chunkr-ai 0.0.36__tar.gz → 0.0.38__tar.gz

{chunkr_ai-0.0.36/src/chunkr_ai.egg-info → chunkr_ai-0.0.38}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: chunkr-ai
-Version: 0.0.36
+Version: 0.0.38
 Summary: Python client for Chunkr: open source document intelligence
 Author-email: Ishaan Kapoor <ishaan@lumina.sh>
 License: MIT License

{chunkr_ai-0.0.36 → chunkr_ai-0.0.38}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "chunkr-ai"
-version = "0.0.36"
+version = "0.0.38"
 authors = [{"name" = "Ishaan Kapoor", "email" = "ishaan@lumina.sh"}]
 description = "Python client for Chunkr: open source document intelligence"
 readme = "README.md"

{chunkr_ai-0.0.36 → chunkr_ai-0.0.38}/src/chunkr_ai/api/chunkr.py

@@ -1,6 +1,6 @@
 from pathlib import Path
 from PIL import Image
-from typing import Union, BinaryIO
+from typing import Union, BinaryIO, Optional
 
 from .configuration import Configuration
 from .decorators import anywhere, ensure_client, retry_on_429
@@ -17,8 +17,9 @@ class Chunkr(ChunkrBase):
         self,
         file: Union[str, Path, BinaryIO, Image.Image],
         config: Configuration = None,
+        filename: Optional[str] = None,
     ) -> TaskResponse:
-        task = await self.create_task(file, config)
+        task = await self.create_task(file, config, filename)
         return await task.poll()
 
     @anywhere()
@@ -34,10 +35,12 @@ class Chunkr(ChunkrBase):
         self,
         file: Union[str, Path, BinaryIO, Image.Image],
         config: Configuration = None,
+        filename: Optional[str] = None,
     ) -> TaskResponse:
-        files = await prepare_upload_data(file, config, self._client)
+        """Create a new task with the given file and configuration."""
+        data = await prepare_upload_data(file, filename, config)
         r = await self._client.post(
-            f"{self.url}/api/v1/task", files=files, headers=self._headers()
+            f"{self.url}/api/v1/task/parse", json=data, headers=self._headers()
         )
         r.raise_for_status()
         return TaskResponse(**r.json()).with_client(self, True, False)
@@ -46,10 +49,11 @@ class Chunkr(ChunkrBase):
     @ensure_client()
     @retry_on_429()
     async def update_task(self, task_id: str, config: Configuration) -> TaskResponse:
-        files = await prepare_upload_data(None, config, self._client)
+        """Update an existing task with new configuration."""
+        data = await prepare_upload_data(None, None, config)
         r = await self._client.patch(
-            f"{self.url}/api/v1/task/{task_id}",
-            files=files,
+            f"{self.url}/api/v1/task/{task_id}/parse",
+            json=data,
             headers=self._headers(),
         )
         r.raise_for_status()

{chunkr_ai-0.0.36 → chunkr_ai-0.0.38}/src/chunkr_ai/api/chunkr_base.py

@@ -7,8 +7,7 @@ import httpx
 import os
 from pathlib import Path
 from PIL import Image
-from typing import BinaryIO, Union
-
+from typing import BinaryIO, Union, Optional
 
 class ChunkrBase(HeadersMixin):
     """Base class with shared functionality for Chunkr API clients.
@@ -20,7 +19,7 @@ class ChunkrBase(HeadersMixin):
     """
 
     def __init__(self, url: str = None, api_key: str = None, raise_on_failure: bool = False):
-        load_dotenv()
+        load_dotenv(override=True)
         self.url = url or os.getenv("CHUNKR_URL") or "https://api.chunkr.ai"
         self._api_key = api_key or os.getenv("CHUNKR_API_KEY")
         self.raise_on_failure = raise_on_failure
@@ -38,13 +37,15 @@ class ChunkrBase(HeadersMixin):
         self,
         file: Union[str, Path, BinaryIO, Image.Image],
         config: Configuration = None,
+        filename: Optional[str] = None,
     ) -> TaskResponse:
         """Upload a file and wait for processing to complete.
 
