retab 0.0.39__py3-none-any.whl → 0.0.41__py3-none-any.whl

retab/_utils/_model_cards/gemini.yaml

@@ -108,7 +108,7 @@
       prompt: 0.50
       completion: 0.40
   capabilities:
-    modalities: ["text", "image", "audio", "video"]
+    modalities: ["text", "image", "audio"]
     endpoints: ["chat_completions"]
     features: ["streaming", "structured_outputs"]
   temperature_support: true

retab/resources/consensus/completions.py

@@ -5,7 +5,7 @@ from openai.types.shared_params.response_format_json_schema import ResponseForma
 from pydantic import BaseModel as ResponseFormatT
 
 from ..._resource import AsyncAPIResource, SyncAPIResource
-from ..._utils.ai_models import assert_valid_model_extraction
+from ...utils.ai_models import assert_valid_model_extraction
 from ...types.chat import ChatCompletionRetabMessage
 from ...types.completions import RetabChatCompletionsRequest
 from ...types.documents.extractions import RetabParsedChatCompletion

retab/resources/consensus/completions_stream.py

@@ -9,9 +9,9 @@ from openai.types.shared_params.response_format_json_schema import ResponseForma
 from pydantic import BaseModel as ResponseFormatT
 
 from ..._resource import AsyncAPIResource, SyncAPIResource
-from ..._utils.ai_models import assert_valid_model_extraction
-from ..._utils.json_schema import unflatten_dict
-from ..._utils.stream_context_managers import as_async_context_manager, as_context_manager
+from ...utils.ai_models import assert_valid_model_extraction
+from ...utils.json_schema import unflatten_dict
+from ...utils.stream_context_managers import as_async_context_manager, as_context_manager
 from ...types.chat import ChatCompletionRetabMessage
 from ...types.completions import RetabChatCompletionsRequest
 from ...types.documents.extractions import RetabParsedChatCompletion, RetabParsedChatCompletionChunk, RetabParsedChoice

retab/resources/consensus/responses.py

@@ -7,7 +7,7 @@ from openai.types.shared_params.reasoning import Reasoning
 from pydantic import BaseModel
 
 from ..._resource import AsyncAPIResource, SyncAPIResource
-from ..._utils.ai_models import assert_valid_model_extraction
+from ...utils.ai_models import assert_valid_model_extraction
 from ...types.completions import RetabChatResponseCreateRequest
 from ...types.documents.extractions import UiResponse
 from ...types.schemas.object import Schema

retab/resources/consensus/responses_stream.py

@@ -6,8 +6,8 @@ from openai.types.shared_params.reasoning import Reasoning
 from pydantic import BaseModel
 
 from ..._resource import AsyncAPIResource, SyncAPIResource
-from ..._utils.ai_models import assert_valid_model_extraction
-from ..._utils.stream_context_managers import as_async_context_manager, as_context_manager
+from ...utils.ai_models import assert_valid_model_extraction
+from ...utils.stream_context_managers import as_async_context_manager, as_context_manager
 from ...types.completions import RetabChatResponseCreateRequest
 from ...types.documents.extractions import UiResponse
 from ...types.schemas.object import Schema

retab/resources/documents/client.py

@@ -8,9 +8,9 @@ from pydantic_core import PydanticUndefined
 from openai.types.chat.chat_completion_reasoning_effort import ChatCompletionReasoningEffort
 
 from ..._resource import AsyncAPIResource, SyncAPIResource
-from ..._utils.json_schema import load_json_schema, filter_auxiliary_fields_json
-from ..._utils.mime import convert_mime_data_to_pil_image, prepare_mime_document
-from ..._utils.ai_models import assert_valid_model_extraction
+from ...utils.json_schema import load_json_schema, filter_auxiliary_fields_json
+from ...utils.mime import convert_mime_data_to_pil_image, prepare_mime_document
+from ...utils.ai_models import assert_valid_model_extraction
 from ...types.documents.create_messages import DocumentCreateInputRequest, DocumentCreateMessageRequest, DocumentMessage
 from ...types.documents.extractions import DocumentExtractRequest, RetabParsedChatCompletion
 from ...types.documents.parse import ParseRequest, ParseResult, TableParsingFormat

