retab 0.0.36__py3-none-any.whl → 0.0.38__py3-none-any.whl

{uiform → retab}/_utils/usage/usage.py

@@ -1,4 +1,4 @@
-from typing import Optional, Dict
+from typing import Optional
 
 from openai.types.completion_usage import CompletionUsage
 from pydantic import BaseModel, Field
@@ -71,16 +71,14 @@ def compute_api_call_cost(pricing: Pricing, usage: CompletionUsage, is_ft: bool
     total_cost = (total_text_cost + total_audio_cost) / 1e6
 
     # Apply fine-tuning price hike if applicable
-    if is_ft and hasattr(pricing, 'ft_price_hike'):
+    if is_ft and hasattr(pricing, "ft_price_hike"):
         total_cost *= pricing.ft_price_hike
 
     return Amount(value=total_cost, currency="USD")
 
 
-
-
 def compute_cost_from_model(model: str, usage: CompletionUsage) -> Amount:
-    # Extract base model name for fine-tuned models like "ft:gpt-4o:uiform:4389573"
+    # Extract base model name for fine-tuned models like "ft:gpt-4o:retab:4389573"
     is_ft = False
     if model.startswith("ft:"):
         # Split by colon and take the second part (index 1) which contains the base model
@@ -93,7 +91,7 @@ def compute_cost_from_model(model: str, usage: CompletionUsage) -> Amount:
     try:
         model_card = get_model_card(model)
         pricing = model_card.pricing
-    except ValueError as e:
+    except ValueError:
         raise ValueError(f"No pricing information found for model: {model}")
 
     return compute_api_call_cost(pricing, usage, is_ft)
@@ -124,46 +122,48 @@ class CompletionsUsage(BaseModel):
     model: Optional[str] = Field(default=None, description="When group_by=model, this field provides the model name of the grouped usage result.")
     batch: Optional[bool] = Field(default=None, description="When group_by=batch, this field tells whether the grouped usage result is batch or not.")
 
+
 ########################
 # DETAILED COST BREAKDOWN
 ########################
 
+
 class TokenCounts(BaseModel):
     """Detailed breakdown of token counts by type and category."""
-    
+
     # Prompt token counts
     prompt_regular_text: int
     prompt_cached_text: int
     prompt_audio: int
-    
+
     # Completion token counts
     completion_regular_text: int
     completion_audio: int
-    
+
     # Total tokens (should match sum of all components)
     total_tokens: int
 
 
 class CostBreakdown(BaseModel):
     """Detailed breakdown of API call costs by token type and usage category."""
-    
+
     # Total cost amount
     total: Amount
-    
+
     # Text token costs broken down by category
     text_prompt_cost: Amount
     text_cached_cost: Amount
     text_completion_cost: Amount
     text_total_cost: Amount
-    
+
     # Audio token costs broken down by category (if applicable)
     audio_prompt_cost: Optional[Amount] = None
     audio_completion_cost: Optional[Amount] = None
     audio_total_cost: Optional[Amount] = None
-    
+
     # Token counts for reference
     token_counts: TokenCounts
-    
+
     # Model and fine-tuning information
     model: str
     is_fine_tuned: bool = False
@@ -172,7 +172,7 @@ class CostBreakdown(BaseModel):
 def compute_api_call_cost_with_breakdown(pricing: Pricing, usage: CompletionUsage, model: str, is_ft: bool = False) -> CostBreakdown:
     """
     Computes a detailed price breakdown for the given token usage, based on the pricing.
-    
+
     Returns a CostBreakdown object containing costs broken down by token type and category.
     """
     # ----- Process prompt tokens -----
@@ -211,7 +211,7 @@ def compute_api_call_cost_with_breakdown(pricing: Pricing, usage: CompletionUsag
     cost_audio_prompt = 0.0
     cost_audio_completion = 0.0
     total_audio_cost = 0.0
-    
+
     if pricing.audio and (prompt_audio > 0 or completion_audio > 0):
         cost_audio_prompt = prompt_audio * pricing.audio.prompt
         cost_audio_completion = completion_audio * pricing.audio.completion
@@ -219,27 +219,27 @@ def compute_api_call_cost_with_breakdown(pricing: Pricing, usage: CompletionUsag
 
