chunkr-ai 0.0.12__tar.gz → 0.0.14__tar.gz

{chunkr_ai-0.0.12/src/chunkr_ai.egg-info → chunkr_ai-0.0.14}/PKG-INFO

@@ -1,13 +1,12 @@
 Metadata-Version: 2.2
 Name: chunkr-ai
-Version: 0.0.12
+Version: 0.0.14
 Summary: Python client for Chunkr: open source document intelligence
 Author-email: Ishaan Kapoor <ishaan@lumina.sh>
 Project-URL: Homepage, https://chunkr.ai
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: httpx>=0.25.0
-Requires-Dist: httpx>=0.25.0
 Requires-Dist: pillow>=10.0.0
 Requires-Dist: pydantic>=2.0.0
 Requires-Dist: pytest-asyncio>=0.21.0
@@ -81,7 +80,7 @@ async def process_document():
     # If you want to upload without waiting for processing
     task = await chunkr.start_upload("document.pdf")
     # ... do other things ...
-    await task.poll_async()  # Check status when needed
+    await task.poll()  # Check status when needed
 ```
 
 ### Additional Features

{chunkr_ai-0.0.12 → chunkr_ai-0.0.14}/README.md

@@ -62,7 +62,7 @@ async def process_document():
     # If you want to upload without waiting for processing
     task = await chunkr.start_upload("document.pdf")
     # ... do other things ...
-    await task.poll_async()  # Check status when needed
+    await task.poll()  # Check status when needed
 ```
 
 ### Additional Features

{chunkr_ai-0.0.12 → chunkr_ai-0.0.14}/pyproject.toml

@@ -4,14 +4,13 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "chunkr-ai"
-version = "0.0.12"
+version = "0.0.14"
 authors = [{"name" = "Ishaan Kapoor", "email" = "ishaan@lumina.sh"}]
 description = "Python client for Chunkr: open source document intelligence"
 readme = "README.md"
 license = {"file" = "LICENSE"}
 urls = {Homepage = "https://chunkr.ai"} 
 dependencies = [
-    "httpx>=0.25.0",
     "httpx>=0.25.0",
     "pillow>=10.0.0",
     "pydantic>=2.0.0",

chunkr_ai-0.0.14/src/chunkr_ai/api/base.py

@@ -0,0 +1,173 @@
+from .config import Configuration
+from .task import TaskResponse
+from .auth import HeadersMixin
+from abc import abstractmethod
+from dotenv import load_dotenv
+import io
+import json
+import os
+from pathlib import Path
+from PIL import Image
+import requests
+from typing import BinaryIO, Tuple, Union
+
+class ChunkrBase(HeadersMixin):
+    """Base class with shared functionality for Chunkr API clients."""
+
+    def __init__(self, url: str = None, api_key: str = None):
+        load_dotenv()
+        self.url = (
+            url or 
+            os.getenv('CHUNKR_URL') or 
+            'https://api.chunkr.ai' 
+        )
+        self._api_key = (
+            api_key or 
+            os.getenv('CHUNKR_API_KEY')
+        )
+        if not self._api_key:
+            raise ValueError("API key must be provided either directly, in .env file, or as CHUNKR_API_KEY environment variable. You can get an api key at: https://www.chunkr.ai")
+            
+        self.url = self.url.rstrip("/")
+
+    def _prepare_file(
+        self,
+        file: Union[str, Path, BinaryIO, Image.Image]
+    ) -> 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://')):
+            response = requests.get(file)
+            response.raise_for_status()
+            file_obj = io.BytesIO(response.content)
+            filename = Path(file.split('/')[-1]).name or 'downloaded_file'
+            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)}")
+
+    def _prepare_upload_data(
+        self,
+        file: Union[str, Path, BinaryIO, Image.Image],
+        config: Configuration = None
+    ) -> Tuple[dict, dict]:
+        """Prepare files and data dictionaries for upload.
+        
+        Args:
+            file: The file to upload
+            config: Optional configuration settings
+            
+        Returns:
+            Tuple[dict, dict]: (files dict, data dict) ready for upload
+        """
+        filename, file_obj = self._prepare_file(file)
+        files = {"file": (filename, file_obj)}
+        data = {}
+        
+        if config:
+            config_dict = config.model_dump(mode="json", exclude_none=True)
+            for key, value in config_dict.items():
+                if isinstance(value, dict):
+                    files[key] = (None, json.dumps(value), 'application/json')
+                else:
+                    data[key] = value
+                    
+        return files, data
+    
+    @abstractmethod
+    def upload(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = None) -> TaskResponse:
+        """Upload a file and wait for processing to complete.
+        
+        Must be implemented by subclasses.
+        """
+        pass
+
+    @abstractmethod
+    def start_upload(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = None) -> TaskResponse:
+        """Upload a file for processing and immediately return the task response.
+        
+        Must be implemented by subclasses.
+        """
+        pass
+
+    @abstractmethod
+    def get_task(self, task_id: str) -> TaskResponse:
+        """Get a task response by its ID.
+        
+        Must be implemented by subclasses.
+        """
+        pass