retab/resources/documents/extractions.py

@@ -14,10 +14,10 @@ from pydantic_core import PydanticUndefined
 from pydantic import HttpUrl
 
 from ..._resource import AsyncAPIResource, SyncAPIResource
-from ..._utils.ai_models import assert_valid_model_extraction
-from ..._utils.json_schema import filter_auxiliary_fields_json, load_json_schema, unflatten_dict
-from ..._utils.mime import MIMEData, prepare_mime_document
-from ..._utils.stream_context_managers import as_async_context_manager, as_context_manager
+from ...utils.ai_models import assert_valid_model_extraction
+from ...utils.json_schema import filter_auxiliary_fields_json, load_json_schema, unflatten_dict
+from ...utils.mime import MIMEData, prepare_mime_document
+from ...utils.stream_context_managers import as_async_context_manager, as_context_manager
 from ...types.chat import ChatCompletionRetabMessage
 from ...types.documents.extractions import DocumentExtractRequest, LogExtractionRequest, RetabParsedChatCompletion, RetabParsedChatCompletionChunk, RetabParsedChoice
 from ...types.browser_canvas import BrowserCanvas

retab/resources/evals.py

@@ -7,7 +7,7 @@ from openai.types.chat.chat_completion_reasoning_effort import ChatCompletionRea
 from pydantic import HttpUrl
 
 from .._resource import AsyncAPIResource, SyncAPIResource
-from .._utils.mime import prepare_mime_document
+from ..utils.mime import prepare_mime_document
 from ..types.evals import (
     CreateIterationRequest,
     DistancesResult,

retab/resources/evaluations/documents.py

@@ -6,7 +6,7 @@ import PIL.Image
 from pydantic import HttpUrl
 
 from ..._resource import AsyncAPIResource, SyncAPIResource
-from ..._utils.mime import prepare_mime_document
+from ...utils.mime import prepare_mime_document
 from ...types.evaluations import DocumentItem, EvaluationDocument, PatchEvaluationDocumentRequest
 from ...types.mime import MIMEData
 from ...types.standards import PreparedRequest, DeleteResponse, FieldUnset

retab/resources/jsonlUtils.py

@@ -18,10 +18,10 @@ from pydantic_core import PydanticUndefined
 from tqdm import tqdm
 
 from .._resource import AsyncAPIResource, SyncAPIResource
-from .._utils.ai_models import assert_valid_model_extraction, get_provider_for_model
-from .._utils.chat import convert_to_anthropic_format, convert_to_openai_format, separate_messages
-from .._utils.display import Metrics, display_metrics, process_dataset_and_compute_metrics
-from .._utils.json_schema import load_json_schema
+from ..utils.ai_models import assert_valid_model_extraction, get_provider_for_model
+from ..utils.chat import convert_to_anthropic_format, convert_to_openai_format, separate_messages
+from ..utils.display import Metrics, display_metrics, process_dataset_and_compute_metrics
+from ..utils.json_schema import load_json_schema
 from ..types.chat import ChatCompletionRetabMessage
 from ..types.modalities import Modality
 from ..types.schemas.object import Schema

retab/resources/processors/automations/endpoints.py

@@ -3,7 +3,7 @@ from typing import Literal, Optional
 from pydantic_core import PydanticUndefined
 
 from ...._resource import AsyncAPIResource, SyncAPIResource
-from ...._utils.ai_models import assert_valid_model_extraction
+from ....utils.ai_models import assert_valid_model_extraction
 from ....types.automations.endpoints import Endpoint, ListEndpoints, UpdateEndpointRequest
 from ....types.standards import PreparedRequest
 