         Args:
             file: The file to upload.
             config: Configuration options for processing. Optional.
-
+            filename: The filename to use for the file. Optional.
+            
         Examples:
         ```python
         # Upload from file path
@@ -58,7 +59,7 @@ class ChunkrBase(HeadersMixin):
         await chunkr.upload("https://example.com/document.pdf")
 
         # Upload from base64 string (must include MIME type header)
-        await chunkr.upload("data:application/pdf;base64,JVBERi0...")
+        await chunkr.upload("data:application/pdf;base64,JVBERi0...", filename="document.pdf")
 
         # Upload an image
         from PIL import Image
@@ -90,13 +91,14 @@ class ChunkrBase(HeadersMixin):
         self,
         file: Union[str, Path, BinaryIO, Image.Image],
         config: Configuration = None,
+        filename: Optional[str] = None,
     ) -> TaskResponse:
         """Upload a file for processing and immediately return the task response. It will not wait for processing to complete. To wait for the full processing to complete, use `task.poll()`.
 
         Args:
             file: The file to upload.
             config: Configuration options for processing. Optional.
-
+            filename: The filename to use for the file. Optional.
         Examples:
         ```
         # Upload from file path
@@ -110,7 +112,7 @@ class ChunkrBase(HeadersMixin):
         task = await chunkr.create_task("https://example.com/document.pdf")
 
         # Upload from base64 string (must include MIME type header)
-        task = await chunkr.create_task("data:application/pdf;base64,JVBERi0xLjcKCjEgMCBvYmo...")
+        task = await chunkr.create_task("data:application/pdf;base64,JVBERi0xLjcKCjEgMCBvYmo...", filename="document.pdf")
 
         # Upload an image
         from PIL import Image

{chunkr_ai-0.0.36 → chunkr_ai-0.0.38}/src/chunkr_ai/api/configuration.py

@@ -19,20 +19,18 @@ class GenerationConfig(BaseModel):
 class SegmentProcessing(BaseModel):
     model_config = ConfigDict(populate_by_name=True, alias_generator=str.title)
 
-    title: Optional[GenerationConfig] = Field(default=None, alias="Title")
-    section_header: Optional[GenerationConfig] = Field(
-        default=None, alias="SectionHeader"
-    )
-    text: Optional[GenerationConfig] = Field(default=None, alias="Text")
-    list_item: Optional[GenerationConfig] = Field(default=None, alias="ListItem")
-    table: Optional[GenerationConfig] = Field(default=None, alias="Table")
-    picture: Optional[GenerationConfig] = Field(default=None, alias="Picture")
     caption: Optional[GenerationConfig] = Field(default=None, alias="Caption")
-    formula: Optional[GenerationConfig] = Field(default=None, alias="Formula")
     footnote: Optional[GenerationConfig] = Field(default=None, alias="Footnote")
-    page_header: Optional[GenerationConfig] = Field(default=None, alias="PageHeader")
-    page_footer: Optional[GenerationConfig] = Field(default=None, alias="PageFooter")
+    formula: Optional[GenerationConfig] = Field(default=None, alias="Formula")
+    list_item: Optional[GenerationConfig] = Field(default=None, alias="ListItem")
     page: Optional[GenerationConfig] = Field(default=None, alias="Page")
+    page_footer: Optional[GenerationConfig] = Field(default=None, alias="PageFooter")
+    page_header: Optional[GenerationConfig] = Field(default=None, alias="PageHeader")
+    picture: Optional[GenerationConfig] = Field(default=None, alias="Picture")
+    section_header: Optional[GenerationConfig] = Field(default=None, alias="SectionHeader")
+    table: Optional[GenerationConfig] = Field(default=None, alias="Table")
+    text: Optional[GenerationConfig] = Field(default=None, alias="Text")
+    title: Optional[GenerationConfig] = Field(default=None, alias="Title")
 
 class ChunkProcessing(BaseModel):
     ignore_headers_and_footers: Optional[bool] = None
@@ -84,11 +82,13 @@ class Segment(BaseModel):
     page_width: float
     segment_id: str
     segment_type: SegmentType
+    confidence: Optional[float]
 
 class Chunk(BaseModel):
     chunk_id: str
     chunk_length: int
     segments: List[Segment]
+    embed: Optional[str] = None
 
 class OutputResponse(BaseModel):
     chunks: List[Chunk]