chunkr_ai-0.0.14/src/chunkr_ai/api/chunkr.py

@@ -0,0 +1,77 @@
+from .chunkr_base import ChunkrBase
+from .config import Configuration
+from .task import TaskResponse
+from pathlib import Path
+from PIL import Image
+import requests
+from typing import Union, BinaryIO
+from .misc import prepare_upload_data
+
+class Chunkr(ChunkrBase):
+    """Chunkr API client"""
+
+    def __init__(self, url: str = None, api_key: str = None):
+        super().__init__(url, api_key)
+        self._session = requests.Session()
+
+    def upload(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = None) -> TaskResponse:
+        task = self.create_task(file, config)
+        return task.poll()
+    
+    def update(self, task_id: str, config: Configuration) -> TaskResponse:
+        task = self.update_task(task_id, config)
+        return task.poll()
+
+    def create_task(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = None) -> TaskResponse:
+        files= prepare_upload_data(file, config)
+        if not self._session:
+            raise ValueError("Session not found")
+        r = self._session.post(
+            f"{self.url}/api/v1/task",
+            files=files,
+            headers=self._headers()
+        )
+        r.raise_for_status()
+        return TaskResponse(**r.json()).with_client(self)
+    
+    def update_task(self, task_id: str, config: Configuration) -> TaskResponse:
+        files = prepare_upload_data(None, config)
+        if not self._session:
+            raise ValueError("Session not found")
+        r = self._session.patch(
+            f"{self.url}/api/v1/task/{task_id}",
+            files=files,
+            headers=self._headers()
+        )
+  
+        r.raise_for_status()
+        return TaskResponse(**r.json()).with_client(self)
+    
+    def get_task(self, task_id: str) -> TaskResponse:
+        if not self._session:
+            raise ValueError("Session not found")
+        r = self._session.get(
+            f"{self.url}/api/v1/task/{task_id}",
+            headers=self._headers()
+        )
+        r.raise_for_status()
+        return TaskResponse(**r.json()).with_client(self)
+
+
+    def delete_task(self, task_id: str) -> None:
+        if not self._session:
+            raise ValueError("Session not found")
+        r = self._session.delete(
+            f"{self.url}/api/v1/task/{task_id}",
+            headers=self._headers()
+        )
+        r.raise_for_status()
+
+    def cancel_task(self, task_id: str) -> None:
+        if not self._session:
+            raise ValueError("Session not found")
+        r = self._session.get(
+            f"{self.url}/api/v1/task/{task_id}/cancel",
+            headers=self._headers()
+        )
+        r.raise_for_status()

chunkr_ai-0.0.14/src/chunkr_ai/api/chunkr_async.py

@@ -0,0 +1,114 @@
+from .chunkr_base import ChunkrBase
+from .config import Configuration
+from .misc import prepare_upload_data
+from .task_async import TaskResponseAsync
+import httpx
+from pathlib import Path
+from PIL import Image
+from typing import Union, BinaryIO
+
+class ChunkrAsync(ChunkrBase):
+    """Asynchronous Chunkr API client"""
+    
+    def __init__(self, url: str = None, api_key: str = None):
+        super().__init__(url, api_key)
+        self._client = httpx.AsyncClient()
+
+    async def upload(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = None) -> TaskResponseAsync:
+        if not self._client or self._client.is_closed:
+            self._client = httpx.AsyncClient()
+        try:
+            task = await self.create_task(file, config)
+            return await task.poll()
+        except Exception as e:
+            await self._client.aclose()
+            raise e
+    
+    async def update(self, task_id: str, config: Configuration) -> TaskResponseAsync:
+        if not self._client or self._client.is_closed:
+            self._client = httpx.AsyncClient()
+        try:
+            task = await self.update_task(task_id, config)
+            return await task.poll()
+        except Exception as e:
+            await self._client.aclose()
+            raise e
+
+    async def create_task(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = None) -> TaskResponseAsync:
+        if not self._client or self._client.is_closed:
+            self._client = httpx.AsyncClient()
+        try:
+            files = prepare_upload_data(file, config)
+            r = await self._client.post(
+                f"{self.url}/api/v1/task",
+                files=files,
+                headers=self._headers()
+            )
+            r.raise_for_status()
+            return TaskResponseAsync(**r.json()).with_client(self)
+        except Exception as e:
+            await self._client.aclose()
+            raise e
+
+    async def update_task(self, task_id: str, config: Configuration) -> TaskResponseAsync:
+        if not self._client or self._client.is_closed:
+            self._client = httpx.AsyncClient()
+        try:
+            files = prepare_upload_data(None, config)
+            r = await self._client.patch(
+                f"{self.url}/api/v1/task/{task_id}",
+                files=files,
+                headers=self._headers()
+            )
+     
+            r.raise_for_status()
+            return TaskResponseAsync(**r.json()).with_client(self)
+        except Exception as e:
+            await self._client.aclose()
+            raise e
+    
+    async def get_task(self, task_id: str) -> TaskResponseAsync:
+        if not self._client or self._client.is_closed:
+            self._client = httpx.AsyncClient()
+        try:
+            r = await self._client.get(
+                f"{self.url}/api/v1/task/{task_id}",
+                headers=self._headers()
+            )
+            r.raise_for_status()
+            return TaskResponseAsync(**r.json()).with_client(self)
+        except Exception as e:
+            await self._client.aclose()
+            raise e
+    
+    async def delete_task(self, task_id: str) -> None:
+        if not self._client or self._client.is_closed:
+            self._client = httpx.AsyncClient()
+        try:
+            r = await self._client.delete(
+                f"{self.url}/api/v1/task/{task_id}",
+                headers=self._headers()
+            )
+            r.raise_for_status()
+        except Exception as e:
+            await self._client.aclose()
+            raise e
+    
+    async def cancel_task(self, task_id: str) -> None:
+        if not self._client or self._client.is_closed:
+            self._client = httpx.AsyncClient()
+        try:
+            r = await self._client.get(
+                f"{self.url}/api/v1/task/{task_id}/cancel",
+                headers=self._headers()
+            )
+            r.raise_for_status()
+        except Exception as e:
+            await self._client.aclose()
+            raise e
+
+    async def __aenter__(self):
+        return self
+
+    async def __aexit__(self, exc_type, exc_val, exc_tb):
+        await self._client.aclose()