@@ -120,7 +120,7 @@ class Endpoints(SyncAPIResource, EndpointsMixin):
             need_validation=need_validation,
         )
         response = self._client._prepared_request(request)
-        print(f"Endpoint ID: {response['id']}. Endpoint available at https://www.retab.dev/dashboard/processors/{response['id']}")
+        print(f"Endpoint Created. Url: https://www.retab.dev/dashboard/processors/automations/{response['id']}")
         return Endpoint.model_validate(response)
 
     def list(
@@ -229,7 +229,7 @@ class AsyncEndpoints(AsyncAPIResource, EndpointsMixin):
             need_validation=need_validation,
         )
         response = await self._client._prepared_request(request)
-        print(f"Endpoint ID: {response['id']}. Endpoint available at https://www.retab.dev/dashboard/processors/{response['id']}")
+        print(f"Endpoint Created. Url: https://www.retab.dev/dashboard/processors/automations/{response['id']}")
 
         return Endpoint.model_validate(response)
 

retab/resources/processors/automations/links.py

@@ -128,7 +128,7 @@ class Links(SyncAPIResource, LinksMixin):
         )
         response = self._client._prepared_request(request)
 
-        print(f"Extraction Link Created. Link available at https://www.retab.dev/dashboard/processors/{response['id']}")
+        print(f"Link Created. Url: https://www.retab.dev/dashboard/processors/automations/{response['id']}")
         return Link.model_validate(response)
 
     def list(
@@ -248,7 +248,7 @@ class AsyncLinks(AsyncAPIResource, LinksMixin):
             password=password,
         )
         response = await self._client._prepared_request(request)
-        print(f"Extraction Link Created. Link available at https://www.retab.dev/dashboard/processors/{response['id']}")
+        print(f"Link Created. Url: https://www.retab.dev/dashboard/processors/automations/{response['id']}")
         return Link.model_validate(response)
 
     async def list(

retab/resources/processors/automations/mailboxes.py

@@ -6,7 +6,7 @@ from pydantic import EmailStr, HttpUrl
 from pydantic_core import PydanticUndefined
 
 from ...._resource import AsyncAPIResource, SyncAPIResource
-from ...._utils.mime import prepare_mime_document
+from ....utils.mime import prepare_mime_document
 from ....types.automations.mailboxes import ListMailboxes, Mailbox, UpdateMailboxRequest
 from ....types.mime import EmailData, MIMEData
 from ....types.standards import PreparedRequest
@@ -144,7 +144,7 @@ class Mailboxes(SyncAPIResource, MailBoxesMixin):
         )
         response = self._client._prepared_request(request)
 
-        print(f"Email automation created. Mailbox available at https://www.retab.dev/dashboard/processors/{response['id']}")
+        print(f"Mailbox {response['email']} created. Url: https://www.retab.dev/dashboard/processors/automations/{response['id']}")
 
         return Mailbox.model_validate(response)
 
@@ -278,7 +278,7 @@ class AsyncMailboxes(AsyncAPIResource, MailBoxesMixin):
         )
         response = await self._client._prepared_request(request)
 
-        print(f"Email automation created. Mailbox available at https://www.retab.dev/dashboard/processors/{response['id']}")
+        print(f"Mailbox {response['email']} created. Url: https://www.retab.dev/dashboard/processors/automations/{response['id']}")
 
         return Mailbox.model_validate(response)
 

retab/resources/processors/automations/outlook.py

@@ -149,7 +149,7 @@ class Outlooks(SyncAPIResource, OutlooksMixin):
         )
         response = self._client._prepared_request(request)
 
-        print(f"Outlook automation created. Outlook available at https://www.retab.dev/dashboard/processors/{response['id']}")
+        print(f"Outlook plugin created. Url: https://www.retab.dev/dashboard/processors/automations/{response['id']}")
 
         return Outlook.model_validate(response)
 
