chunkr-ai 0.0.22__py3-none-any.whl → 0.0.24__py3-none-any.whl

chunkr_ai/api/chunkr.py

@@ -34,7 +34,7 @@ class Chunkr(ChunkrBase):
         file: Union[str, Path, BinaryIO, Image.Image],
         config: Configuration = None,
     ) -> TaskResponse:
-        files = prepare_upload_data(file, config)
+        files = await prepare_upload_data(file, config, self._client)
         r = await self._client.post(
             f"{self.url}/api/v1/task", files=files, headers=self._headers()
         )
@@ -44,7 +44,7 @@ class Chunkr(ChunkrBase):
     @anywhere()
     @ensure_client()
     async def update_task(self, task_id: str, config: Configuration) -> TaskResponse:
-        files = prepare_upload_data(None, config)
+        files = await prepare_upload_data(None, config, self._client)
         r = await self._client.patch(
             f"{self.url}/api/v1/task/{task_id}",
             files=files,

chunkr_ai/api/chunkr_base.py

@@ -4,22 +4,27 @@ from .auth import HeadersMixin
 from abc import abstractmethod
 from dotenv import load_dotenv
 import httpx
-import io
-import json
 import os
 from pathlib import Path
 from PIL import Image
-import requests
-from typing import BinaryIO, Tuple, Union
+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):
+    """Base class with shared functionality for Chunkr API clients.
+    
+    Args:
+        url: The base URL of the Chunkr API. Defaults to the value of the CHUNKR_URL environment variable, or "https://api.chunkr.ai" if not set.
+        api_key: The API key to use for authentication. Defaults to the value of the CHUNKR_API_KEY environment variable, or None if not set.
+        raise_on_failure: Whether to raise an exception if the task fails. Defaults to False.
+    """
+
+    def __init__(self, url: str = None, api_key: str = None, raise_on_failure: bool = False):
         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")
+        self.raise_on_failure = raise_on_failure
+        
         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"
@@ -28,130 +33,6 @@ class ChunkrBase(HeadersMixin):
         self.url = self.url.rstrip("/")
         self._client = httpx.AsyncClient()
 
-    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,

chunkr_ai/api/misc.py

@@ -3,16 +3,36 @@ import io
 import json
 from pathlib import Path
 from PIL import Image
-import requests
+import httpx
 from typing import Union, Tuple, BinaryIO, Optional
 
-def prepare_file(file: Union[str, Path, BinaryIO, Image.Image]) -> Tuple[str, BinaryIO]:
-    """Convert various file types into a tuple of (filename, file-like object)."""
+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://")
     ):
-        response = requests.get(file)
+        if not client:
+            raise ValueError("Client must be provided to download files from URLs")
+        response = client.get(file)
         response.raise_for_status()
 
         # Try to get filename from Content-Disposition header first
@@ -108,9 +128,10 @@ def prepare_file(file: Union[str, Path, BinaryIO, Image.Image]) -> Tuple[str, Bi
     raise TypeError(f"Unsupported file type: {type(file)}")
 
 
-def prepare_upload_data(
+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.
 
@@ -123,7 +144,7 @@ def prepare_upload_data(
     """
     files = {}
     if file:
-        filename, file_obj = prepare_file(file)
+        filename, file_obj = await prepare_file(file, client)
         files = {"file": (filename, file_obj)}
 
     if config:

chunkr_ai/api/protocol.py

@@ -5,15 +5,10 @@ from httpx import AsyncClient
 @runtime_checkable
 class ChunkrClientProtocol(Protocol):
     """Protocol defining the interface for Chunkr clients"""
-
-    url: str
-    _api_key: str
+    
+    raise_on_failure: bool = True
     _client: Optional[AsyncClient] = None
 
-    def get_api_key(self) -> str:
-        """Get the API key"""
-        ...
-
     def _headers(self) -> dict:
         """Return headers required for API requests"""
         ...

chunkr_ai/api/task_response.py

@@ -30,15 +30,16 @@ class TaskResponse(BaseModel, Generic[T]):
     def _check_status(self) -> Optional[T]:
         """Helper method to check task status and handle completion/failure"""
         if self.status == "Failed":
-            raise ValueError(self.message)
+            if getattr(self._client, 'raise_on_failure', True):
+                raise ValueError(self.message)
+            return self
         if self.status not in ("Starting", "Processing"):
             return self
         return None
 
+    @require_task()
     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()
             )
@@ -51,7 +52,6 @@ class TaskResponse(BaseModel, Generic[T]):
             raise
 
     @anywhere()
-    @require_task()
     async def poll(self) -> T:
         """Poll the task for completion."""
         while True:
@@ -66,7 +66,7 @@ class TaskResponse(BaseModel, Generic[T]):
     @require_task()
     async def update(self, config: Configuration) -> T:
         """Update the task configuration."""
-        f = prepare_upload_data(None, config)
+        f = await prepare_upload_data(None, config, self._client._client)
         r = await self._client._client.patch(
             self.task_url, files=f, headers=self._client._headers()
         )