chunkr_ai-0.0.12/src/chunkr_ai/api/chunkr.py → chunkr_ai-0.0.14/src/chunkr_ai/api/chunkr_base.py

@@ -1,20 +1,35 @@
-from .chunkr_base import ChunkrBase
 from .config import Configuration
 from .task import TaskResponse
+from .task_async import TaskResponseAsync
+from .auth import HeadersMixin
+from abc import abstractmethod
+from dotenv import load_dotenv
+import os
 from pathlib import Path
 from PIL import Image
-import requests
-from typing import Union, BinaryIO
-from .misc import prepare_upload_data
+from typing import BinaryIO, Union
 
-class Chunkr(ChunkrBase):
-    """Chunkr API client"""
+class ChunkrBase(HeadersMixin):
+    """Base class with shared functionality for Chunkr API clients."""
 
     def __init__(self, url: str = None, api_key: str = None):
-        super().__init__(url, api_key)
-        self._session = requests.Session()
+        load_dotenv()
+        self.url = (
+            url or 
+            os.getenv('CHUNKR_URL') or 
+            'https://api.chunkr.ai' 
+        )
+        self._api_key = (
+            api_key or 
+            os.getenv('CHUNKR_API_KEY')
+        )
+        if not self._api_key:
+            raise ValueError("API key must be provided either directly, in .env file, or as CHUNKR_API_KEY environment variable. You can get an api key at: https://www.chunkr.ai")
+            
+        self.url = self.url.rstrip("/")
 
-    def upload(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = None) -> TaskResponse:
+    @abstractmethod
+    def upload(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = None) -> Union[TaskResponse, TaskResponseAsync]:
         """Upload a file and wait for processing to complete.
 
         Args:
@@ -22,32 +37,32 @@ class Chunkr(ChunkrBase):
             config: Configuration options for processing. Optional.
 
         Examples:
-        ```
+        ```python
         # Upload from file path
-        chunkr.upload("document.pdf")
+        await chunkr.upload("document.pdf")
 
+        # Upload from opened file
+        with open("document.pdf", "rb") as f:
+            await chunkr.upload(f)
+        
         # Upload from URL
-        chunkr.upload("https://example.com/document.pdf")
+        await chunkr.upload("https://example.com/document.pdf")
 
         # Upload from base64 string (must include MIME type header)
-        chunkr.upload("data:application/pdf;base64,JVBERi0xLjcKCjEgMCBvYmo...")
-
-        # Upload from opened file
-        with open("document.pdf", "rb") as f:
-            chunkr.upload(f)
+        await chunkr.upload("data:application/pdf;base64,JVBERi0...")
 
         # Upload an image
         from PIL import Image
         img = Image.open("photo.jpg")
-        chunkr.upload(img)
+        await chunkr.upload(img)
         ```
         Returns:
             TaskResponse: The completed task response
         """
-        task = self.create_task(file, config)
-        return task.poll()
+        pass
     
-    def update(self, task_id: str, config: Configuration) -> TaskResponse:
+    @abstractmethod
+    def update(self, task_id: str, config: Configuration) -> Union[TaskResponse, TaskResponseAsync]:
         """Update a task by its ID and wait for processing to complete.
         
