chunkr-ai 0.0.50__py3-none-any.whl → 0.3.0__py3-none-any.whl

chunkr_ai/api/configuration.py

@@ -3,27 +3,34 @@ from enum import Enum
 from typing import Any, List, Optional, Union
 from pydantic import field_validator, field_serializer
 
-class GenerationStrategy(str, Enum):
-    LLM = "LLM"
-    AUTO = "Auto"
-
 class CroppingStrategy(str, Enum):
     ALL = "All"
     AUTO = "Auto"
 
-class EmbedSource(str, Enum):
-    HTML = "HTML"
+class SegmentFormat(str, Enum):
+    HTML = "Html"
     MARKDOWN = "Markdown"
-    LLM = "LLM"
+
+class EmbedSource(str, Enum):
     CONTENT = "Content"
+    HTML = "HTML"  # Deprecated
+    MARKDOWN = "Markdown"  # Deprecated
+    LLM = "LLM"
+
+class GenerationStrategy(str, Enum):
+    LLM = "LLM"
+    AUTO = "Auto"
 
 class GenerationConfig(BaseModel):
-    html: Optional[GenerationStrategy] = None
+    format: Optional[SegmentFormat] = None
+    strategy: Optional[GenerationStrategy] = None
     llm: Optional[str] = None
-    markdown: Optional[GenerationStrategy] = None
     crop_image: Optional[CroppingStrategy] = None
-    embed_sources: Optional[List[EmbedSource]] = Field(default_factory=lambda: [EmbedSource.MARKDOWN])
+    embed_sources: Optional[List[EmbedSource]] = None
     extended_context: Optional[bool] = None
+    # Deprecated fields for backwards compatibility
+    html: Optional[GenerationStrategy] = None  # Deprecated: Use format=SegmentFormat.HTML and strategy instead
+    markdown: Optional[GenerationStrategy] = None  # Deprecated: Use format=SegmentFormat.MARKDOWN and strategy instead
 
 class SegmentProcessing(BaseModel):
     model_config = ConfigDict(populate_by_name=True, alias_generator=str.title)
@@ -244,9 +251,45 @@ class SegmentType(str, Enum):
     TEXT = "Text"
     TITLE = "Title"
 
+class Alignment(str, Enum):
+    LEFT = "Left"
+    CENTER = "Center"
+    RIGHT = "Right"
+    JUSTIFY = "Justify"
+
+class VerticalAlignment(str, Enum):
+    TOP = "Top"
+    MIDDLE = "Middle"
+    BOTTOM = "Bottom"
+    BASELINE = "Baseline"
+
+class CellStyle(BaseModel):
+    bg_color: Optional[str] = None
+    text_color: Optional[str] = None
+    font_face: Optional[str] = None
+    is_bold: Optional[bool] = None
+    align: Optional[Alignment] = None
+    valign: Optional[VerticalAlignment] = None
+
+class Cell(BaseModel):
+    cell_id: str
+    text: str
+    range: str
+    formula: Optional[str] = None
+    value: Optional[str] = None
+    hyperlink: Optional[str] = None
+    style: Optional[CellStyle] = None
+
+class Page(BaseModel):
+    image: str
+    page_number: int
+    page_height: float
+    page_width: float
+    ss_sheet_name: Optional[str] = None
+
 class Segment(BaseModel):
     bbox: BoundingBox
-    content: str
+    content: str = ""
     page_height: float
     llm: Optional[str] = None
     html: Optional[str] = None
@@ -258,6 +301,16 @@ class Segment(BaseModel):
     segment_id: str
     segment_type: SegmentType
     confidence: Optional[float]
+    text: str = ""
+    segment_length: Optional[int] = None
+    # Spreadsheet-specific fields
+    ss_cells: Optional[List[Cell]] = None
+    ss_header_bbox: Optional[BoundingBox] = None
+    ss_header_ocr: Optional[List[OCRResult]] = None
+    ss_header_text: Optional[str] = None
+    ss_header_range: Optional[str] = None
+    ss_range: Optional[str] = None
+    ss_sheet_name: Optional[str] = None
 
 class Chunk(BaseModel):
     chunk_id: str
@@ -268,6 +321,8 @@ class Chunk(BaseModel):
 class OutputResponse(BaseModel):
     chunks: List[Chunk]
     file_name: Optional[str]
+    mime_type: Optional[str] = None
+    pages: Optional[List[Page]] = None
     page_count: Optional[int]
     pdf_url: Optional[str]
 

chunkr_ai/models.py

@@ -1,5 +1,8 @@
 from .api.configuration import (
+    Alignment,
     BoundingBox,
+    Cell,
+    CellStyle,
     Chunk,
     ChunkProcessing,
     Configuration,
@@ -14,7 +17,9 @@ from .api.configuration import (
     OCRResult,
     OcrStrategy,
     OutputResponse,
+    Page,
     Segment,
+    SegmentFormat,
     SegmentProcessing,
     SegmentType,
     SegmentationStrategy,
@@ -22,11 +27,15 @@ from .api.configuration import (
     Pipeline,
     Tokenizer,
     TokenizerType,
+    VerticalAlignment,
 )
 from .api.task_response import TaskResponse
 