     # Convert to dollars (divide by 1M) and create Amount objects
     ft_multiplier = pricing.ft_price_hike if is_ft else 1.0
-    
+
     # Create Amount objects for each cost category
     text_prompt_amount = Amount(value=(cost_text_prompt / 1e6) * ft_multiplier, currency="USD")
     text_cached_amount = Amount(value=(cost_text_cached / 1e6) * ft_multiplier, currency="USD")
     text_completion_amount = Amount(value=(cost_text_completion / 1e6) * ft_multiplier, currency="USD")
     text_total_amount = Amount(value=(total_text_cost / 1e6) * ft_multiplier, currency="USD")
-    
+
     # Audio amounts (if applicable)
     audio_prompt_amount = None
     audio_completion_amount = None
     audio_total_amount = None
-    
+
     if pricing.audio and (prompt_audio > 0 or completion_audio > 0):
         audio_prompt_amount = Amount(value=(cost_audio_prompt / 1e6) * ft_multiplier, currency="USD")
         audio_completion_amount = Amount(value=(cost_audio_completion / 1e6) * ft_multiplier, currency="USD")
         audio_total_amount = Amount(value=(total_audio_cost / 1e6) * ft_multiplier, currency="USD")
-    
+
     # Total cost
     total_cost = (total_text_cost + total_audio_cost) / 1e6 * ft_multiplier
     total_amount = Amount(value=total_cost, currency="USD")
-    
+
     # Create TokenCounts object with token usage breakdown
     token_counts = TokenCounts(
         prompt_regular_text=prompt_regular_text,
@@ -247,9 +247,9 @@ def compute_api_call_cost_with_breakdown(pricing: Pricing, usage: CompletionUsag
         prompt_audio=prompt_audio,
         completion_regular_text=completion_regular_text,
         completion_audio=completion_audio,
-        total_tokens=usage.total_tokens
+        total_tokens=usage.total_tokens,
     )
-    
+
     return CostBreakdown(
         total=total_amount,
         text_prompt_cost=text_prompt_amount,
@@ -261,28 +261,28 @@ def compute_api_call_cost_with_breakdown(pricing: Pricing, usage: CompletionUsag
         audio_total_cost=audio_total_amount,
         token_counts=token_counts,
         model=model,
-        is_fine_tuned=is_ft
+        is_fine_tuned=is_ft,
     )
 
 
 def compute_cost_from_model_with_breakdown(model: str, usage: CompletionUsage) -> CostBreakdown:
     """
     Computes a detailed cost breakdown for an API call using the specified model and usage.
-    
+
     Args:
-        model: The model name (can be a fine-tuned model like "ft:gpt-4o:uiform:4389573")
+        model: The model name (can be a fine-tuned model like "ft:gpt-4o:retab:4389573")
         usage: Token usage statistics for the API call
-        
+
     Returns:
         CostBreakdown object with detailed cost information
-        
+
     Raises:
         ValueError: If no pricing information is found for the model
     """
-    # Extract base model name for fine-tuned models like "ft:gpt-4o:uiform:4389573"
+    # Extract base model name for fine-tuned models like "ft:gpt-4o:retab:4389573"
     original_model = model
     is_ft = False
-    
+
     if model.startswith("ft:"):
         # Split by colon and take the second part (index 1) which contains the base model
         parts = model.split(":")
@@ -294,7 +294,7 @@ def compute_cost_from_model_with_breakdown(model: str, usage: CompletionUsage) -
     try:
         model_card = get_model_card(model)
         pricing = model_card.pricing
-    except ValueError as e:
+    except ValueError:
         raise ValueError(f"No pricing information found for model: {original_model}")
 
     return compute_api_call_cost_with_breakdown(pricing, usage, original_model, is_ft)

{uiform → retab}/client.py

@@ -1,16 +1,16 @@
 import json
 import os
 from types import TracebackType
-from typing import Any, AsyncIterator, BinaryIO, Iterator, List, Optional
+from typing import Any, AsyncIterator, Iterator, Optional
 
 import backoff
 import backoff.types
 import httpx
+import truststore
 from pydantic_core import PydanticUndefined
 
-from .resources import processors, consensus, documents, evals, files, finetuning, models, schemas, secrets, usage
+from .resources import consensus, documents, evals, files, finetuning, models, processors, schemas, secrets, usage, evaluations
 from .types.standards import PreparedRequest
-import truststore
 
 
 class MaxRetriesExceeded(Exception):
@@ -26,15 +26,15 @@ def raise_max_tries_exceeded(details: backoff.types.Details) -> None:
         raise Exception(f"Max tries exceeded after {tries} tries.")
 