@@ -280,7 +280,7 @@ class AsyncOutlooks(AsyncAPIResource, OutlooksMixin):
             fetch_params=fetch_params,
         )
         response = await self._client._prepared_request(request)
-        print(f"Outlook automation created. Outlook available at https://www.retab.dev/dashboard/processors/{response['id']}")
+        print(f"Outlook plugin created. Url: https://www.retab.dev/dashboard/processors/automations/{response['id']}")
         return Outlook.model_validate(response)
 
     async def list(

retab/resources/processors/automations/tests.py

@@ -6,7 +6,7 @@ from PIL.Image import Image
 from pydantic import HttpUrl
 
 from ...._resource import AsyncAPIResource, SyncAPIResource
-from ...._utils.mime import prepare_mime_document
+from ....utils.mime import prepare_mime_document
 from ....types.logs import AutomationLog
 from ....types.mime import MIMEData
 from ....types.standards import PreparedRequest

retab/resources/processors/client.py

@@ -1,7 +1,7 @@
 import base64
 from io import IOBase
 from pathlib import Path
-from typing import Any, Dict, List, Literal
+from typing import Any, List, Literal
 
 import PIL.Image
 from openai.types.chat.chat_completion_reasoning_effort import ChatCompletionReasoningEffort
@@ -9,8 +9,9 @@ from pydantic import BaseModel, HttpUrl
 from pydantic_core import PydanticUndefined
 
 from ..._resource import AsyncAPIResource, SyncAPIResource
-from ..._utils.ai_models import assert_valid_model_extraction
-from ..._utils.mime import MIMEData, prepare_mime_document
+from ...utils.ai_models import assert_valid_model_extraction
+from ...utils.json_schema import load_json_schema
+from ...utils.mime import MIMEData, prepare_mime_document
 from ...types.browser_canvas import BrowserCanvas
 from ...types.documents.extractions import RetabParsedChatCompletion
 from ...types.logs import ProcessorConfig, UpdateProcessorRequest
@@ -31,7 +32,7 @@ class ProcessorsMixin:
     def prepare_create(
         self,
         name: str,
-        json_schema: dict[str, Any],
+        json_schema: dict[str, Any] | Path | str,
         modality: Modality = "native",
         model: str = "gpt-4o-mini",
         temperature: float = PydanticUndefined,  # type: ignore[assignment]
@@ -42,9 +43,12 @@ class ProcessorsMixin:
     ) -> PreparedRequest:
         assert_valid_model_extraction(model)
 
+        # Load the JSON schema from file path, string, or dict
+        loaded_schema = load_json_schema(json_schema)
+
         processor_config = ProcessorConfig(
             name=name,
-            json_schema=json_schema,
+            json_schema=loaded_schema,
             modality=modality,
             model=model,
             temperature=temperature,
@@ -104,7 +108,7 @@ class ProcessorsMixin:
         image_resolution_dpi: int | None = None,
         browser_canvas: BrowserCanvas | None = None,
         model: str | None = None,
-        json_schema: dict[str, Any] | None = None,
+        json_schema: dict[str, Any] | Path | str | None = None,
         temperature: float | None = None,
         reasoning_effort: ChatCompletionReasoningEffort | None = None,
         n_consensus: int | None = None,
@@ -112,13 +116,18 @@ class ProcessorsMixin:
         if model is not None:
             assert_valid_model_extraction(model)
 
+        # Load the JSON schema from file path, string, or dict if provided
+        loaded_schema = None
+        if json_schema is not None:
+            loaded_schema = load_json_schema(json_schema)
+
         update_request = UpdateProcessorRequest(
             name=name,
             modality=modality,
             image_resolution_dpi=image_resolution_dpi,
             browser_canvas=browser_canvas,
             model=model,
-            json_schema=json_schema,
+            json_schema=loaded_schema,
             temperature=temperature,
             reasoning_effort=reasoning_effort,
             n_consensus=n_consensus,
@@ -203,7 +212,7 @@ class Processors(SyncAPIResource, ProcessorsMixin):
     def create(
         self,
         name: str,
-        json_schema: Dict[str, Any],
+        json_schema: dict[str, Any] | Path | str,
         modality: Modality = "native",
         model: str = "gpt-4o-mini",
         temperature: float = PydanticUndefined,  # type: ignore[assignment]
@@ -216,7 +225,7 @@ class Processors(SyncAPIResource, ProcessorsMixin):
 
