Update app.py
Browse files
app.py
CHANGED
|
@@ -9,14 +9,12 @@ from typing import Optional
|
|
| 9 |
from pydantic import BaseModel
|
| 10 |
from rapidocr_onnxruntime import RapidOCR
|
| 11 |
|
| 12 |
-
# Khởi tạo FastAPI và OCR engine
|
| 13 |
app = FastAPI(title="OCR API", description="RapidOCR API Service")
|
| 14 |
engine = RapidOCR()
|
| 15 |
|
| 16 |
def process_ocr(image: np.ndarray, use_det: bool, use_cls: bool, use_rec: bool) -> list:
|
| 17 |
"""Xử lý OCR và trả về kết quả"""
|
| 18 |
img_bgr = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
|
| 19 |
-
print(use_det, use_cls, use_rec)
|
| 20 |
result, _ = engine(img_bgr, use_det=use_det, use_cls=use_cls, use_rec=use_rec)
|
| 21 |
|
| 22 |
if not result:
|
|
@@ -41,22 +39,6 @@ def process_ocr(image: np.ndarray, use_det: bool, use_cls: bool, use_rec: bool)
|
|
| 41 |
|
| 42 |
return texts
|
| 43 |
|
| 44 |
-
@app.get("/")
|
| 45 |
-
async def root():
|
| 46 |
-
"""Root endpoint - hướng dẫn sử dụng API"""
|
| 47 |
-
return {
|
| 48 |
-
"message": "OCR API is running",
|
| 49 |
-
"endpoints": {
|
| 50 |
-
"POST /ocr": "Upload image file for OCR",
|
| 51 |
-
"POST /ocr/url": "Provide image URL for OCR",
|
| 52 |
-
"GET /health": "Health check"
|
| 53 |
-
},
|
| 54 |
-
"usage": {
|
| 55 |
-
"curl_example": 'curl -X POST https://your-space.hf.space/ocr -F "file=@image.jpg"',
|
| 56 |
-
"python_example": "requests.post('https://your-space.hf.space/ocr', files={'file': open('image.jpg', 'rb')})"
|
| 57 |
-
}
|
| 58 |
-
}
|
| 59 |
-
|
| 60 |
@app.post("/ocr")
|
| 61 |
async def ocr_from_file(
|
| 62 |
file: UploadFile = File(..., description="Image file to process"),
|
|
@@ -64,14 +46,10 @@ async def ocr_from_file(
|
|
| 64 |
use_cls: bool = Form(True, description="Use classification"),
|
| 65 |
use_rec: bool = Form(True, description="Use recognition")
|
| 66 |
):
|
| 67 |
-
"""OCR from uploaded image file"""
|
| 68 |
-
|
| 69 |
-
# Validate file type
|
| 70 |
if not file.content_type or not file.content_type.startswith('image/'):
|
| 71 |
raise HTTPException(status_code=400, detail="File must be an image")
|
| 72 |
|
| 73 |
try:
|
| 74 |
-
# Read and process image
|
| 75 |
contents = await file.read()
|
| 76 |
image = Image.open(io.BytesIO(contents))
|
| 77 |
|
|
@@ -91,52 +69,11 @@ async def ocr_from_file(
|
|
| 91 |
except Exception as e:
|
| 92 |
raise HTTPException(status_code=500, detail=f"Error: {str(e)}")
|
| 93 |
|
| 94 |
-
class OCRUrlRequest(BaseModel):
|
| 95 |
-
url: str
|
| 96 |
-
use_det: Optional[bool] = True
|
| 97 |
-
use_cls: Optional[bool] = True
|
| 98 |
-
use_rec: Optional[bool] = True
|
| 99 |
-
|
| 100 |
-
@app.post("/ocr/url")
|
| 101 |
-
async def ocr_from_url(request: OCRUrlRequest):
|
| 102 |
-
"""OCR from image URL"""
|
| 103 |
-
|
| 104 |
-
try:
|
| 105 |
-
# Download image from URL
|
| 106 |
-
response = requests.get(request.url, timeout=30)
|
| 107 |
-
response.raise_for_status()
|
| 108 |
