scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-09-21 00:21:49 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 74

remove input.py

Browse source
This commit is contained in:
Yujong Lee 2026-09-15 19:31:57 -07:00
parent 2f25875ce6
commit 6c893a7dba
4 changed files with 23 additions and 118 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

112
litellm/ocr/input.py
View file

@ -1,112 +0,0 @@
from collections.abc import Mapping
from os import PathLike
from typing import Final, Literal, Protocol, cast # noqa: TID251 # native callables are validated when loaded
from typing_extensions import NotRequired, ReadOnly, TypedDict
from litellm.rust_bridge.bindings import NativeBinding
from litellm.rust_bridge.configuration import rust_ocr_enabled
class FileReader(Protocol):
def read(self) -> bytes | str: ...
class FileDocument(TypedDict):
type: ReadOnly[Literal["file"]]
file: ReadOnly[bytes | PathLike[str] | FileReader]
mime_type: ReadOnly[NotRequired[str]]
class NativeFileDocument(Protocol):
def __call__(self, document: Mapping[str, object]) -> dict[str, str]: ...
class NativeUploadDocument(Protocol):
def __call__(self, file_content: bytes, file_name: str | None, content_type: str | None) -> dict[str, str]: ...
class NativeMimeType(Protocol):
def __call__(self, file_name: str) -> str: ...
_FILE_DOCUMENT: Final = NativeBinding(
"_ocr_file_document",
validate=lambda value: (
cast( # cast-ok: native export owns the callable signature
NativeFileDocument, value
)
if callable(value)
else None
),
)
_UPLOAD_DOCUMENT: Final = NativeBinding(
"_ocr_upload_document",
validate=lambda value: (
cast( # cast-ok: native export owns the callable signature
NativeUploadDocument, value
)
if callable(value)
else None
),
)
_MAX_FILE_BYTES: Final = NativeBinding(
"_OCR_MAX_FILE_BYTES", validate=lambda value: value if isinstance(value, int) and value > 0 else None
)
_MIME_TYPE: Final = NativeBinding(
"_ocr_mime_type",
validate=lambda value: (
cast( # cast-ok: native export owns the callable signature
NativeMimeType, value
)
if callable(value)
else None
),
)
_PYTHON_MAX_FILE_BYTES: Final = 50 * 1024 * 1024
def get_mime_type(file_path: str) -> str:
native: Final = _MIME_TYPE.load() if rust_ocr_enabled() else None
if native is None:
from litellm.ocr import legacy
return legacy.get_mime_type(file_path)
return native(file_path)
def get_max_file_bytes() -> int:
limit: Final = _MAX_FILE_BYTES.load() if rust_ocr_enabled() else None
if limit is None:
return _PYTHON_MAX_FILE_BYTES
return limit
def convert_file_document_to_url_document(document: FileDocument) -> dict[str, str]:
native: Final = _FILE_DOCUMENT.load() if rust_ocr_enabled() else None
if native is None:
from litellm.ocr import legacy
return legacy.convert_file_document_to_url_document(document)
return native(document)
def convert_upload_to_url_document(
file_content: bytes, filename: str | None, content_type: str | None
) -> dict[str, str]:
native: Final = _UPLOAD_DOCUMENT.load() if rust_ocr_enabled() else None
if native is None:
from litellm.ocr import legacy
if len(file_content) > _PYTHON_MAX_FILE_BYTES:
raise ValueError("OCR file exceeds the size limit")
content_mime: Final = content_type.split(";")[0].strip() if content_type else None
mime_type: Final = (
legacy.get_mime_type(filename)
if filename and (not content_mime or content_mime == "application/octet-stream")
else content_mime or "application/octet-stream"
)
return legacy.convert_file_document_to_url_document(
{"type": "file", "file": file_content, "mime_type": mime_type}
)
return native(file_content, filename, content_type)

7
litellm/ocr/legacy.py
View file

@ -11,7 +11,7 @@ from collections.abc import Coroutine, Mapping
from dataclasses import dataclass
from io import IOBase
from types import MappingProxyType
from typing import Final, cast # noqa: TID251 # adapters preserve the legacy untyped contracts
from typing import Final, Protocol, cast # noqa: TID251 # adapters preserve the legacy untyped contracts
import httpx
@ -26,7 +26,6 @@ from litellm.llms.base_llm.ocr.transformation import (
parse_ocr_request_format,
)
from litellm.llms.custom_httpx.llm_http_handler import BaseLLMHTTPHandler
from litellm.ocr.input import FileReader
from litellm.types.router import GenericLiteLLMParams
from litellm.types.utils import CustomPricingLiteLLMParams
from litellm.utils import ProviderConfigManager, client
@ -34,6 +33,10 @@ from litellm.utils import ProviderConfigManager, client
base_llm_http_handler: Final = BaseLLMHTTPHandler()
class FileReader(Protocol):
def read(self) -> bytes | str: ...
@dataclass(frozen=True, slots=True)
class _PreparedOCRRequest:
model: str

2
litellm/ocr/main.py
View file

@ -5,7 +5,7 @@ import httpx
from litellm.llms.base_llm.ocr.transformation import OCRResponse
from litellm.ocr import legacy
from litellm.ocr.input import convert_file_document_to_url_document, get_mime_type
from litellm.ocr.legacy import convert_file_document_to_url_document, get_mime_type
from litellm.rust_bridge.bindings import native_exception_types
from litellm.rust_bridge.configuration import rust_ocr_enabled
from litellm.rust_bridge.ocr import LiteLLMOcrRequest

20
litellm/proxy/ocr_endpoints/endpoints.py
View file

@ -15,12 +15,13 @@ from litellm.llms.base_llm.ocr.transformation import (
OCRResponse,
parse_ocr_request_format,
)
from litellm.ocr.input import convert_upload_to_url_document, get_max_file_bytes
from litellm.ocr.legacy import convert_file_document_to_url_document, get_mime_type
from litellm.proxy._types import *
from litellm.proxy.auth.user_api_key_auth import UserAPIKeyAuth, user_api_key_auth
from litellm.proxy.common_request_processing import ProxyBaseLLMRequestProcessing
router: Final = APIRouter()
_MAX_FILE_BYTES: Final = 50 * 1024 * 1024
def _build_document_from_upload(
@ -28,7 +29,18 @@ def _build_document_from_upload(
filename: str | None,
content_type: str | None,
) -> dict[str, str]:
return convert_upload_to_url_document(file_content, filename, content_type)
mime_type: Final = content_type.split(";")[0].strip() if content_type else None
if not mime_type or mime_type == "application/octet-stream":
if filename:
mime_type = get_mime_type(filename)
return convert_file_document_to_url_document(
{
"type": "file",
"file": file_content,
"mime_type": mime_type or "application/octet-stream",
}
)
def _with_request_format(data: Mapping[str, Any], request: Request) -> Mapping[str, Any]:
@ -103,9 +115,11 @@ async def _parse_multipart_form(request: Request) -> dict[str, Any]:
# Seek to start in case the file was already partially read by middleware
await uploaded_file.seek(0)
file_content: Final = await uploaded_file.read(get_max_file_bytes() + 1)
file_content: Final = await uploaded_file.read(_MAX_FILE_BYTES + 1)
if not file_content:
raise ValueError("Uploaded file is empty")
if len(file_content) > _MAX_FILE_BYTES:
raise ValueError("OCR file exceeds the size limit")
document: Final = _build_document_from_upload(
file_content=file_content,