         Args:
             name: Name of the processor
-            json_schema: JSON schema for the processor
+            json_schema: JSON schema for the processor. Can be a dictionary, file path (Path or str), or JSON string.
             image_resolution_dpi: Optional image resolution DPI
             browser_canvas: Optional browser canvas size
             modality: Processing modality (currently only "native" supported)
@@ -295,7 +304,7 @@ class Processors(SyncAPIResource, ProcessorsMixin):
         image_resolution_dpi: int | None = None,
         browser_canvas: BrowserCanvas | None = None,
         model: str | None = None,
-        json_schema: dict[str, Any] | None = None,
+        json_schema: dict[str, Any] | Path | str | None = None,
         temperature: float | None = None,
         reasoning_effort: ChatCompletionReasoningEffort | None = None,
         n_consensus: int | None = None,
@@ -309,7 +318,7 @@ class Processors(SyncAPIResource, ProcessorsMixin):
             image_resolution_dpi: New image resolution DPI
             browser_canvas: New browser canvas size
             model: New AI model
-            json_schema: New JSON schema for the processor
+            json_schema: New JSON schema for the processor. Can be a dictionary, file path (Path or str), or JSON string.
             temperature: New temperature setting
             reasoning_effort: The effort level for the model to reason about the input data.
             n_consensus: New number of consensus required
@@ -378,7 +387,7 @@ class AsyncProcessors(AsyncAPIResource, ProcessorsMixin):
     async def create(
         self,
         name: str,
-        json_schema: Dict[str, Any],
+        json_schema: dict[str, Any] | Path | str,
         modality: Modality = "native",
         model: str = "gpt-4o-mini",
         temperature: float = PydanticUndefined,  # type: ignore[assignment]
@@ -432,7 +441,7 @@ class AsyncProcessors(AsyncAPIResource, ProcessorsMixin):
         image_resolution_dpi: int | None = None,
         browser_canvas: BrowserCanvas | None = None,
         model: str | None = None,
-        json_schema: dict[str, Any] | None = None,
+        json_schema: dict[str, Any] | Path | str | None = None,
         temperature: float | None = None,
         reasoning_effort: ChatCompletionReasoningEffort | None = None,
         n_consensus: int | None = None,
@@ -446,7 +455,7 @@ class AsyncProcessors(AsyncAPIResource, ProcessorsMixin):
             image_resolution_dpi: New image resolution DPI
             browser_canvas: New browser canvas size
             model: New AI model
-            json_schema: New JSON schema for the processor
+            json_schema: New JSON schema for the processor. Can be a dictionary, file path (Path or str), or JSON string.
             temperature: New temperature setting
             reasoning_effort: The effort level for the model to reason about the input data.
             n_consensus: New number of consensus required

retab/resources/prompt_optimization.py

@@ -3,7 +3,7 @@
 # import json
 
 # from .._resource import SyncAPIResource, AsyncAPIResource
-# from .._utils.json_schema import load_json_schema
+# from ..utils.json_schema import load_json_schema
 # from ..types.jobs import JobResponse
 # from ..types.jobs.prompt_optimization import PromptOptimizationObject, PromptOptimizationProps, PromptOptimizationJobInputData, PromptOptimizationJob
 

retab/resources/schemas.py

@@ -7,9 +7,9 @@ from openai.types.chat.chat_completion_reasoning_effort import ChatCompletionRea
 from pydantic import BaseModel
 
