions.Literal["VERTEXAI_EMBEDDING"]
+
+    class Config:
+        frozen = True
+        smart_union = True
+        allow_population_by_field_name = True
+
+
 PipelineEmbeddingConfig = typing.Union[
     PipelineEmbeddingConfig_OpenaiEmbedding,
     PipelineEmbeddingConfig_AzureEmbedding,
@@ -75,4 +85,5 @@ PipelineEmbeddingConfig = typing.Union[
     PipelineEmbeddingConfig_BedrockEmbedding,
     PipelineEmbeddingConfig_GeminiEmbedding,
     PipelineEmbeddingConfig_CohereEmbedding,
+    PipelineEmbeddingConfig_VertexaiEmbedding,
 ]

llama_cloud/types/pipeline_file.py

@@ -4,6 +4,7 @@ import datetime as dt
 import typing
 
 from ..core.datetime_utils import serialize_datetime
+from .pipeline_file_config_hash_value import PipelineFileConfigHashValue
 from .pipeline_file_custom_metadata_value import PipelineFileCustomMetadataValue
 from .pipeline_file_resource_info_value import PipelineFileResourceInfoValue
 
@@ -40,6 +41,9 @@ class PipelineFile(pydantic.BaseModel):
     custom_metadata: typing.Optional[typing.Dict[str, PipelineFileCustomMetadataValue]] = pydantic.Field(
         description="Custom metadata for the file"
     )
+    config_hash: typing.Optional[typing.Dict[str, PipelineFileConfigHashValue]] = pydantic.Field(
+        description="Hashes for the configuration of the pipeline."
+    )
 
     def json(self, **kwargs: typing.Any) -> str:
         kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}

llama_cloud/types/pipeline_file_config_hash_value.py

@@ -0,0 +1,5 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing
+
+PipelineFileConfigHashValue = typing.Union[typing.Dict[str, typing.Any], typing.List[typing.Any], str, int, float, bool]

llama_cloud/types/preset_retrieval_params.py

@@ -33,6 +33,7 @@ class PresetRetrievalParams(pydantic.BaseModel):
         description="Number of files to retrieve (only for retrieval mode files_via_metadata and files_via_content)."
     )
     retrieval_mode: typing.Optional[RetrievalMode] = pydantic.Field(description="The retrieval mode for the query.")
+    retrieve_image_nodes: typing.Optional[bool] = pydantic.Field(description="Whether to retrieve image nodes.")
 
     def json(self, **kwargs: typing.Any) -> str:
         kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}

llama_cloud/types/retrieve_results.py

@@ -4,6 +4,7 @@ import datetime as dt
 import typing
 
 from ..core.datetime_utils import serialize_datetime
+from .page_screenshot_node_with_score import PageScreenshotNodeWithScore
 from .text_node_with_score import TextNodeWithScore
 
 try:
@@ -24,6 +25,9 @@ class RetrieveResults(pydantic.BaseModel):
     retrieval_nodes: typing.List[TextNodeWithScore] = pydantic.Field(
         description="The nodes retrieved by the pipeline for the given query."
     )
+    image_nodes: typing.Optional[typing.List[PageScreenshotNodeWithScore]] = pydantic.Field(
+        description="The image nodes retrieved by the pipeline for the given query."
+    )
     retrieval_latency: typing.Dict[str, float] = pydantic.Field(
         description="The end-to-end latency for retrieval and reranking."
     )

llama_cloud/types/vertex_ai_embedding_config.py

@@ -0,0 +1,34 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import datetime as dt
+import typing
+
+from ..core.datetime_utils import serialize_datetime
+from .extend_vertex_text_embedding import ExtendVertexTextEmbedding
+
+try:
+    import pydantic
+    if pydantic.__version__.startswith("1."):
+        raise ImportError
+    import pydantic.v1 as pydantic  # type: ignore
+except ImportError:
+    import pydantic  # type: ignore
+
+
+class VertexAiEmbeddingConfig(pydantic.BaseModel):
+    component: typing.Optional[ExtendVertexTextEmbedding] = pydantic.Field(
+        description="Configuration for the VertexAI embedding model."
+    )
+
+    def json(self, **kwargs: typing.Any) -> str:
+        kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}
+        return super().json(**kwargs_with_defaults)
+
+    def dict(self, **kwargs: typing.Any) -> typing.Dict[str, typing.Any]:
+        kwargs_with_defaults: typing.Any = {"by_alias": True, "exclude_unset": True, **kwargs}
+        return super().dict(**kwargs_with_defaults)
+
+    class Config:
+        frozen = True
+        smart_union = True
+        json_encoders = {dt.datetime: serialize_datetime}

llama_cloud/types/vertex_embedding_mode.py

@@ -0,0 +1,45 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import enum
+import typing
+
+T_Result = typing.TypeVar("T_Result")
+
+
+class VertexEmbeddingMode(str, enum.Enum):
+    """
+    VertexAI embedding mode.
+
+    Attributes:
+    DEFAULT_MODE (str): The default embedding mode, for older models before August 2023,
+    that does not support task_type
+    CLASSIFICATION_MODE (str): Optimizes embeddings for classification tasks.
+    CLUSTERING_MODE (str): Optimizes embeddings for clustering tasks.
+    SEMANTIC_SIMILARITY_MODE (str): Optimizes embeddings for tasks that require assessments of semantic similarity.
+    RETRIEVAL_MODE (str): Optimizes embeddings for retrieval tasks, including search and document retrieval.
+    """
+
+    DEFAULT = "default"
+    CLASSIFICATION = "classification"
+    CLUSTERING = "clustering"
+    SIMILARITY = "similarity"
+    RETRIEVAL = "retrieval"
+
+    def visit(
+        self,
+        default: typing.Callable[[], T_Result],
+        classification: typing.Callable[[], T_Result],
+        clustering: typing.Callable[[], T_Result],
+        similarity: typing.Callable[[], T_Result],
+        retrieval: typing.Callable[[], T_Result],
+    ) -> T_Result:
+        if self is VertexEmbeddingMode.DEFAULT:
+            return default()
+        if self is VertexEmbeddingMode.CLASSIFICATION:
+            return classification()
+        if self is VertexEmbeddingMode.CLUSTERING:
+            return clustering()
+        if self is VertexEmbeddingMode.SIMILARITY:
+            return similarity()
+        if self is VertexEmbeddingMode.RETRIEVAL:
+            return retrieval()

{llama_cloud-0.0.15.dist-info → llama_cloud-0.0.16.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: llama-cloud
-Version: 0.0.15
+Version: 0.0.16
 Summary: 
 Author: Logan Markewich
 Author-email: logan@runllama.ai