 __all__ = [
+    "Alignment",
     "BoundingBox",
+    "Cell",
+    "CellStyle",
     "Chunk",
     "ChunkProcessing",
     "Configuration",
@@ -41,7 +50,9 @@ __all__ = [
     "OCRResult",
     "OcrStrategy",
     "OutputResponse",
+    "Page",
     "Segment",
+    "SegmentFormat",
     "SegmentProcessing",
     "SegmentType",
     "SegmentationStrategy",
@@ -50,4 +61,5 @@ __all__ = [
     "Pipeline",
     "Tokenizer",
     "TokenizerType",
+    "VerticalAlignment",
 ]

{chunkr_ai-0.0.50.dist-info → chunkr_ai-0.3.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: chunkr-ai
-Version: 0.0.50
+Version: 0.3.0
 Summary: Python client for Chunkr: open source document intelligence
 Author-email: Ishaan Kapoor <ishaan@lumina.sh>
 License: MIT License
@@ -28,6 +28,7 @@ Project-URL: Homepage, https://chunkr.ai
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: httpx>=0.25.0
+Requires-Dist: matplotlib>=3.10.3
 Requires-Dist: nest-asyncio>=1.6.0
 Requires-Dist: pillow>=10.0.0
 Requires-Dist: pydantic>=2.0.0

{chunkr_ai-0.0.50.dist-info → chunkr_ai-0.3.0.dist-info}/RECORD

@@ -1,16 +1,16 @@
 chunkr_ai/__init__.py,sha256=6KpYv2lmD6S5z2kc9pqwuLP5VDHmOuu2qDZArUIhb1s,53
-chunkr_ai/models.py,sha256=L0L9CjY8SgSh9_Fzvo_nJXqKf_2urZHngMWtBVlAQAo,1006
+chunkr_ai/models.py,sha256=NvFJOpsgzEyYHhE-flp7Yr9tpTDvFmF4T87jttFRquU,1202
 chunkr_ai/api/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 chunkr_ai/api/auth.py,sha256=0RSNFPvHt4Nrg8qtP2xvA2KbR0J_KUe1B_tKynbq9Fc,436
 chunkr_ai/api/chunkr.py,sha256=uSNYtB_mcs4-QRKsX7wZb8yv6ayXgRrJSDNZ-EbAyvc,3857
 chunkr_ai/api/chunkr_base.py,sha256=8roSPoCADmaXM2r7zz2iHfZzIcY9NopOfa4j-dfk8RA,6310
-chunkr_ai/api/configuration.py,sha256=M0oAXJmGBakOAKywg_Eyoz4GLginBrXuSvCtG0_BznQ,9926
+chunkr_ai/api/configuration.py,sha256=y_jd3K5GB-P8N3uym4wqHDVq-Rq-VT_bhqJqgKs0PVg,11586
 chunkr_ai/api/decorators.py,sha256=w1l_ZEkl99C-BO3qRTbi74sYwHDFspB1Bjt1Arv9lPc,4384
 chunkr_ai/api/misc.py,sha256=AaGLxZlMzNgVPwErskDRKc2UVGkC0JwxLXU-enPwzA0,5354
 chunkr_ai/api/protocol.py,sha256=LjPrYSq52m1afIlAo0yVGXlGZxPRh8J6g7S4PAit3Zo,388
 chunkr_ai/api/task_response.py,sha256=VYa62E08VlZUyjn2YslnY4cohdK9e53HbEzsaYIXKXM,8028
-chunkr_ai-0.0.50.dist-info/licenses/LICENSE,sha256=w3R12yNDyZpMiy2lxy_hvNbsldC75ww79sF0u11rkho,1069
-chunkr_ai-0.0.50.dist-info/METADATA,sha256=ZlM0XHhykHUOLEsNCF52OmMEw2QlvgSUy972FQjgD2s,7053
-chunkr_ai-0.0.50.dist-info/WHEEL,sha256=zaaOINJESkSfm_4HQVc5ssNzHCPXhJm0kEUakpsEHaU,91
-chunkr_ai-0.0.50.dist-info/top_level.txt,sha256=0IZY7PZIiS8bw5r4NUQRUQ-ATi-L_3vLQVq3ZLouOW8,10
-chunkr_ai-0.0.50.dist-info/RECORD,,
+chunkr_ai-0.3.0.dist-info/licenses/LICENSE,sha256=w3R12yNDyZpMiy2lxy_hvNbsldC75ww79sF0u11rkho,1069
+chunkr_ai-0.3.0.dist-info/METADATA,sha256=RCgp4cjj3CduWuM1ycLczz7iBHohtaEXLv8CZseephw,7086
+chunkr_ai-0.3.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+chunkr_ai-0.3.0.dist-info/top_level.txt,sha256=0IZY7PZIiS8bw5r4NUQRUQ-ATi-L_3vLQVq3ZLouOW8,10
+chunkr_ai-0.3.0.dist-info/RECORD,,

{chunkr_ai-0.0.50.dist-info → chunkr_ai-0.3.0.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (80.8.0)
+Generator: setuptools (80.9.0)
 Root-Is-Purelib: true
 Tag: py3-none-any