 
-class BaseUiForm:
-    """Base class for UiForm clients that handles authentication and configuration.
+class BaseRetab:
+    """Base class for Retab clients that handles authentication and configuration.
 
     This class provides core functionality for API authentication, configuration, and common HTTP operations
     used by both synchronous and asynchronous clients.
 
     Args:
-        api_key (str, optional): UiForm API key. If not provided, will look for UIFORM_API_KEY env variable.
-        base_url (str, optional): Base URL for API requests. Defaults to https://api.uiform.com
+        api_key (str, optional): Retab API key. If not provided, will look for RETAB_API_KEY env variable.
+        base_url (str, optional): Base URL for API requests. Defaults to https://api.retab.dev
         timeout (float): Request timeout in seconds. Defaults to 240.0
         max_retries (int): Maximum number of retries for failed requests. Defaults to 3
         openai_api_key (str, optional): OpenAI API key. Will look for OPENAI_API_KEY env variable if not provided
@@ -59,17 +59,16 @@ class BaseUiForm:
         xai_api_key: Optional[str] = PydanticUndefined,  # type: ignore[assignment]
     ) -> None:
         if api_key is None:
-            api_key = os.environ.get("UIFORM_API_KEY")
+            api_key = os.environ.get("RETAB_API_KEY")
 
         if api_key is None:
             raise ValueError(
-                "No API key provided. You can create an API key at https://uiform.com\n"
-                "Then either pass it to the client (api_key='your-key') or set the UIFORM_API_KEY environment variable"
+                "No API key provided. You can create an API key at https://retab.dev\n"
+                "Then either pass it to the client (api_key='your-key') or set the RETAB_API_KEY environment variable"
             )
 
         if base_url is None:
-            base_url = os.environ.get("UIFORM_API_BASE_URL", "https://api.uiform.com")
-
+            base_url = os.environ.get("RETAB_API_BASE_URL", "https://api.retab.dev")
 
         truststore.inject_into_ssl()
         self.api_key = api_key
@@ -84,26 +83,26 @@ class BaseUiForm:
         # Only check environment variables if the value is PydanticUndefined
         if openai_api_key is PydanticUndefined:
             openai_api_key = os.environ.get("OPENAI_API_KEY")
-        
+
         # if claude_api_key is PydanticUndefined:
         #    claude_api_key = os.environ.get("CLAUDE_API_KEY")
-        
+
         # if xai_api_key is PydanticUndefined:
         #    xai_api_key = os.environ.get("XAI_API_KEY")
-        
+
         if gemini_api_key is PydanticUndefined:
             gemini_api_key = os.environ.get("GEMINI_API_KEY")
 
         # Only add headers if the values are actual strings (not None or PydanticUndefined)
         if openai_api_key and openai_api_key is not PydanticUndefined:
             self.headers["OpenAI-Api-Key"] = openai_api_key
-        
+
         # if claude_api_key and claude_api_key is not PydanticUndefined:
         #    self.headers["Anthropic-Api-Key"] = claude_api_key
-        
+
         if xai_api_key and xai_api_key is not PydanticUndefined:
             self.headers["XAI-Api-Key"] = xai_api_key
-        
+
         if gemini_api_key and gemini_api_key is not PydanticUndefined:
             self.headers["Gemini-Api-Key"] = gemini_api_key
 
@@ -126,12 +125,12 @@ class BaseUiForm:
 
     def _parse_response(self, response: httpx.Response) -> Any:
         """Parse response based on content-type.
-        
+
         Returns:
             Any: Parsed JSON object for JSON responses, raw text string for text responses
         """
         content_type = response.headers.get("content-type", "")
-        
+
         # Check if it's a JSON response
         if "application/json" in content_type or "application/stream+json" in content_type:
             return response.json()
@@ -147,15 +146,15 @@ class BaseUiForm:
                 return response.text
 
 
-class UiForm(BaseUiForm):
-    """Synchronous client for interacting with the UiForm API.
+class Retab(BaseRetab):
+    """Synchronous client for interacting with the Retab API.
 