chunkr_ai-0.0.38/src/chunkr_ai/api/misc.py

@@ -0,0 +1,103 @@
+from .configuration import Configuration
+import base64
+import io
+from pathlib import Path
+from PIL import Image
+from typing import Union, Tuple, BinaryIO, Optional
+
+async def prepare_file(file: Union[str, Path, BinaryIO, Image.Image]) -> Tuple[Optional[str], str]:
+    """Convert various file types into a tuple of (filename, file content).
+
+    Args:
+        file: Input file, can be:
+            - URL string starting with http:// or https://
+            - Base64 string
+            - Local file path (will be converted to base64)
+            - Opened binary file (will be converted to base64)
+            - PIL/Pillow Image object (will be converted to base64)
+
+    Returns:
+        Tuple[Optional[str], str]: (filename, content) where content is either a URL or base64 string
+        The filename may be None for URLs, base64 strings, and PIL Images
+
+    Raises:
+        FileNotFoundError: If the file path doesn't exist
+        TypeError: If the file type is not supported
+        ValueError: If the URL is invalid or unreachable
+        ValueError: If the MIME type is unsupported
+    """
+    # Handle strings
+    if isinstance(file, str):
+        if file.startswith(('http://', 'https://')):
+            return None, file
+        try:
+            base64.b64decode(file)
+            return None, file
+        except:
+            try:
+                file = Path(file)
+            except:
+                raise ValueError("File must be a valid path, URL, or base64 string")
+
+    # Handle file paths - convert to base64
+    if isinstance(file, Path):
+        path = Path(file).resolve()
+        if not path.exists():
+            raise FileNotFoundError(f"File not found: {file}")
+        
+        with open(path, "rb") as f:
+            file_content = f.read()
+            file_ext = path.suffix.lower().lstrip('.')
+            if not file_ext:
+                raise ValueError("File must have an extension")
+            base64_str = base64.b64encode(file_content).decode()
+            return path.name, base64_str
+
+    # Handle PIL Images - convert to base64
+    if isinstance(file, Image.Image):
+        img_byte_arr = io.BytesIO()
+        format = file.format or "PNG"
+        file.save(img_byte_arr, format=format)
+        img_byte_arr.seek(0)
+        base64_str = base64.b64encode(img_byte_arr.getvalue()).decode()
+        return None, base64_str
+
+    # Handle file-like objects - convert to base64
+    if hasattr(file, "read") and hasattr(file, "seek"):
+        file.seek(0)
+        file_content = file.read()
+        name = getattr(file, "name", "document")
+        file_ext = Path(name).suffix.lower().lstrip('.')
+        if not file_ext:
+            raise ValueError("File must have an extension")
+        base64_str = base64.b64encode(file_content).decode()
+        return Path(name).name, base64_str
+
+    raise TypeError(f"Unsupported file type: {type(file)}")
+
+
+async def prepare_upload_data(
+    file: Optional[Union[str, Path, BinaryIO, Image.Image]] = None,
+    filename: Optional[str] = None,
+    config: Optional[Configuration] = None,
+) -> dict:
+    """Prepare data dictionary for upload.
+
+    Args:
+        file: The file to upload
+        config: Optional configuration settings
+        client: HTTP client for downloading remote files
+
+    Returns:
+        dict: JSON-serializable data dictionary ready for upload
+    """
+    data = {}
+    if file:
+        processed_filename, processed_file = await prepare_file(file)
+        data["file"] = processed_file
+        data["file_name"] = filename or processed_filename
+
+    if config:
+        data.update(config.model_dump(mode="json", exclude_none=True))
+        
+    return data

{chunkr_ai-0.0.36 → chunkr_ai-0.0.38}/src/chunkr_ai/api/task_response.py

@@ -74,9 +74,11 @@ class TaskResponse(BaseModel, Generic[T]):
     @retry_on_429()
     async def update(self, config: Configuration) -> T:
         """Update the task configuration."""
-        f = await prepare_upload_data(None, config, self._client._client)
+        data = await prepare_upload_data(None, None, config)
         r = await self._client._client.patch(
-            self.task_url, files=f, headers=self._client._headers()
+            f"{self.task_url}/parse",
+            json=data,
+            headers=self._client._headers()
         )
         r.raise_for_status()
         updated = TaskResponse(**r.json()).with_client(self._client)
@@ -142,7 +144,7 @@ class TaskResponse(BaseModel, Generic[T]):
         Args:
             output_file (str, optional): Path to save the markdown content. Defaults to None.
         """
-        content = self._get_content("markdown")
+        content = self._get_content("markdown", separator="\n\n")
         self._write_to_file(content, output_file)
         return content
 
@@ -166,7 +168,7 @@ class TaskResponse(BaseModel, Generic[T]):
         self._write_to_file(data, output_file, is_json=True)
         return data
 