         Args:
@@ -57,11 +72,11 @@ class Chunkr(ChunkrBase):
         Returns:
             TaskResponse: The updated task response
         """
-        task = self.update_task(task_id, config)
-        return task.poll()
+        pass
 
-    def create_task(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = 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()`
+    @abstractmethod
+    def create_task(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = None) -> Union[TaskResponse, TaskResponseAsync]:
+        """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.
@@ -70,60 +85,44 @@ class Chunkr(ChunkrBase):
         Examples:
         ```
         # Upload from file path
-        task = chunkr.start_upload("document.pdf")
+        task = await chunkr.create_task("document.pdf")
 
         # Upload from opened file
         with open("document.pdf", "rb") as f:
-            task = chunkr.start_upload(f)
-
+            task = await chunkr.create_task(f)
+    
         # Upload from URL
-        task = chunkr.start_upload("https://example.com/document.pdf")
+        task = await chunkr.create_task("https://example.com/document.pdf")
 
         # Upload from base64 string (must include MIME type header)
-        task = chunkr.start_upload("data:application/pdf;base64,JVBERi0xLjcKCjEgMCBvYmo...")
+        task = await chunkr.create_task("data:application/pdf;base64,JVBERi0xLjcKCjEgMCBvYmo...")
 
         # Upload an image
         from PIL import Image
         img = Image.open("photo.jpg")
-        task = chunkr.start_upload(img)
+        task = await chunkr.create_task(img)
 
         # Wait for the task to complete - this can be done when needed
-        task.poll()
+        await task.poll()
         ```
-
-        Returns:
-            TaskResponse: The initial task response
         """
-        files= prepare_upload_data(file, config)
-        r = self._session.post(
-            f"{self.url}/api/v1/task",
-            files=files,
-            headers=self._headers()
-        )
-        r.raise_for_status()
-        return TaskResponse(**r.json()).with_client(self)
-    
-    def update_task(self, task_id: str, config: Configuration) -> TaskResponse:
-        """Update a task by its ID.
+        pass
+
+    @abstractmethod
+    def update_task(self, task_id: str, config: Configuration) -> Union[TaskResponse, TaskResponseAsync]:
+        """Update a task by its ID 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:
             task_id: The ID of the task to update
-            config: The new configuration to use
+            config: Configuration options for processing. Optional.
 
         Returns:
             TaskResponse: The updated task response
         """
-        files = prepare_upload_data(None, config)
-        r = self._session.patch(
-            f"{self.url}/api/v1/task/{task_id}",
-            files=files,
-            headers=self._headers()
-        )
-  
-        r.raise_for_status()
-        return TaskResponse(**r.json()).with_client(self)
+        pass
     
-    def get_task(self, task_id: str) -> TaskResponse:
+    @abstractmethod
+    def get_task(self, task_id: str) -> Union[TaskResponse, TaskResponseAsync]:
         """Get a task response by its ID.
         
         Args:
@@ -132,34 +131,23 @@ class Chunkr(ChunkrBase):
         Returns:
             TaskResponse: The task response
         """
-        r = self._session.get(
-            f"{self.url}/api/v1/task/{task_id}",
-            headers=self._headers()
-        )
-        r.raise_for_status()
-        return TaskResponse(**r.json()).with_client(self)
-
+        pass
 
+    @abstractmethod
     def delete_task(self, task_id: str) -> None:
         """Delete a task by its ID.
         
         Args:
             task_id: The ID of the task to delete
         """
-        r = self._session.delete(
-            f"{self.url}/api/v1/task/{task_id}",
-            headers=self._headers()
-        )
-        r.raise_for_status()
-
+        pass
+    
+    @abstractmethod
     def cancel_task(self, task_id: str) -> None:
         """Cancel a task by its ID.
         
         Args:
             task_id: The ID of the task to cancel
         """
-        r = self._session.get(
-            f"{self.url}/api/v1/task/{task_id}/cancel",
-            headers=self._headers()
-        )
-        r.raise_for_status()
+        pass
+

{chunkr_ai-0.0.12 → chunkr_ai-0.0.14}/src/chunkr_ai/api/protocol.py

@@ -1,14 +1,14 @@
-from typing import runtime_checkable, Protocol
+from typing import Optional, runtime_checkable, Protocol
 from requests import Session
-from httpx import AsyncClient
+from aiohttp import ClientSession
 
 @runtime_checkable
 class ChunkrClientProtocol(Protocol):
     """Protocol defining the interface for Chunkr clients"""
     url: str
     _api_key: str
-    _session: Session
-    _client: AsyncClient
+    _session: Optional[Session] = None
+    _client: Optional[ClientSession] = None
 
     def get_api_key(self) -> str:
         """Get the API key"""

