retab 0.0.90__py3-none-any.whl → 0.0.91__py3-none-any.whl

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/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/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/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.dist-info → retab-0.0.91.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: retab
-Version: 0.0.90
+Version: 0.0.91
 Summary: Retab official python library
 Home-page: https://github.com/retab-dev/retab
 Author: Retab

{retab-0.0.90.dist-info → retab-0.0.91.dist-info}/RECORD

@@ -7,7 +7,7 @@ retab/resources/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 retab/resources/models.py,sha256=4WidFBnTGZEA65DSn2pLP2SRnCVXkMTw7o_m8xVCFC4,2469
 retab/resources/schemas.py,sha256=rZ6OzfmoYv-mGaRVzvXjO09dD-KxP74mZhOO8sMgcDQ,4632
 retab/resources/documents/__init__.py,sha256=OjXmngFN0RKqO4SI-mJBNzr6Ex6rMxfq0DxaqzP0RQs,89
-retab/resources/documents/client.py,sha256=0ZOJojT4M9QZ53nheS_vuNZWcnmwTnKx3YqYyJ7_sGY,48912
+retab/resources/documents/client.py,sha256=GrVgOM37M3K62PdEXyBQNo1J4oyqXHhafR-hmHwe3Lw,49049
 retab/resources/edit/__init__.py,sha256=yycIstpTSKsz2qXbrY3Buzd35UDcPWvb5hw6Eb2rLow,69
 retab/resources/edit/client.py,sha256=DJKlwh8xui7IDRjwPmiGKTC1_HshXLYXX-xr93FhSbo,1270
 retab/resources/edit/agent/__init__.py,sha256=i5IdOMhwOOQmnhPFeBbh7-ChqwQh5q7oLow1zJ0ZAwM,74
@@ -32,14 +32,14 @@ retab/types/mime.py,sha256=ZLNCD3pvgn5cbGfJwzrdkjgB9dMHCbN67YEV9bx47zE,10063
 retab/types/modality.py,sha256=4B8LctdUBZVgIjtS2FjrJpljn2Eyse0XE1bpFsGb9O4,131
 retab/types/pagination.py,sha256=A0Fw06baPTfEaYwo3kvNs4vaupzlqylBc6tQH-2DFuY,279
 retab/types/standards.py,sha256=7aGtuvzzkKidvqY8JB2Cjfn43V80FeKwrTtp162kjKc,1477
-retab/types/documents/__init__.py,sha256=t1jXdpYqi-zQMC_9uM0m7eA1hRU0MCROwUx89ccD2-c,418
-retab/types/documents/classify.py,sha256=ZCEkuNESHqHIF4Iw5v-jCqt3d238QiqMNKPoUaKBWUg,1342
+retab/types/documents/__init__.py,sha256=7vRh1a5WRy1Ue0Tg69CsXeOltGmPawJ6GgAKpRy8huk,451
+retab/types/documents/classify.py,sha256=PBeF5bmCc3GUHxR8hcQ-49DBndO84WWFiuGRcAcpBBo,1626
 retab/types/documents/correct_orientation.py,sha256=e-ivsslI6L6Gl0YkcslXw_DH620xMGEYVp4tdeviXeM,261
 retab/types/documents/create_messages.py,sha256=Uym0SnVUGkyt1C5AOD37BsZ3puyeu_igR6X9SboojfA,7267
 retab/types/documents/edit.py,sha256=b6UcYLOJkClpMu4QyYmdp-X4WtN8U_3oiMBc1KLklVY,5663
 retab/types/documents/extract.py,sha256=x_59fm69-icsxxGRgpFd0NN-SLRoMYqbvfCZuG7zyGc,18033
 retab/types/documents/parse.py,sha256=MXe7zh3DusWQhGe0Sr95nPy6cB8DRX8MA4Hmjj_AP7E,1300
-retab/types/documents/split.py,sha256=Bjk5iJdS3v7I3rCvqpFUPlzgO4HINqh3uMPQJg-MqPc,2166
+retab/types/documents/split.py,sha256=z670Ppg7tiu_eBf4zPcsy5l57f_2T5yTYTzacquZpn4,2346
 retab/types/edit/__init__.py,sha256=M8hF97h7fX8RP9IsB6qpkw0eyvO0DFQvP6FmWL8caCQ,331
 retab/types/edit/templates.py,sha256=RLRIMdXzU-5_3XPf0iMSozjRTAP5Tliq0nrjlZn0l8E,2412
 retab/types/extractions/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -62,7 +62,7 @@ retab/utils/hashing.py,sha256=_BMVUvftOcJav68QL0rLkH2dbhW9RRJPzeGC2akR0fc,757
 retab/utils/json_schema.py,sha256=zP4pQLpVHBKWo_abCjb_dU4kA0azhHopd-1TFUgVEvc,20655
 retab/utils/mime.py,sha256=mTP_lqSPttOP5DYJxopiWaeFXrUCPjhwd7y53nCVGO4,6189
 retab/utils/stream_context_managers.py,sha256=gI1gVQSj3nWz6Mvjz7Ix5AiY0g6vSL-c2tPfuP04izo,2314
-retab-0.0.90.dist-info/METADATA,sha256=ea40JF37wk3xxuQRUv17mvliwsqiOHP8ufyna9ScpyY,4532
-retab-0.0.90.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-retab-0.0.90.dist-info/top_level.txt,sha256=waQR0EGdhLIQtztoE3AXg7ik5ONQ9q_bsKVpyFuJdq0,6
-retab-0.0.90.dist-info/RECORD,,
+retab-0.0.91.dist-info/METADATA,sha256=f3TPkuDOc9Wmpqjmx8zJpEySKkSq3Re8EXjB7ev92oE,4532
+retab-0.0.91.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+retab-0.0.91.dist-info/top_level.txt,sha256=waQR0EGdhLIQtztoE3AXg7ik5ONQ9q_bsKVpyFuJdq0,6
+retab-0.0.91.dist-info/RECORD,,