-    def _get_content(self, t: str) -> str:
+    def _get_content(self, t: str, separator: str = "\n") -> str:
         if not self.output:
             return ""
         parts = []
@@ -175,4 +177,4 @@ class TaskResponse(BaseModel, Generic[T]):
                 v = getattr(s, t)
                 if v:
                     parts.append(v)
-        return "\n".join(parts)
+        return separator.join(parts)

{chunkr_ai-0.0.36 → chunkr_ai-0.0.38/src/chunkr_ai.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: chunkr-ai
-Version: 0.0.36
+Version: 0.0.38
 Summary: Python client for Chunkr: open source document intelligence
 Author-email: Ishaan Kapoor <ishaan@lumina.sh>
 License: MIT License

{chunkr_ai-0.0.36 → chunkr_ai-0.0.38}/tests/test_chunkr.py

@@ -2,6 +2,7 @@ import pytest
 from pathlib import Path
 from PIL import Image
 import asyncio
+import base64
 
 from chunkr_ai import Chunkr
 from chunkr_ai.models import (
@@ -68,6 +69,7 @@ async def test_send_pil_image(client, sample_image):
     assert response.task_id is not None
     assert response.status == "Succeeded"
     assert response.output is not None
+    assert response.output is not None
 
 @pytest.mark.asyncio
 async def test_ocr_auto(client, sample_path):
@@ -222,15 +224,41 @@ async def test_task_operations_after_client_close(client, sample_path):
     result = await task.poll()
     assert result.status == "Succeeded"
 
+@pytest.mark.asyncio
+async def test_send_base64_file(client, sample_path):
+    # Read file and convert to base64
+    with open(sample_path, "rb") as f:
+        base64_content = base64.b64encode(f.read()).decode('utf-8')
+    response = await client.upload(base64_content)
+    assert response.task_id is not None
+    assert response.status == "Succeeded"
+    assert response.output is not None
+
+@pytest.mark.asyncio
+async def test_send_base64_file_with_filename(client, sample_path):
+    # Read file and convert to base64
+    with open(sample_path, "rb") as f:
+        base64_content = base64.b64encode(f.read()).decode('utf-8')
+    
+    response = await client.upload(base64_content, filename="test.pdf")
+    assert response.task_id is not None
+    assert response.status == "Succeeded"
+    assert response.output is not None
+    
 @pytest.mark.asyncio
 async def test_output_files_no_dir(client, sample_path, tmp_path):
-    await client.upload(sample_path)
+    task = await client.upload(sample_path)
     
     html_file = tmp_path / "output.html"
     md_file = tmp_path / "output.md"
     content_file = tmp_path / "output.txt"
     json_file = tmp_path / "output.json"
     
+    task.html(html_file)
+    task.markdown(md_file)
+    task.content(content_file)
+    task.json(json_file)
+    
     assert html_file.exists()
     assert md_file.exists()
     assert content_file.exists()
@@ -238,13 +266,18 @@ async def test_output_files_no_dir(client, sample_path, tmp_path):
 
 @pytest.mark.asyncio
 async def test_output_files_with_dirs(client, sample_path, tmp_path):
-    await client.upload(sample_path)
+    task = await client.upload(sample_path)
     
     nested_dir = tmp_path / "nested" / "output" / "dir"
     html_file = nested_dir / "output.html"
     md_file = nested_dir / "output.md"
     content_file = nested_dir / "output.txt"
     json_file = nested_dir / "output.json"
+    
+    task.html(html_file)
+    task.markdown(md_file)
+    task.content(content_file)
+    task.json(json_file)
 
     assert html_file.exists()
     assert md_file.exists()