-    This client provides synchronous access to all UiForm API resources including files, fine-tuning,
+    This client provides synchronous access to all Retab API resources including files, fine-tuning,
     prompt optimization, documents, models, datasets, and schemas.
 
     Args:
-        api_key (str, optional): UiForm API key. If not provided, will look for UIFORM_API_KEY env variable.
-        base_url (str, optional): Base URL for API requests. Defaults to https://api.uiform.com
+        api_key (str, optional): Retab API key. If not provided, will look for RETAB_API_KEY env variable.
+        base_url (str, optional): Base URL for API requests. Defaults to https://api.retab.dev
         timeout (float): Request timeout in seconds. Defaults to 240.0
         max_retries (int): Maximum number of retries for failed requests. Defaults to 3
         openai_api_key (str, optional): OpenAI API key. Will look for OPENAI_API_KEY env variable if not provided
@@ -198,6 +197,7 @@ class UiForm(BaseUiForm):
 
         self.client = httpx.Client(timeout=self.timeout)
         self.evals = evals.Evals(client=self)
+        self.evaluations = evaluations.Evaluations(client=self)
         self.files = files.Files(client=self)
         self.fine_tuning = finetuning.FineTuning(client=self)
         # self.prompt_optimization = prompt_optimization.PromptOptimization(client=self)
@@ -247,7 +247,7 @@ class UiForm(BaseUiForm):
                 "params": params,
                 "headers": self._get_headers(idempotency_key),
             }
-            
+
             # Handle different content types
             if files or form_data:
                 # For multipart/form-data requests
@@ -262,7 +262,7 @@ class UiForm(BaseUiForm):
             elif data:
                 # For JSON requests
                 request_kwargs["json"] = data
-            
+
             response = self.client.request(**request_kwargs)
             self._validate_response(response)
             return self._parse_response(response)
@@ -314,7 +314,7 @@ class UiForm(BaseUiForm):
                 "params": params,
                 "headers": self._get_headers(idempotency_key),
             }
-            
+
             # Handle different content types
             if files or form_data:
                 # For multipart/form-data requests
@@ -329,10 +329,10 @@ class UiForm(BaseUiForm):
             elif data:
                 # For JSON requests
                 stream_kwargs["json"] = data
-            
+
             with self.client.stream(**stream_kwargs) as response_ctx_manager:
                 self._validate_response(response_ctx_manager)
-                
+
                 content_type = response_ctx_manager.headers.get("content-type", "")
                 is_json_stream = "application/json" in content_type or "application/stream+json" in content_type
                 is_text_stream = "text/plain" in content_type or ("text/" in content_type and not is_json_stream)
@@ -340,7 +340,7 @@ class UiForm(BaseUiForm):
                 for chunk in response_ctx_manager.iter_lines():
                     if not chunk:
                         continue
-                    
+
                     if is_json_stream:
                         try:
                             yield json.loads(chunk)
@@ -375,7 +375,7 @@ class UiForm(BaseUiForm):
             form_data=request.form_data,
             files=request.files,
             idempotency_key=request.idempotency_key,
-            raise_for_status=request.raise_for_status
+            raise_for_status=request.raise_for_status,
         )
 
     def _prepared_request_stream(self, request: PreparedRequest) -> Iterator[Any]:
@@ -387,7 +387,7 @@ class UiForm(BaseUiForm):
             form_data=request.form_data,
             files=request.files,
             idempotency_key=request.idempotency_key,
-            raise_for_status=request.raise_for_status
+            raise_for_status=request.raise_for_status,
         ):
             yield item
 
@@ -395,11 +395,11 @@ class UiForm(BaseUiForm):
         """Closes the HTTP client session."""
         self.client.close()
 
-    def __enter__(self) -> "UiForm":
+    def __enter__(self) -> "Retab":
         """Context manager entry point.
 
         Returns:
-            UiForm: The client instance
+            Retab: The client instance
         """
         return self
 
@@ -414,15 +414,15 @@ class UiForm(BaseUiForm):
         self.close()
 
 
-class AsyncUiForm(BaseUiForm):
-    """Asynchronous client for interacting with the UiForm API.
+class AsyncRetab(BaseRetab):
+    """Asynchronous client for interacting with the Retab API.
 