{chunkr_ai-0.0.12 → chunkr_ai-0.0.14}/src/chunkr_ai/api/task.py

@@ -7,6 +7,10 @@ class TaskResponse(TaskBase):
     def _poll_request(self) -> dict:
         while True:
             try:
+                if not self.task_url:
+                   raise ValueError("Task URL not found in response")
+                if not self._client._session:
+                    raise ValueError("Client session not found")
                 r = self._client._session.get(self.task_url, headers=self._client._headers())
                 r.raise_for_status()
                 return r.json()
@@ -17,10 +21,8 @@ class TaskResponse(TaskBase):
                 raise
 
     def poll(self) -> 'TaskResponse':
-        if not self.task_url:
-            raise ValueError("Task URL not found in response")
         while True:
-            response = self._poll_request_sync()
+            response = self._poll_request()
             updated_task = TaskResponse(**response).with_client(self._client)
             self.__dict__.update(updated_task.__dict__)
             if result := self._check_status():
@@ -30,9 +32,11 @@ class TaskResponse(TaskBase):
     def update(self, config: Configuration) -> 'TaskResponse':
         if not self.task_url:
             raise ValueError("Task URL not found")
+        if not self._client._session:
+            raise ValueError("Client session not found")
         files = prepare_upload_data(None, config)
         r = self._client._session.patch(
-            f"{self.task_url}",
+            self.task_url,
             files=files,
             headers=self._client._headers()
         )
@@ -44,6 +48,8 @@ class TaskResponse(TaskBase):
     def cancel(self):
         if not self.task_url:
             raise ValueError("Task URL not found")
+        if not self._client._session:
+            raise ValueError("Client session not found")
         r = self._client._session.get(
             f"{self.task_url}/cancel",
             headers=self._client._headers()
@@ -54,6 +60,8 @@ class TaskResponse(TaskBase):
     def delete(self):
         if not self.task_url:
             raise ValueError("Task URL not found")
+        if not self._client._session:
+            raise ValueError("Client session not found")
         r = self._client._session.delete(
             self.task_url,
             headers=self._client._headers()

{chunkr_ai-0.0.12 → chunkr_ai-0.0.14}/src/chunkr_ai/api/task_async.py

@@ -6,6 +6,8 @@ import asyncio
 class TaskResponseAsync(TaskBase):
     async def _poll_request(self) -> dict:
         try:
+            if not self._client._client:
+                raise ValueError("Client not found")
             r = await self._client._client.get(self.task_url, headers=self._client._headers())
             r.raise_for_status()
             return r.json()
@@ -18,6 +20,8 @@ class TaskResponseAsync(TaskBase):
     async def poll(self) -> 'TaskResponseAsync':
         if not self.task_url:
             raise ValueError("Task URL not found")
+        if not self._client._client:
+            raise ValueError("Client not found")
         while True:
             j = await self._poll_request()
             updated = TaskResponseAsync(**j).with_client(self._client)
@@ -29,6 +33,8 @@ class TaskResponseAsync(TaskBase):
     async def update(self, config: Configuration) -> 'TaskResponseAsync':
         if not self.task_url:
             raise ValueError("Task URL not found")
+        if not self._client._client:
+            raise ValueError("Client not found")
         f = prepare_upload_data(None, config)
         r = await self._client._client.patch(self.task_url, files=f, headers=self._client._headers())
         r.raise_for_status()
@@ -39,6 +45,8 @@ class TaskResponseAsync(TaskBase):
     async def cancel(self):
         if not self.task_url:
             raise ValueError("Task URL not found")
+        if not self._client._client:
+            raise ValueError("Client not found")
         r = await self._client._client.get(f"{self.task_url}/cancel", headers=self._client._headers())
         r.raise_for_status()
         return await self.poll()
@@ -46,5 +54,7 @@ class TaskResponseAsync(TaskBase):
     async def delete(self):
         if not self.task_url:
             raise ValueError("Task URL not found")
+        if not self._client._client:
+            raise ValueError("Client not found")
         r = await self._client._client.delete(self.task_url, headers=self._client._headers())
-        r.raise_for_status()
+        r.raise_for_status()

{chunkr_ai-0.0.12 → chunkr_ai-0.0.14}/src/chunkr_ai/api/task_base.py

@@ -1,8 +1,7 @@
-from .config import Configuration
+from .config import Configuration, Status, OutputResponse
 from .protocol import ChunkrClientProtocol
-from ..models import Status, OutputResponse
 from abc import ABC, abstractmethod
-from typing import TypeVar, Optional, Generic, Union
+from typing import TypeVar, Optional, Generic
 from pydantic import BaseModel, PrivateAttr
 from datetime import datetime
 
@@ -23,7 +22,7 @@ class TaskBase(BaseModel, ABC, Generic[T]):
     status: Status
     task_id: str
     task_url: Optional[str]
-    _client: Optional[Union[ChunkrClientProtocol]] = PrivateAttr(default=None)
+    _client: Optional[ChunkrClientProtocol] = PrivateAttr(default=None)
 
     @abstractmethod
     def _poll_request(self) -> dict:
@@ -50,7 +49,7 @@ class TaskBase(BaseModel, ABC, Generic[T]):
         """Delete the task."""
         pass
 
-    def with_client(self, client: Union[ChunkrClientProtocol]) -> T:
+    def with_client(self, client: ChunkrClientProtocol) -> T:
         self._client = client
         return self
 