chunkr_ai-0.0.36/src/chunkr_ai/api/misc.py

@@ -1,155 +0,0 @@
-from .configuration import Configuration
-import io
-import json
-from pathlib import Path
-from PIL import Image
-import httpx
-from typing import Union, Tuple, BinaryIO, Optional
-
-async def prepare_file(file: Union[str, Path, BinaryIO, Image.Image], client: httpx.AsyncClient = None) -> Tuple[str, BinaryIO]:
-    """Convert various file types into a tuple of (filename, file-like object).
-
-        Args:
-            file: Input file, can be:
-                - String or Path to a file
-                - URL string starting with http:// or https://
-                - Base64 string
-                - Opened binary file (mode='rb')
-                - PIL/Pillow Image object
-
-        Returns:
-            Tuple[str, BinaryIO]: (filename, file-like object) ready for upload
-
-        Raises:
-            FileNotFoundError: If the file path doesn't exist
-            TypeError: If the file type is not supported
-            ValueError: If the URL is invalid or unreachable
-            ValueError: If the MIME type is unsupported
-    """  
-    # Handle URLs
-    if isinstance(file, str) and (
-        file.startswith("http://") or file.startswith("https://")
-    ):
-        if not client:
-            raise ValueError("Client must be provided to download files from URLs")
-        response = await client.get(file)
-        response.raise_for_status()
-
-        # Try to get filename from Content-Disposition header first
-        filename = None
-        content_disposition = response.headers.get("Content-Disposition")
-        if content_disposition and "filename=" in content_disposition:
-            filename = content_disposition.split("filename=")[-1].strip("\"'")
-
-        # If no Content-Disposition, try to get clean filename from URL path
-        if not filename:
-            from urllib.parse import urlparse, unquote
-
-            parsed_url = urlparse(file)
-            path = unquote(parsed_url.path)
-            filename = Path(path).name if path else None
-
-        # Fallback to default name if we couldn't extract one
-        filename = filename or "downloaded_file"
-
-        # Sanitize filename: remove invalid characters and limit length
-        import re
-
-        filename = re.sub(
-            r'[<>:"/\\|?*%]', "_", filename
-        )  # Replace invalid chars with underscore
-        filename = re.sub(r"\s+", "_", filename)  # Replace whitespace with underscore
-        filename = filename.strip("._")  # Remove leading/trailing dots and underscores
-        filename = filename[:255]  # Limit length to 255 characters
-
-        file_obj = io.BytesIO(response.content)
-        return filename, file_obj
-
-    # Handle base64 strings
-    if isinstance(file, str) and "," in file and ";base64," in file:
-        try:
-            # Split header and data
-            header, base64_data = file.split(",", 1)
-            import base64
-
-            file_bytes = base64.b64decode(base64_data)
-            file_obj = io.BytesIO(file_bytes)
-
-            # Try to determine format from header
-            format = "bin"
-            mime_type = header.split(":")[-1].split(";")[0].lower()
-
-            # Map MIME types to file extensions
-            mime_to_ext = {
-                "application/pdf": "pdf",
-                "application/vnd.openxmlformats-officedocument.wordprocessingml.document": "docx",
-                "application/msword": "doc",
-                "application/vnd.openxmlformats-officedocument.presentationml.presentation": "pptx",
-                "application/vnd.ms-powerpoint": "ppt",
-                "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet": "xlsx",
-                "application/vnd.ms-excel": "xls",
-                "image/jpeg": "jpg",
-                "image/png": "png",
-                "image/jpg": "jpg",
-            }
-
-            if mime_type in mime_to_ext:
-                format = mime_to_ext[mime_type]
-            else:
-                raise ValueError(f"Unsupported MIME type: {mime_type}")
-
-            return f"file.{format}", file_obj
-        except Exception as e:
-            raise ValueError(f"Invalid base64 string: {str(e)}")
-
-    # Handle file paths
-    if isinstance(file, (str, Path)):
-        path = Path(file).resolve()
-        if not path.exists():
-            raise FileNotFoundError(f"File not found: {file}")
-        return path.name, open(path, "rb")
-
-    # Handle PIL Images
-    if isinstance(file, Image.Image):
-        img_byte_arr = io.BytesIO()
-        format = file.format or "PNG"
-        file.save(img_byte_arr, format=format)
-        img_byte_arr.seek(0)
-        return f"image.{format.lower()}", img_byte_arr
-
-    # Handle file-like objects
-    if hasattr(file, "read") and hasattr(file, "seek"):
-        # Try to get the filename from the file object if possible
-        name = (
-            getattr(file, "name", "document") if hasattr(file, "name") else "document"
-        )
-        return Path(name).name, file
-
-    raise TypeError(f"Unsupported file type: {type(file)}")
-
-
-async def prepare_upload_data(
-    file: Optional[Union[str, Path, BinaryIO, Image.Image]] = None,
-    config: Optional[Configuration] = None,
-    client: httpx.AsyncClient = None,
-) -> dict:
-    """Prepare files and data dictionaries for upload.
-
-    Args:
-        file: The file to upload
-        config: Optional configuration settings
-
-    Returns:
-        dict: (files dict) ready for upload
-    """
-    files = {}
-    if file:
-        filename, file_obj = await prepare_file(file, client)
-        files = {"file": (filename, file_obj)}
-
-    if config:
-        config_dict = config.model_dump(mode="json", exclude_none=True)
-        for key, value in config_dict.items():
-            files[key] = (None, json.dumps(value), "application/json")
-
-    return files