retab 0.0.90__tar.gz → 0.0.92__tar.gz

{retab-0.0.90 → retab-0.0.92}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: retab
-Version: 0.0.90
+Version: 0.0.92
 Summary: Retab official python library
 Home-page: https://github.com/retab-dev/retab
 Author: Retab

{retab-0.0.90 → retab-0.0.92}/retab/resources/documents/client.py

@@ -16,7 +16,8 @@ from ...types.chat import ChatCompletionRetabMessage
 from ...types.documents.edit import EditRequest, EditResponse
 from ...types.documents.extract import DocumentExtractRequest, RetabParsedChatCompletion, RetabParsedChatCompletionChunk, RetabParsedChoice, maybe_parse_to_pydantic
 from ...types.documents.parse import ParseRequest, ParseResult, TableParsingFormat
-from ...types.documents.split import Category, SplitRequest, SplitResponse
+from ...types.documents.split import Subdocument, SplitRequest, SplitResponse
+from ...types.documents.classify import Category
 from ...types.documents.classify import ClassifyRequest, ClassifyResponse
 from ...types.mime import MIMEData
 from ...types.standards import PreparedRequest, FieldUnset
@@ -148,21 +149,21 @@ class BaseDocumentsMixin:
     def _prepare_split(
         self,
         document: Path | str | IOBase | MIMEData | PIL.Image.Image | HttpUrl,
-        categories: list[Category] | list[dict[str, str]],
+        subdocuments: list[Subdocument] | list[dict[str, str]],
         model: str,
         **extra_body: Any,
     ) -> PreparedRequest:
         mime_document = prepare_mime_document(document)
         
-        # Convert dict categories to Category objects if needed
-        category_objects = [
-            Category(**cat) if isinstance(cat, dict) else cat
-            for cat in categories
+        # Convert dict subdocuments to Subdocument objects if needed
+        subdocument_objects = [
+            Subdocument(**subdoc) if isinstance(subdoc, dict) else subdoc
+            for subdoc in subdocuments
         ]
         
         request_dict: dict[str, Any] = {
             "document": mime_document,
-            "categories": category_objects,
+            "subdocuments": subdocument_objects,
             "model": model,
         }
 
@@ -644,20 +645,20 @@ class Documents(SyncAPIResource, BaseDocumentsMixin):
     def split(
         self,
         document: Path | str | IOBase | MIMEData | PIL.Image.Image | HttpUrl,
-        categories: list[Category] | list[dict[str, str]],
+        subdocuments: list[Subdocument] | list[dict[str, str]],
         model: str,
         **extra_body: Any,
     ) -> SplitResponse:
         """
-        Split a document into sections based on provided categories.
+        Split a document into sections based on provided subdocuments.
 
         This method analyzes a multi-page document and classifies pages into 
-        user-defined categories, returning the page ranges for each section.
+        user-defined subdocuments, returning the page ranges for each section.
 
         Args:
             document: The document to split. Can be a file path (Path or str), file-like object, MIMEData, PIL Image, or URL.
-            categories: List of categories to split the document into. Each category should have a 'name' and 'description'.
-                Can be Category objects or dicts with 'name' and 'description' keys.
+            subdocuments: List of subdocuments to split the document into. Each subdocument should have a 'name' and 'description'.
+                Can be Subdocument objects or dicts with 'name' and 'description' keys.
             model: The AI model to use for document splitting (e.g., "gemini-2.5-flash").
 
         Returns:
@@ -672,7 +673,7 @@ class Documents(SyncAPIResource, BaseDocumentsMixin):
             response = retab.documents.split(
                 document="invoice_batch.pdf",
                 model="gemini-2.5-flash",
-                categories=[
+                subdocuments=[
                     {"name": "invoice", "description": "Invoice documents with billing information"},
                     {"name": "receipt", "description": "Receipt documents for payments"},
                     {"name": "contract", "description": "Legal contract documents"},
@@ -684,7 +685,7 @@ class Documents(SyncAPIResource, BaseDocumentsMixin):
         """
         request = self._prepare_split(
             document=document,
-            categories=categories,
+            subdocuments=subdocuments,
             model=model,
             **extra_body,
         )
@@ -1039,20 +1040,20 @@ class AsyncDocuments(AsyncAPIResource, BaseDocumentsMixin):
     async def split(
         self,
         document: Path | str | IOBase | MIMEData | PIL.Image.Image | HttpUrl,
-        categories: list[Category] | list[dict[str, str]],
+        subdocuments: list[Subdocument] | list[dict[str, str]],
         model: str,
         **extra_body: Any,
     ) -> SplitResponse:
         """
-        Split a document into sections based on provided categories asynchronously.
+        Split a document into sections based on provided subdocuments asynchronously.
 
         This method analyzes a multi-page document and classifies pages into 
-        user-defined categories, returning the page ranges for each section.
+        user-defined subdocuments, returning the page ranges for each section.
 
         Args:
             document: The document to split. Can be a file path (Path or str), file-like object, MIMEData, PIL Image, or URL.
-            categories: List of categories to split the document into. Each category should have a 'name' and 'description'.
-                Can be Category objects or dicts with 'name' and 'description' keys.
+            subdocuments: List of subdocuments to split the document into. Each subdocument should have a 'name' and 'description'.
+                Can be Subdocument objects or dicts with 'name' and 'description' keys.
             model: The AI model to use for document splitting (e.g., "gemini-2.5-flash").
 