{chunkr_ai-0.0.22.dist-info → chunkr_ai-0.0.24.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: chunkr-ai
-Version: 0.0.22
+Version: 0.0.24
 Summary: Python client for Chunkr: open source document intelligence
 Author-email: Ishaan Kapoor <ishaan@lumina.sh>
 License: MIT License
@@ -35,6 +35,7 @@ Provides-Extra: test
 Requires-Dist: pytest>=7.0.0; extra == "test"
 Requires-Dist: pytest-xdist>=3.0.0; extra == "test"
 Requires-Dist: pytest-asyncio>=0.21.0; extra == "test"
+Requires-Dist: ruff>=0.9.3; extra == "test"
 
 # Chunkr Python Client
 

chunkr_ai-0.0.24.dist-info/RECORD

@@ -0,0 +1,17 @@
+chunkr_ai/__init__.py,sha256=6KpYv2lmD6S5z2kc9pqwuLP5VDHmOuu2qDZArUIhb1s,53
+chunkr_ai/models.py,sha256=MK8FPbWDj1ynvSHaYuslKCPybxLuAlrsVIM3Eym3kKI,750
+chunkr_ai/api/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+chunkr_ai/api/api.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+chunkr_ai/api/auth.py,sha256=hlv0GiUmlsbFO1wLL9sslqOnsBSoBqkL_6Mk2SDvxgE,413
+chunkr_ai/api/chunkr.py,sha256=XTXJFs0xjYY3w3N4fSQcxtJFBtNfzFYYkh6nDlFz4cY,2714
+chunkr_ai/api/chunkr_base.py,sha256=4SXA-gdZd1w2zZeeMdy4xog0NKOrKjmo6IMvSl9KSBg,5538
+chunkr_ai/api/config.py,sha256=NmPTsDvcjkvNx0gNzDTz-oFG5rQC7jm-H70O_crJCw8,4478
+chunkr_ai/api/decorators.py,sha256=y_Z9z0O2XXiX9z6jWDwdbCPdQyMLnjE0pGkJjHQEv_Q,2652
+chunkr_ai/api/misc.py,sha256=5Q2K713VPwf3S2519KTzjT9PKhTEBgBMk1d8NNnmpZ0,5717
+chunkr_ai/api/protocol.py,sha256=LjPrYSq52m1afIlAo0yVGXlGZxPRh8J6g7S4PAit3Zo,388
+chunkr_ai/api/task_response.py,sha256=hcHsBgX-2C5Px5Bu0IKk33K_AkqHSEM1Wu2zkcPh9to,3935
+chunkr_ai-0.0.24.dist-info/LICENSE,sha256=w3R12yNDyZpMiy2lxy_hvNbsldC75ww79sF0u11rkho,1069
+chunkr_ai-0.0.24.dist-info/METADATA,sha256=JyDI8EkFaJQQ7vIo2osHxXmeuNqhQ0UWjgUMHSFIYow,6996
+chunkr_ai-0.0.24.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+chunkr_ai-0.0.24.dist-info/top_level.txt,sha256=0IZY7PZIiS8bw5r4NUQRUQ-ATi-L_3vLQVq3ZLouOW8,10
+chunkr_ai-0.0.24.dist-info/RECORD,,

chunkr_ai-0.0.22.dist-info/RECORD

@@ -1,17 +0,0 @@
-chunkr_ai/__init__.py,sha256=6KpYv2lmD6S5z2kc9pqwuLP5VDHmOuu2qDZArUIhb1s,53
-chunkr_ai/models.py,sha256=MK8FPbWDj1ynvSHaYuslKCPybxLuAlrsVIM3Eym3kKI,750
-chunkr_ai/api/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-chunkr_ai/api/api.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-chunkr_ai/api/auth.py,sha256=hlv0GiUmlsbFO1wLL9sslqOnsBSoBqkL_6Mk2SDvxgE,413
-chunkr_ai/api/chunkr.py,sha256=ZjDVUeR1KPiLLHbj7AJipt-prBTw7a_GvjJ0nd7JV7c,2674
-chunkr_ai/api/chunkr_base.py,sha256=TDqEwCCfgshggi_Mzv76FhPj5z21QP8EVj7siczvfao,9826
-chunkr_ai/api/config.py,sha256=NmPTsDvcjkvNx0gNzDTz-oFG5rQC7jm-H70O_crJCw8,4478
-chunkr_ai/api/decorators.py,sha256=y_Z9z0O2XXiX9z6jWDwdbCPdQyMLnjE0pGkJjHQEv_Q,2652
-chunkr_ai/api/misc.py,sha256=wUG4SpfEEo7NcVK47gmw42dRy9zT5F9S2DtVC4T4ERs,4877
-chunkr_ai/api/protocol.py,sha256=Nt8aWr4ouVwCvoLqVI5vnXJhT2cvxt0sQC-svUk2G5w,458
-chunkr_ai/api/task_response.py,sha256=yjlUOADqf0O9X-yF7lbZuv39ttvitZHQpbGy3DtrJ80,3909
-chunkr_ai-0.0.22.dist-info/LICENSE,sha256=w3R12yNDyZpMiy2lxy_hvNbsldC75ww79sF0u11rkho,1069
-chunkr_ai-0.0.22.dist-info/METADATA,sha256=ro8wcqBnNGLkIhh7XrbSxdl4zaWXhA0d5SUgBQteAec,6952
-chunkr_ai-0.0.22.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
-chunkr_ai-0.0.22.dist-info/top_level.txt,sha256=0IZY7PZIiS8bw5r4NUQRUQ-ATi-L_3vLQVq3ZLouOW8,10
-chunkr_ai-0.0.22.dist-info/RECORD,,