-    This client provides asynchronous access to all UiForm API resources including files, fine-tuning,
+    This client provides asynchronous access to all Retab API resources including files, fine-tuning,
     prompt optimization, documents, models, datasets, and schemas.
 
     Args:
-        api_key (str, optional): UiForm API key. If not provided, will look for UIFORM_API_KEY env variable.
-        base_url (str, optional): Base URL for API requests. Defaults to https://api.uiform.com
+        api_key (str, optional): Retab API key. If not provided, will look for RETAB_API_KEY env variable.
+        base_url (str, optional): Base URL for API requests. Defaults to https://api.retab.dev
         timeout (float): Request timeout in seconds. Defaults to 240.0
         max_retries (int): Maximum number of retries for failed requests. Defaults to 3
         openai_api_key (str, optional): OpenAI API key. Will look for OPENAI_API_KEY env variable if not provided
@@ -466,6 +466,7 @@ class AsyncUiForm(BaseUiForm):
         self.client = httpx.AsyncClient(timeout=self.timeout)
 
         self.evals = evals.AsyncEvals(client=self)
+        self.evaluations = evaluations.AsyncEvaluations(client=self)
         self.files = files.AsyncFiles(client=self)
         self.fine_tuning = finetuning.AsyncFineTuning(client=self)
         # self.prompt_optimization = prompt_optimization.AsyncPromptOptimization(client=self)
@@ -479,12 +480,12 @@ class AsyncUiForm(BaseUiForm):
 
     def _parse_response(self, response: httpx.Response) -> Any:
         """Parse response based on content-type.
-        
+
         Returns:
             Any: Parsed JSON object for JSON responses, raw text string for text responses
         """
         content_type = response.headers.get("content-type", "")
-        
+
         # Check if it's a JSON response
         if "application/json" in content_type or "application/stream+json" in content_type:
             return response.json()
@@ -536,7 +537,7 @@ class AsyncUiForm(BaseUiForm):
                 "params": params,
                 "headers": self._get_headers(idempotency_key),
             }
-            
+
             # Handle different content types
             if files or form_data:
                 # For multipart/form-data requests
@@ -551,7 +552,7 @@ class AsyncUiForm(BaseUiForm):
             elif data:
                 # For JSON requests
                 request_kwargs["json"] = data
-            
+
             response = await self.client.request(**request_kwargs)
             self._validate_response(response)
             return self._parse_response(response)
@@ -602,7 +603,7 @@ class AsyncUiForm(BaseUiForm):
                 "params": params,
                 "headers": self._get_headers(idempotency_key),
             }
-            
+
             # Handle different content types
             if files or form_data:
                 # For multipart/form-data requests
@@ -617,18 +618,18 @@ class AsyncUiForm(BaseUiForm):
             elif data:
                 # For JSON requests
                 stream_kwargs["json"] = data
-            
+
             async with self.client.stream(**stream_kwargs) as response_ctx_manager:
                 self._validate_response(response_ctx_manager)
-                
+
                 content_type = response_ctx_manager.headers.get("content-type", "")
                 is_json_stream = "application/json" in content_type or "application/stream+json" in content_type
                 is_text_stream = "text/plain" in content_type or ("text/" in content_type and not is_json_stream)
-                
+
                 async for chunk in response_ctx_manager.aiter_lines():
                     if not chunk:
                         continue
-                    
+
                     if is_json_stream:
                         try:
                             yield json.loads(chunk)
@@ -662,7 +663,7 @@ class AsyncUiForm(BaseUiForm):
             form_data=request.form_data,
             files=request.files,
             idempotency_key=request.idempotency_key,
-            raise_for_status=request.raise_for_status
+            raise_for_status=request.raise_for_status,
         )
 
     async def _prepared_request_stream(self, request: PreparedRequest) -> AsyncIterator[Any]:
@@ -674,7 +675,7 @@ class AsyncUiForm(BaseUiForm):
             form_data=request.form_data,
             files=request.files,
             idempotency_key=request.idempotency_key,
-            raise_for_status=request.raise_for_status
+            raise_for_status=request.raise_for_status,
         ):
             yield item
 
@@ -682,11 +683,11 @@ class AsyncUiForm(BaseUiForm):
         """Closes the async HTTP client session."""
         await self.client.aclose()
 
-    async def __aenter__(self) -> "AsyncUiForm":
+    async def __aenter__(self) -> "AsyncRetab":
         """Async context manager entry point.
 