         Returns:
@@ -1067,7 +1068,7 @@ class AsyncDocuments(AsyncAPIResource, BaseDocumentsMixin):
             response = await retab.documents.split(
                 document="invoice_batch.pdf",
                 model="gemini-2.5-flash",
-                categories=[
+                subdocuments=[
                     {"name": "invoice", "description": "Invoice documents with billing information"},
                     {"name": "receipt", "description": "Receipt documents for payments"},
                     {"name": "contract", "description": "Legal contract documents"},
@@ -1079,7 +1080,7 @@ class AsyncDocuments(AsyncAPIResource, BaseDocumentsMixin):
         """
         request = self._prepare_split(
             document=document,
-            categories=categories,
+            subdocuments=subdocuments,
             model=model,
             **extra_body,
         )

{retab-0.0.90 → retab-0.0.92}/retab/types/documents/__init__.py

@@ -1,12 +1,14 @@
 from .parse import ParseRequest, ParseResult, RetabUsage
-from .split import Category, SplitRequest, SplitResult, SplitResponse
-from .classify import ClassifyRequest, ClassifyResult, ClassifyResponse
+from .split import Subdocument, SplitRequest, SplitResult, SplitResponse
+from .classify import ClassifyRequest, ClassifyResult, ClassifyResponse, Category
+
 
 __all__ = [
     "ParseRequest", 
     "ParseResult", 
     "RetabUsage",
     "Category",
+    "Subdocument",
     "SplitRequest",
     "SplitResult",
     "SplitResponse",

{retab-0.0.90 → retab-0.0.92}/retab/types/documents/classify.py

@@ -1,13 +1,16 @@
 from pydantic import BaseModel, Field
 from ..mime import MIMEData
-from .split import Category
 
+class Category(BaseModel):
+    name: str = Field(..., description="The name of the category")
+    description: str = Field(..., description="The description of the category")
 
 class ClassifyRequest(BaseModel):
     document: MIMEData = Field(..., description="The document to classify")
     categories: list[Category] = Field(..., description="The categories to classify the document into")
     model: str = Field(default="retab-small", description="The model to use for classification")
     first_n_pages: int | None = Field(default=None, description="Only use the first N pages of the document for classification. Useful for large documents where classification can be determined from early pages.")
+    context: str | None = Field(default=None, description="Additional context for classification (e.g., iteration context from a loop)")
 
 
 class ClassifyResult(BaseModel):

{retab-0.0.90 → retab-0.0.92}/retab/types/documents/split.py

@@ -1,17 +1,17 @@
 from pydantic import BaseModel, Field
 from ..mime import MIMEData
 
-
-class Category(BaseModel):
-    name: str = Field(..., description="The name of the category")
-    description: str = Field(..., description="The description of the category")
-    partition_key: str | None = Field(default=None, description="The key to partition the category")
+class Subdocument(BaseModel):
+    name: str = Field(..., description="The name of the subdocument")
+    description: str = Field(..., description="The description of the subdocument")
+    partition_key: str | None = Field(default=None, description="The key to partition the subdocument")
 
 
 class SplitRequest(BaseModel):
     document: MIMEData = Field(..., description="The document to split")
-    categories: list[Category] = Field(..., description="The categories to split the document into")
+    subdocuments: list[Subdocument] = Field(..., description="The subdocuments to split the document into")
     model: str = Field(default="retab-small", description="The model to use to split the document")
+    context: str | None = Field(default=None, description="Additional context for the split operation (e.g., iteration context from a loop)")
 
 
 class Partition(BaseModel):
@@ -21,9 +21,9 @@ class Partition(BaseModel):
     last_page_y_end: float = Field(default=1.0, description="The y coordinate of the last page of the partition")
 
 class SplitResult(BaseModel):
-    name: str = Field(..., description="The name of the category")
-    pages: list[int] = Field(..., description="The pages of the category (1-indexed)")
-    partitions: list[Partition] = Field(default_factory=list, description="The partitions of the category")
+    name: str = Field(..., description="The name of the subdocument")
+    pages: list[int] = Field(..., description="The pages of the subdocument (1-indexed)")
+    partitions: list[Partition] = Field(default_factory=list, description="The partitions of the subdocument")
 
 
 class SplitResponse(BaseModel):
@@ -32,14 +32,14 @@ class SplitResponse(BaseModel):
 
 class SplitOutputItem(BaseModel):
     """Internal schema item for LLM structured output validation."""
-    name: str = Field(..., description="The name of the category")
-    start_page: int = Field(..., description="The start page of the category (1-indexed)")
-    end_page: int = Field(..., description="The end page of the category (1-indexed, inclusive)")
+    name: str = Field(..., description="The name of the subdocument")
+    start_page: int = Field(..., description="The start page of the subdocument (1-indexed)")
+    end_page: int = Field(..., description="The end page of the subdocument (1-indexed, inclusive)")
 
 
 class SplitOutputSchema(BaseModel):
     """Schema for LLM structured output."""
     splits: list[SplitOutputItem] = Field(
         ...,
-        description="List of document sections, each classified into one of the provided categories with their page ranges"
+        description="List of document sections, each classified into one of the provided subdocuments with their page ranges"
     )

{retab-0.0.90 → retab-0.0.92}/retab.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: retab
-Version: 0.0.90
+Version: 0.0.92
 Summary: Retab official python library
 Home-page: https://github.com/retab-dev/retab
 Author: Retab

{retab-0.0.90 → retab-0.0.92}/setup.py

@@ -6,7 +6,7 @@ with open("requirements.txt") as f:
 
 setup(
     name="retab",
-    version="0.0.90",
+    version="0.0.92",
     author="Retab",
     author_email="contact@retab.com",
     description="Retab official python library",