{chunkr_ai-0.0.12 → chunkr_ai-0.0.14/src/chunkr_ai.egg-info}/PKG-INFO

@@ -1,13 +1,12 @@
 Metadata-Version: 2.2
 Name: chunkr-ai
-Version: 0.0.12
+Version: 0.0.14
 Summary: Python client for Chunkr: open source document intelligence
 Author-email: Ishaan Kapoor <ishaan@lumina.sh>
 Project-URL: Homepage, https://chunkr.ai
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: httpx>=0.25.0
-Requires-Dist: httpx>=0.25.0
 Requires-Dist: pillow>=10.0.0
 Requires-Dist: pydantic>=2.0.0
 Requires-Dist: pytest-asyncio>=0.21.0
@@ -81,7 +80,7 @@ async def process_document():
     # If you want to upload without waiting for processing
     task = await chunkr.start_upload("document.pdf")
     # ... do other things ...
-    await task.poll_async()  # Check status when needed
+    await task.poll()  # Check status when needed
 ```
 
 ### Additional Features

{chunkr_ai-0.0.12 → chunkr_ai-0.0.14}/src/chunkr_ai.egg-info/SOURCES.txt

@@ -9,7 +9,9 @@ src/chunkr_ai.egg-info/dependency_links.txt
 src/chunkr_ai.egg-info/requires.txt
 src/chunkr_ai.egg-info/top_level.txt
 src/chunkr_ai/api/__init__.py
+src/chunkr_ai/api/api.py
 src/chunkr_ai/api/auth.py
+src/chunkr_ai/api/base.py
 src/chunkr_ai/api/chunkr.py
 src/chunkr_ai/api/chunkr_async.py
 src/chunkr_ai/api/chunkr_base.py

{chunkr_ai-0.0.12 → chunkr_ai-0.0.14}/src/chunkr_ai.egg-info/requires.txt

@@ -1,5 +1,4 @@
 httpx>=0.25.0
-httpx>=0.25.0
 pillow>=10.0.0
 pydantic>=2.0.0
 pytest-asyncio>=0.21.0

{chunkr_ai-0.0.12 → chunkr_ai-0.0.14}/tests/test_chunkr.py

@@ -38,7 +38,7 @@ async def test_send_file_path(chunkr_client, sample_path):
     client_type, client = chunkr_client
     response = await client.upload(sample_path) if client_type == "async" else client.upload(sample_path)
     
-    assert isinstance(response, TaskResponse)
+    
     assert response.task_id is not None
     assert response.status == "Succeeded"
     assert response.output is not None
@@ -48,7 +48,7 @@ async def test_send_file_path_str(chunkr_client, sample_path):
     client_type, client = chunkr_client
     response = await client.upload(str(sample_path)) if client_type == "async" else client.upload(str(sample_path))
     
-    assert isinstance(response, TaskResponse)
+    
     assert response.task_id is not None
     assert response.status == "Succeeded"
     assert response.output is not None
@@ -59,7 +59,7 @@ async def test_send_opened_file(chunkr_client, sample_path):
     with open(sample_path, 'rb') as f:
         response = await client.upload(f) if client_type == "async" else client.upload(f)
     
-    assert isinstance(response, TaskResponse)
+    
     assert response.task_id is not None
     assert response.status == "Succeeded"
     assert response.output is not None
@@ -69,7 +69,6 @@ async def test_send_pil_image(chunkr_client, sample_image):
     client_type, client = chunkr_client
     response = await client.upload(sample_image) if client_type == "async" else client.upload(sample_image)
     
-    assert isinstance(response, TaskResponse)
     assert response.task_id is not None
     assert response.status == "Succeeded"
 
@@ -82,7 +81,6 @@ async def test_ocr_auto(chunkr_client, sample_path):
         ocr_strategy=OcrStrategy.AUTO
     ))
     
-    assert isinstance(response, TaskResponse)
     assert response.task_id is not None
     assert response.status == "Succeeded"
     assert response.output is not None
@@ -96,7 +94,7 @@ async def test_expires_in(chunkr_client, sample_path):
         expires_in=10
     ))
     
-    assert isinstance(response, TaskResponse)
+    
     assert response.task_id is not None
     assert response.status == "Succeeded"
     assert response.output is not None
@@ -114,7 +112,7 @@ async def test_chunk_processing(chunkr_client, sample_path):
         )
     ))
     
-    assert isinstance(response, TaskResponse)
+    
     assert response.task_id is not None
     assert response.status == "Succeeded"
     assert response.output is not None
@@ -128,7 +126,6 @@ async def test_segmentation_strategy_page(chunkr_client, sample_path):
         segmentation_strategy=SegmentationStrategy.PAGE
     ))
     
-    assert isinstance(response, TaskResponse)
     assert response.task_id is not None
     assert response.status == "Succeeded"
     assert response.output is not None
@@ -152,7 +149,7 @@ async def test_page_llm_html(chunkr_client, sample_path):
         )
     ))
     
-    assert isinstance(response, TaskResponse)
+    
     assert response.task_id is not None
     assert response.status == "Succeeded"
     assert response.output is not None
@@ -160,15 +157,7 @@ async def test_page_llm_html(chunkr_client, sample_path):
 @pytest.mark.asyncio
 async def test_page_llm(chunkr_client, sample_path):
     client_type, client = chunkr_client
-    response = await client.upload(sample_path, Configuration(
-        segmentation_strategy=SegmentationStrategy.PAGE,
-        segment_processing=SegmentProcessing(
-            page=GenerationConfig(
-                html=GenerationStrategy.LLM,
-                markdown=GenerationStrategy.LLM
-            )
-        )
