Spaces:
Running
Running
Delete app.py
Browse files
app.py
DELETED
|
@@ -1,116 +0,0 @@
|
|
| 1 |
-
from fastapi import FastAPI, UploadFile, File, Query
|
| 2 |
-
from fastapi.responses import JSONResponse
|
| 3 |
-
import easyocr
|
| 4 |
-
import numpy as np
|
| 5 |
-
import cv2
|
| 6 |
-
import time
|
| 7 |
-
import requests
|
| 8 |
-
from io import BytesIO
|
| 9 |
-
from PIL import Image
|
| 10 |
-
import os
|
| 11 |
-
|
| 12 |
-
app = FastAPI(title="EasyOCR API")
|
| 13 |
-
|
| 14 |
-
# Use /tmp for writable storage
|
| 15 |
-
tmp_dir = "/tmp/EasyOCR"
|
| 16 |
-
model_dir = os.path.join(tmp_dir, "models")
|
| 17 |
-
user_net_dir = os.path.join(tmp_dir, "user_network")
|
| 18 |
-
os.makedirs(model_dir, exist_ok=True)
|
| 19 |
-
os.makedirs(user_net_dir, exist_ok=True)
|
| 20 |
-
|
| 21 |
-
# Initialize reader with custom directories
|
| 22 |
-
reader = easyocr.Reader(
|
| 23 |
-
['en'],
|
| 24 |
-
gpu=False,
|
| 25 |
-
model_storage_directory=model_dir,
|
| 26 |
-
user_network_directory=user_net_dir
|
| 27 |
-
)
|
| 28 |
-
|
| 29 |
-
def read_image_from_url(url: str):
|
| 30 |
-
response = requests.get(url)
|
| 31 |
-
response.raise_for_status()
|
| 32 |
-
img = np.array(Image.open(BytesIO(response.content)).convert("RGB"))
|
| 33 |
-
return img
|
| 34 |
-
|
| 35 |
-
def preprocess_image(image: np.ndarray) -> np.ndarray:
|
| 36 |
-
# Convert to grayscale
|
| 37 |
-
gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
|
| 38 |
-
|
| 39 |
-
# Denoise slightly
|
| 40 |
-
gray = cv2.GaussianBlur(gray, (3, 3), 0)
|
| 41 |
-
|
| 42 |
-
# Adaptive threshold (makes text sharper)
|
| 43 |
-
thresh = cv2.adaptiveThreshold(
|
| 44 |
-
gray, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
|
| 45 |
-
cv2.THRESH_BINARY, 31, 2
|
| 46 |
-
)
|
| 47 |
-
|
| 48 |
-
return thresh
|
| 49 |
-
|
| 50 |
-
@app.get("/")
|
| 51 |
-
async def ocr_url(url: str = Query(None)):
|
| 52 |
-
start_time = time.time()
|
| 53 |
-
|
| 54 |
-
if not url:
|
| 55 |
-
return JSONResponse(
|
| 56 |
-
status_code=400,
|
| 57 |
-
content={"status": False, "error": "Provide ?url= for OCR."}
|
| 58 |
-
)
|
| 59 |
-
|
| 60 |
-
try:
|
| 61 |
-
image = read_image_from_url(url)
|
| 62 |
-
|
| 63 |
-
# Preprocess for better OCR
|
| 64 |
-
processed = preprocess_image(image)
|
| 65 |
-
|
| 66 |
-
# OCR
|
| 67 |
-
result = reader.readtext(processed)
|
| 68 |
-
|
| 69 |
-
# Format result (only text)
|
| 70 |
-
output = [r[1] for r in result]
|
| 71 |
-
|
| 72 |
-
end_time = time.time()
|
| 73 |
-
time_taken = round(end_time - start_time, 3)
|
| 74 |
-
|
| 75 |
-
return {
|
| 76 |
-
"status": True,
|
| 77 |
-
"time_taken": time_taken,
|
| 78 |
-
"results": output
|
| 79 |
-
}
|
| 80 |
-
|
| 81 |
-
except Exception as e:
|
| 82 |
-
return JSONResponse(
|
| 83 |
-
status_code=500,
|
| 84 |
-
content={"status": False, "error": str(e)}
|
| 85 |
-
)
|
| 86 |
-
|
| 87 |
-
@app.post("/")
|
| 88 |
-
async def ocr_file(file: UploadFile = File(...)):
|
| 89 |
-
start_time = time.time()
|
| 90 |
-
|
| 91 |
-
try:
|
| 92 |
-
image = np.array(Image.open(file.file).convert("RGB"))
|
| 93 |
-
|
| 94 |
-
# Preprocess for better OCR
|
| 95 |
-
processed = preprocess_image(image)
|
| 96 |
-
|
| 97 |
-
# OCR
|
| 98 |
-
result = reader.readtext(processed)
|
| 99 |
-
|
| 100 |
-
# Format result (only text)
|
| 101 |
-
output = [r[1] for r in result]
|
| 102 |
-
|
| 103 |
-
end_time = time.time()
|
| 104 |
-
time_taken = round(end_time - start_time, 3)
|
| 105 |
-
|
| 106 |
-
return {
|
| 107 |
-
"status": True,
|
| 108 |
-
"time_taken": time_taken,
|
| 109 |
-
"results": output
|
| 110 |
-
}
|
| 111 |
-
|
| 112 |
-
except Exception as e:
|
| 113 |
-
return JSONResponse(
|
| 114 |
-
status_code=500,
|
| 115 |
-
content={"status": False, "error": str(e)}
|
| 116 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|