         Returns:
-            AsyncUiForm: The async client instance
+            AsyncRetab: The async client instance
         """
         return self
 

{uiform → retab}/resources/consensus/client.py

@@ -1,15 +1,13 @@
 from typing import Any, Dict, List, Literal, Optional
 
-from pydantic import BaseModel, Field
-
 from ..._resource import AsyncAPIResource, SyncAPIResource
+from ...types.consensus import ReconciliationResponse, ReconciliationRequest
 from ...types.standards import PreparedRequest
 from .completions import AsyncCompletions, Completions
 from .responses import AsyncResponses, Responses
-from ...types.consensus import ReconciliationResponse
 
-class BaseConsensusMixin:
 
+class BaseConsensusMixin:
     def _prepare_reconcile(
         self,
         list_dicts: List[Dict[str, Any]],
@@ -17,21 +15,15 @@ class BaseConsensusMixin:
         mode: Literal["direct", "aligned"] = "direct",
         idempotency_key: str | None = None,
     ) -> PreparedRequest:
-        data = {
-            "list_dicts": list_dicts,
-            "mode": mode,
-        }
-        
-        if reference_schema is not None:
-            data["reference_schema"] = reference_schema
-            
-        return PreparedRequest(
-            method="POST", 
-            url="/v1/consensus/reconcile", 
-            data=data, 
-            idempotency_key=idempotency_key
+        request = ReconciliationRequest(
+            list_dicts=list_dicts,
+            reference_schema=reference_schema,
+            mode=mode,
         )
 
+        return PreparedRequest(method="POST", url="/v1/consensus/reconcile", data=request.model_dump(), idempotency_key=idempotency_key)
+
+
 class Consensus(SyncAPIResource, BaseConsensusMixin):
     """Consensus API wrapper for synchronous operations"""
 
@@ -49,25 +41,20 @@ class Consensus(SyncAPIResource, BaseConsensusMixin):
     ) -> ReconciliationResponse:
         """
         Reconcile multiple dictionaries to produce a single unified consensus dictionary.
-        
+
         Args:
             list_dicts: List of dictionaries to reconcile
             reference_schema: Optional schema to validate dictionaries against
             mode: Mode for consensus computation ("direct" or "aligned")
             idempotency_key: Optional idempotency key for the request
-            
+
         Returns:
             Dict containing the consensus dictionary and consensus likelihoods
-            
+
         Raises:
-            UiformAPIError: If the API request fails
+            RetabAPIError: If the API request fails
         """
-        request = self._prepare_reconcile(
-            list_dicts, 
-            reference_schema, 
-            mode, 
-            idempotency_key
-        )
+        request = self._prepare_reconcile(list_dicts, reference_schema, mode, idempotency_key)
         response = self._client._prepared_request(request)
         return ReconciliationResponse.model_validate(response)
 
@@ -80,7 +67,6 @@ class AsyncConsensus(AsyncAPIResource, BaseConsensusMixin):
         self.completions = AsyncCompletions(client=client)
         self.responses = AsyncResponses(client=client)
 
-    
     async def reconcile(
         self,
         list_dicts: List[Dict[str, Any]],
@@ -90,25 +76,20 @@ class AsyncConsensus(AsyncAPIResource, BaseConsensusMixin):
     ) -> ReconciliationResponse:
         """
         Reconcile multiple dictionaries to produce a single unified consensus dictionary asynchronously.
-        
+
         Args:
             list_dicts: List of dictionaries to reconcile
             reference_schema: Optional schema to validate dictionaries against
             mode: Mode for consensus computation ("direct" or "aligned")
             idempotency_key: Optional idempotency key for the request
-            
+
         Returns:
             Dict containing the consensus dictionary and consensus likelihoods
-            
+
         Raises:
-            UiformAPIError: If the API request fails
+            RetabAPIError: If the API request fails
         """
-        request = self._prepare_reconcile(
-            list_dicts, 
-            reference_schema, 
-            mode, 
-            idempotency_key
-        )
+        request = self._prepare_reconcile(list_dicts, reference_schema, mode, idempotency_key)
         response = await self._client._prepared_request(request)
 
         return ReconciliationResponse.model_validate(response)