-    )) if client_type == "async" else client.upload(sample_path, Configuration(
+    configuration = Configuration(
         segmentation_strategy=SegmentationStrategy.PAGE,
         segment_processing=SegmentProcessing(
             page=GenerationConfig(
@@ -176,9 +165,10 @@ async def test_page_llm(chunkr_client, sample_path):
                 markdown=GenerationStrategy.LLM
             )
         )
-    ))
+    )
+    
+    response = await client.upload(sample_path, configuration) if client_type == "async" else client.upload(sample_path, configuration)
     
-    assert isinstance(response, TaskResponse)
     assert response.task_id is not None
     assert response.status == "Succeeded"
     assert response.output is not None
@@ -204,16 +194,16 @@ async def test_json_schema(chunkr_client, sample_path):
         )
     ))
     
-    assert isinstance(response, TaskResponse)
     assert response.task_id is not None
     if response.status != "Succeeded":
         raise ValueError(f"Task failed with message: {response.message}")
     assert response.output is not None
+    
 @pytest.mark.asyncio
 async def test_delete_task(chunkr_client, sample_path):
     client_type, client = chunkr_client
     response = await client.upload(sample_path) if client_type == "async" else client.upload(sample_path)
-    assert isinstance(response, TaskResponse)
+    
     assert response.task_id is not None
     assert response.status == "Succeeded"
     assert response.output is not None
@@ -249,14 +239,14 @@ async def test_delete_task_direct(chunkr_client, sample_path):
 async def test_cancel_task(chunkr_client, sample_path):
     client_type, client = chunkr_client
     response = await client.create_task(sample_path) if client_type == "async" else client.create_task(sample_path)
-    assert isinstance(response, TaskResponse)
+    
     assert response.task_id is not None
     assert response.status == "Starting"
     
     if client_type == "async":
         await client.cancel_task(response.task_id)
         assert (await client.get_task(response.task_id)).status == "Cancelled"
-        await response.poll_async()
+        await response.poll()
     else:
         client.cancel_task(response.task_id)
         assert client.get_task(response.task_id).status == "Cancelled"
@@ -290,7 +280,7 @@ async def test_update_task(chunkr_client, sample_path):
         segmentation_strategy=SegmentationStrategy.PAGE,
     )
     response = await client.upload(sample_path, original_config) if client_type == "async" else client.upload(sample_path, original_config)
-    assert isinstance(response, TaskResponse)
+    
     assert response.task_id is not None
     assert response.status == "Succeeded"
     assert response.output is not None

chunkr_ai-0.0.12/src/chunkr_ai/api/chunkr_async.py

@@ -1,144 +0,0 @@
-from .chunkr_base import ChunkrBase
-from .task import TaskResponse
-from .config import Configuration
-import httpx
-from pathlib import Path
-from PIL import Image
-from typing import Union, BinaryIO
-from .misc import prepare_upload_data
-
-class ChunkrAsync(ChunkrBase):
-    """Asynchronous Chunkr API client"""
-    
-    def __init__(self, url: str = None, api_key: str = None):
-        super().__init__(url, api_key)
-        self._client = httpx.AsyncClient()
-
-    async def upload(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = None) -> TaskResponse:
-        """Upload a file and wait for processing to complete.
-
-        Args:
-            file: The file to upload. 
-            config: Configuration options for processing. Optional.
-
-        Examples:
-        ```python
-        # Upload from file path
-        await chunkr.upload("document.pdf")
-
-        # Upload from opened file
-        with open("document.pdf", "rb") as f:
-            await chunkr.upload(f)
-        
-        # Upload from URL
-        await chunkr.upload("https://example.com/document.pdf")
-
-        # Upload from base64 string (must include MIME type header)
-        await chunkr.upload("data:application/pdf;base64,JVBERi0xLjcKCjEgMCBvYmo...")
-
-        # Upload an image
-        from PIL import Image
-        img = Image.open("photo.jpg")
-        await chunkr.upload(img)
-        ```
-        Returns:
-            TaskResponse: The completed task response
-        """
-        task = await self.create_task(file, config)
-        return await task.poll_async()
-    
-    async def update(self, task_id: str, config: Configuration) -> TaskResponse:
-        """Update a task by its ID and wait for processing to complete.
-        
-        Args:
-            task_id: The ID of the task to update
-            config: Configuration options for processing. Optional.
-
-        Returns:
-            TaskResponse: The updated task response
-        """
-        task = await self.update_task(task_id, config)
-        return await task.poll_async()
-
-    async def create_task(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = 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_async()`.
-
-        Args:
-            file: The file to upload.
-            config: Configuration options for processing. Optional.