 from .._resource import AsyncAPIResource, SyncAPIResource
-from .._utils.ai_models import assert_valid_model_schema_generation
-from .._utils.json_schema import load_json_schema
-from .._utils.mime import prepare_mime_document_list
+from ..utils.ai_models import assert_valid_model_schema_generation
+from ..utils.json_schema import load_json_schema
+from ..utils.mime import prepare_mime_document_list
 from ..types.mime import MIMEData
 from ..types.modalities import Modality
 from ..types.schemas.enhance import EnhanceSchemaConfig, EnhanceSchemaConfigDict, EnhanceSchemaRequest

retab/types/ai_models.py

@@ -14,7 +14,7 @@ GeminiModel = Literal[
     "gemini-2.5-pro-preview-03-25",
     "gemini-2.5-flash-preview-05-20",
     "gemini-2.5-flash-preview-04-17",
-    "gemini-2.5-flash-lite-preview-06-25",
+    "gemini-2.5-flash-lite-preview-06-17",
     "gemini-2.5-pro-exp-03-25",
     "gemini-2.0-flash-lite",
     "gemini-2.0-flash",

retab/types/completions.py

@@ -7,7 +7,7 @@ from openai.types.shared_params.reasoning import Reasoning
 from openai.types.shared_params.response_format_json_schema import ResponseFormatJSONSchema
 from pydantic import BaseModel, ConfigDict, Field
 
-from .._utils.ai_models import get_provider_for_model
+from ..utils.ai_models import get_provider_for_model
 from .ai_models import AIProvider
 from .chat import ChatCompletionRetabMessage
 

retab/types/documents/create_messages.py

@@ -10,10 +10,10 @@ from openai.types.chat.chat_completion_message_param import ChatCompletionMessag
 from openai.types.responses.response_input_param import ResponseInputItemParam
 from pydantic import BaseModel, Field, computed_field
 
-from ..._utils.chat import convert_to_anthropic_format, convert_to_google_genai_format, str_messages
-from ..._utils.chat import convert_to_openai_format as convert_to_openai_completions_api_format
-from ..._utils.display import count_image_tokens, count_text_tokens
-from ..._utils.responses import convert_to_openai_format as convert_to_openai_responses_api_format
+from ...utils.chat import convert_to_anthropic_format, convert_to_google_genai_format, str_messages
+from ...utils.chat import convert_to_openai_format as convert_to_openai_completions_api_format
+from ...utils.display import count_image_tokens, count_text_tokens
+from ...utils.responses import convert_to_openai_format as convert_to_openai_responses_api_format
 from ..chat import ChatCompletionRetabMessage
 from ..mime import MIMEData
 from ..modalities import Modality

retab/types/documents/extractions.py

@@ -15,7 +15,7 @@ from openai.types.responses.response import Response
 from openai.types.responses.response_input_param import ResponseInputItemParam
 from pydantic import BaseModel, ConfigDict, Field, ValidationInfo, computed_field, field_validator, model_validator
 
-from ..._utils.usage.usage import CostBreakdown, compute_cost_from_model, compute_cost_from_model_with_breakdown
+from ...utils.usage.usage import CostBreakdown, compute_cost_from_model, compute_cost_from_model_with_breakdown
 from ..ai_models import Amount
 from ..chat import ChatCompletionRetabMessage
 from ..mime import MIMEData

retab/types/evals.py

@@ -6,8 +6,8 @@ from typing import Any, List, Literal, Optional
 import nanoid  # type: ignore
 from pydantic import BaseModel, Field, computed_field
 
-from .._utils.json_schema import clean_schema, compute_schema_data_id
-from .._utils.mime import generate_blake2b_hash_from_string
+from ..utils.json_schema import clean_schema, compute_schema_data_id
+from ..utils.mime import generate_blake2b_hash_from_string
 from .ai_models import Amount
 from .inference_settings import InferenceSettings
 from .mime import MIMEData

retab/types/evaluations/iterations.py

@@ -6,8 +6,8 @@ from typing import Any, Optional, Self
 import nanoid  # type: ignore
 from pydantic import BaseModel, Field, computed_field, model_validator
 