-
|
| 109 |
-
image = Image.open(io.BytesIO(response.content))
|
| 110 |
-
|
| 111 |
-
if image.mode != 'RGB':
|
| 112 |
-
image = image.convert('RGB')
|
| 113 |
-
|
| 114 |
-
img_np = np.array(image)
|
| 115 |
-
results = process_ocr(
|
| 116 |
-
img_np,
|
| 117 |
-
request.use_det,
|
| 118 |
-
request.use_cls,
|
| 119 |
-
request.use_rec
|
| 120 |
-
)
|
| 121 |
-
|
| 122 |
-
return JSONResponse(content={
|
| 123 |
-
"success": True,
|
| 124 |
-
"texts": [item["text"] for item in results],
|
| 125 |
-
"details": results,
|
| 126 |
-
"num_texts": len(results)
|
| 127 |
-
})
|
| 128 |
-
|
| 129 |
-
except requests.RequestException as e:
|
| 130 |
-
raise HTTPException(status_code=400, detail=f"Failed to fetch image: {str(e)}")
|
| 131 |
-
except Exception as e:
|
| 132 |
-
raise HTTPException(status_code=500, detail=f"Error: {str(e)}")
|
| 133 |
-
|
| 134 |
@app.get("/health")
|
| 135 |
async def health_check():
|
| 136 |
"""Health check endpoint"""
|
| 137 |
-
return {"status": "healthy", "service": "
|
| 138 |
|
| 139 |
-
# Chạy với uvicorn (Hugging Face sẽ tự gọi cái này)
|
| 140 |
if __name__ == "__main__":
|
| 141 |
import uvicorn
|
| 142 |
uvicorn.run(app, host="0.0.0.0", port=7860)
|
|
|
|
| 9 |
from pydantic import BaseModel
|
| 10 |
from rapidocr_onnxruntime import RapidOCR
|
| 11 |
|
|
|
|
| 12 |
app = FastAPI(title="OCR API", description="RapidOCR API Service")
|
| 13 |
engine = RapidOCR()
|
| 14 |
|
| 15 |
def process_ocr(image: np.ndarray, use_det: bool, use_cls: bool, use_rec: bool) -> list:
|
| 16 |
"""Xử lý OCR và trả về kết quả"""
|
| 17 |
img_bgr = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
|
|
|
|
| 18 |
result, _ = engine(img_bgr, use_det=use_det, use_cls=use_cls, use_rec=use_rec)
|
| 19 |
|
| 20 |
if not result:
|
|
|
|
| 39 |
|
| 40 |
return texts
|
| 41 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 42 |
@app.post("/ocr")
|
| 43 |
async def ocr_from_file(
|
| 44 |
file: UploadFile = File(..., description="Image file to process"),
|
|
|
|
| 46 |
use_cls: bool = Form(True, description="Use classification"),
|
| 47 |
use_rec: bool = Form(True, description="Use recognition")
|
| 48 |
):
|
|
|
|
|
|
|
|
|
|
| 49 |
if not file.content_type or not file.content_type.startswith('image/'):
|
| 50 |
raise HTTPException(status_code=400, detail="File must be an image")
|
| 51 |
|
| 52 |
try:
|
|
|
|
| 53 |
contents = await file.read()
|
| 54 |
image = Image.open(io.BytesIO(contents))
|
| 55 |
|
|
|
|
| 69 |
except Exception as e:
|
| 70 |
raise HTTPException(status_code=500, detail=f"Error: {str(e)}")
|
| 71 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 72 |
@app.get("/health")
|
| 73 |
async def health_check():
|
| 74 |
"""Health check endpoint"""
|
| 75 |
+
return {"status": "healthy", "service": "XOCR AI"}
|
| 76 |
|
|
|
|
| 77 |
if __name__ == "__main__":
|
| 78 |
import uvicorn
|
| 79 |
uvicorn.run(app, host="0.0.0.0", port=7860)
|