-
-        Examples:
-        ```
-        # Upload from file path
-        task = await chunkr.start_upload("document.pdf")
-
-        # Upload from opened file
-        with open("document.pdf", "rb") as f:
-            task = await chunkr.start_upload(f)
-    
-        # Upload from URL
-        task = await chunkr.start_upload("https://example.com/document.pdf")
-
-        # Upload from base64 string (must include MIME type header)
-        task = await chunkr.start_upload("data:application/pdf;base64,JVBERi0xLjcKCjEgMCBvYmo...")
-
-        # Upload an image
-        from PIL import Image
-        img = Image.open("photo.jpg")
-        task = await chunkr.start_upload(img)
-
-        # Wait for the task to complete - this can be done when needed
-        await task.poll_async()
-        ```
-
-        Returns:
-            TaskResponse: The initial task response
-        """
-        files = prepare_upload_data(file, config)
-        r = await self._client.post(
-            f"{self.url}/api/v1/task",
-            files=files,
-            headers=self._headers()
-        )
-        r.raise_for_status()
-        return TaskResponse(**r.json()).with_client(self)
-
-    async def update_task(self, task_id: str, config: Configuration) -> TaskResponse:
-        files = prepare_upload_data(None, config)
-        r = await self._client.patch(
-            f"{self.url}/api/v1/task/{task_id}",
-            files=files,
-            headers=self._headers()
-        )
-     
-        r.raise_for_status()
-        return TaskResponse(**r.json()).with_client(self)
-    
-    async def get_task(self, task_id: str) -> TaskResponse:
-        r = await self._client.get(
-            f"{self.url}/api/v1/task/{task_id}",
-            headers=self._headers()
-        )
-        r.raise_for_status()
-        return TaskResponse(**r.json()).with_client(self)
-    
-    async def delete_task(self, task_id: str) -> None:
-        r = await self._client.delete(
-            f"{self.url}/api/v1/task/{task_id}",
-            headers=self._headers()
-        )
-        r.raise_for_status()
-    
-    async def cancel_task(self, task_id: str) -> None:
-        r = await self._client.get(
-            f"{self.url}/api/v1/task/{task_id}/cancel",
-            headers=self._headers()
-        )
-        r.raise_for_status()
-
-    
-    async def __aenter__(self):
-        return self
-
-    async def __aexit__(self, exc_type, exc_val, exc_tb):
-        await self._client.aclose()

chunkr_ai-0.0.12/src/chunkr_ai/api/chunkr_base.py

@@ -1,85 +0,0 @@
-from .config import Configuration
-from .task import TaskResponse
-from .auth import HeadersMixin
-from abc import abstractmethod
-from dotenv import load_dotenv
-import os
-from pathlib import Path
-from PIL import Image
-from typing import BinaryIO, Union
-
-class ChunkrBase(HeadersMixin):
-    """Base class with shared functionality for Chunkr API clients."""
-
-    def __init__(self, url: str = None, api_key: str = None):
-        load_dotenv()
-        self.url = (
-            url or 
-            os.getenv('CHUNKR_URL') or 
-            'https://api.chunkr.ai' 
-        )
-        self._api_key = (
-            api_key or 
-            os.getenv('CHUNKR_API_KEY')
-        )
-        if not self._api_key:
-            raise ValueError("API key must be provided either directly, in .env file, or as CHUNKR_API_KEY environment variable. You can get an api key at: https://www.chunkr.ai")
-            
-        self.url = self.url.rstrip("/")
-
-    @abstractmethod
-    def upload(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = None) -> TaskResponse:
-        """Upload a file and wait for processing to complete.
-        
-        Must be implemented by subclasses.
-        """
-        pass
-    
-    @abstractmethod
-    def update_task(self, task_id: str, config: Configuration) -> TaskResponse:
-        """Update a task by its ID.
-        
-        Must be implemented by subclasses.
-        """
-        pass
-
-    @abstractmethod
-    def create_task(self, file: Union[str, Path, BinaryIO, Image.Image], config: Configuration = None) -> TaskResponse:
-        """Upload a file for processing and immediately return the task response.
-        
-        Must be implemented by subclasses.
-        """
-        pass
-
-    @abstractmethod
-    def update_task(self, task_id: str, config: Configuration) -> TaskResponse:
-        """Update a task by its ID.
-        
-        Must be implemented by subclasses.
-        """
-        pass
-    
-    @abstractmethod
-    def get_task(self, task_id: str) -> TaskResponse:
-        """Get a task response by its ID.
-        
-        Must be implemented by subclasses.
-        """
-        pass
-
-    @abstractmethod
-    def delete_task(self, task_id: str) -> None:
-        """Delete a task by its ID.
-        
-        Must be implemented by subclasses.
-        """
-        pass
-    
-    @abstractmethod
-    def cancel_task(self, task_id: str) -> None:
-        """Cancel a task by its ID.
-        
-        Must be implemented by subclasses.
-        """
-        pass
-