-from ..._utils.json_schema import clean_schema
-from ..._utils.mime import generate_blake2b_hash_from_string
+from ...utils.json_schema import clean_schema
+from ...utils.mime import generate_blake2b_hash_from_string
 from ..inference_settings import InferenceSettings
 from ..metrics import MetricResult
 from ..predictions import PredictionData

retab/types/evaluations/model.py

@@ -5,8 +5,8 @@ from typing import Any, Optional
 import nanoid  # type: ignore
 from pydantic import BaseModel, Field, computed_field
 
-from ..._utils.json_schema import compute_schema_data_id
-from ..._utils.mime import generate_blake2b_hash_from_string
+from ...utils.json_schema import compute_schema_data_id
+from ...utils.mime import generate_blake2b_hash_from_string
 from ..inference_settings import InferenceSettings
 from .documents import EvaluationDocument
 from .iterations import Iteration

retab/types/extractions.py

@@ -9,7 +9,7 @@ from pydantic import BaseModel, Field, computed_field, model_validator
 from retab.types.chat import ChatCompletionRetabMessage
 from retab.types.documents.extractions import RetabParsedChatCompletion
 
-from .._utils.usage.usage import CostBreakdown, compute_cost_from_model, compute_cost_from_model_with_breakdown
+from ..utils.usage.usage import CostBreakdown, compute_cost_from_model, compute_cost_from_model_with_breakdown
 from .ai_models import Amount
 from .modalities import Modality
 

retab/types/jobs/prompt_optimization.py

@@ -1,7 +1,7 @@
 # from typing import Literal, Any
 # from pydantic import BaseModel, computed_field
 # from ..mime import MIMEData
-# from ..._utils.benchmarking import ExtractionAnalysis
+# from ...utils.benchmarking import ExtractionAnalysis
 
 # MAX_CONCURRENCY = 15
 

retab/types/logs.py

@@ -7,9 +7,9 @@ from openai.types.chat.chat_completion import ChatCompletion
 from openai.types.chat.chat_completion_reasoning_effort import ChatCompletionReasoningEffort
 from pydantic import BaseModel, EmailStr, Field, HttpUrl, computed_field, field_validator
 
-from .._utils.json_schema import compute_schema_data_id
-from .._utils.mime import generate_blake2b_hash_from_string
-from .._utils.usage.usage import CostBreakdown, compute_cost_from_model, compute_cost_from_model_with_breakdown
+from ..utils.json_schema import compute_schema_data_id
+from ..utils.mime import generate_blake2b_hash_from_string
+from ..utils.usage.usage import CostBreakdown, compute_cost_from_model, compute_cost_from_model_with_breakdown
 from .ai_models import Amount
 from .documents.extractions import RetabParsedChatCompletion
 from .mime import BaseMIMEData

retab/types/schemas/object.py

@@ -10,9 +10,9 @@ from openai.types.chat.chat_completion_message_param import ChatCompletionMessag
 from openai.types.responses.response_input_param import ResponseInputItemParam
 from pydantic import BaseModel, Field, PrivateAttr, computed_field, model_validator
 
-from ..._utils.chat import convert_to_anthropic_format, convert_to_google_genai_format
-from ..._utils.chat import convert_to_openai_format as convert_to_openai_completions_api_format
-from ..._utils.json_schema import (
+from ...utils.chat import convert_to_anthropic_format, convert_to_google_genai_format
+from ...utils.chat import convert_to_openai_format as convert_to_openai_completions_api_format
+from ...utils.json_schema import (
     convert_basemodel_to_partial_basemodel,
     convert_json_schema_to_basemodel,
     create_reasoning_schema,
@@ -25,7 +25,7 @@ from ..._utils.json_schema import (
     load_json_schema,
     schema_to_ts_type,
 )
-from ..._utils.responses import convert_to_openai_format as convert_to_openai_responses_api_format
+from ...utils.responses import convert_to_openai_format as convert_to_openai_responses_api_format
 from ...types.standards import StreamingBaseModel
 from ..chat import ChatCompletionRetabMessage
 

retab/types/schemas/templates.py

@@ -4,7 +4,7 @@ from typing import Any, Literal, Optional
 import nanoid  # type: ignore
 from pydantic import BaseModel, Field, PrivateAttr, computed_field
 
-from ..._utils.json_schema import generate_schema_data_id, generate_schema_id
+from ...utils.json_schema import generate_schema_data_id, generate_schema_id
 from ...types.mime import MIMEData
 
 

retab/utils/_model_cards/anthropic.yaml

@@ -0,0 +1,59 @@
+- model: "claude-3-5-sonnet-latest"
+  pricing:
+    text:
+      prompt: 3.00
+      cached_discount: 0.5
+      completion: 15.00
+    audio: null
+  capabilities:
+    modalities: ["text", "image"]
+    endpoints: ["chat_completions"]
+    features: ["streaming", "function_calling"]
+  permissions:
+    show_in_free_picker: true
+    show_in_paid_picker: true
+
+- model: "claude-3-5-sonnet-20241022"
+  inherits: "claude-3-5-sonnet-latest"
+  permissions:
+    show_in_free_picker: false
+    show_in_paid_picker: false
+
+- model: "claude-3-opus-20240229"
+  pricing:
+    text:
+      prompt: 15.00
+      cached_discount: 0.5
+      completion: 75.00
+    audio: null
+  capabilities:
+    modalities: ["text", "image"]
+    endpoints: ["chat_completions"]
+    features: ["streaming", "function_calling"]
+  permissions:
+    show_in_free_picker: true
+    show_in_paid_picker: true
+
+- model: "claude-3-sonnet-20240229"
+  pricing:
+    text:
+      prompt: 3.00
+      cached_discount: 0.5
+      completion: 15.00
+    audio: null
+  capabilities:
+    modalities: ["text", "image"]
+    endpoints: ["chat_completions"]
+    features: ["streaming", "function_calling"]
+
+- model: "claude-3-haiku-20240307"
+  pricing:
+    text:
+      prompt: 0.25
+      cached_discount: 0.5
+      completion: 1.25
+    audio: null
+  capabilities:
+    modalities: ["text", "image"]
+    endpoints: ["chat_completions"]
+    features: ["streaming", "function_calling"]

retab/utils/_model_cards/auto.yaml

@@ -0,0 +1,43 @@
+- model: "auto-large"
+  pricing:
+    text:
+      prompt: 2.00
+      cached_discount: 0.25
+      completion: 8.00
+    audio: null
+    ft_price_hike: 1.5
+  capabilities:
+    modalities: ["text", "image"]
+    endpoints: ["chat_completions"]
+    features: ["streaming", "function_calling", "structured_outputs"]
+  temperature_support: true
+
+- model: "auto-small"
+  pricing:
+    text:
+      prompt: 0.4
+      cached_discount: 0.25
+      completion: 1.6
+    audio:
+      prompt: 0.7
+      cached_discount: 0.175
+      completion: 1000
+  capabilities:
+    modalities: ["text", "image"]
+    endpoints: ["chat_completions"]
+    features: ["streaming", "function_calling", "structured_outputs"]
+  temperature_support: true
+
+- model: "auto-micro"
+  pricing:
+    text:
+      prompt: 0.1
+      cached_discount: 0.25
+      completion: 0.4
+    audio: null
+    ft_price_hike: 1.5
+  capabilities:
+    modalities: ["text", "image"]
+    endpoints: ["chat_completions"]
+    features: ["streaming", "function_calling", "structured_outputs"]
+  temperature_support: true