Upload 2 files
Browse files- app.py +69 -0
- requirements.txt +4 -0
app.py
ADDED
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import os
|
2 |
+
import requests
|
3 |
+
import base64
|
4 |
+
import json
|
5 |
+
import gradio as gr
|
6 |
+
from PIL import Image, ImageDraw, ImageFont
|
7 |
+
|
8 |
+
# Load secrets from environment variables (configure these in your Hugging Face Space)
|
9 |
+
API_KEY = os.environ.get("API_KEY")
|
10 |
+
OCR_API_URL = os.environ.get("OCR_API_URL")
|
11 |
+
OCR_TASK = os.environ.get("OCR_TASK")
|
12 |
+
OCR_FUNCTION_NAME = os.environ.get("OCR_FUNCTION_NAME")
|
13 |
+
|
14 |
+
def call_ocr(image_path: str):
|
15 |
+
# Read and encode the image to base64
|
16 |
+
with open(image_path, "rb") as image_file:
|
17 |
+
base64_string = base64.b64encode(image_file.read()).decode('utf-8')
|
18 |
+
|
19 |
+
payload = {
|
20 |
+
"image": base64_string,
|
21 |
+
"task": OCR_TASK,
|
22 |
+
"function_name": OCR_FUNCTION_NAME
|
23 |
+
}
|
24 |
+
|
25 |
+
headers = {
|
26 |
+
"Content-Type": "application/json",
|
27 |
+
"Accept": "application/json",
|
28 |
+
"Authorization": f"Basic {API_KEY}"
|
29 |
+
}
|
30 |
+
|
31 |
+
response = requests.post(OCR_API_URL, json=payload, headers=headers)
|
32 |
+
response.raise_for_status()
|
33 |
+
response_json = response.json()
|
34 |
+
|
35 |
+
# Extract quad_boxes and labels
|
36 |
+
data = response_json.get('data', {}).get(OCR_TASK, {})
|
37 |
+
quad_boxes = data.get('quad_boxes', [])
|
38 |
+
labels = data.get('labels', [])
|
39 |
+
|
40 |
+
# Clean up labels
|
41 |
+
cleaned_labels = [label.replace("</s>", "").strip() for label in labels]
|
42 |
+
|
43 |
+
# Open the original image and draw boxes
|
44 |
+
image = Image.open(image_path)
|
45 |
+
draw = ImageDraw.Draw(image)
|
46 |
+
|
47 |
+
# Use a true-type font if available
|
48 |
+
try:
|
49 |
+
font = ImageFont.truetype("arial.ttf", 16)
|
50 |
+
except IOError:
|
51 |
+
font = None
|
52 |
+
|
53 |
+
for quad, label in zip(quad_boxes, cleaned_labels):
|
54 |
+
polygon = [(quad[i], quad[i+1]) for i in range(0, len(quad), 2)]
|
55 |
+
draw.polygon(polygon, outline="red", width=2)
|
56 |
+
draw.text((quad[0], quad[1]), label, fill="blue", font=font)
|
57 |
+
|
58 |
+
output_labels = "\n".join(cleaned_labels)
|
59 |
+
return image, output_labels
|
60 |
+
|
61 |
+
iface = gr.Interface(
|
62 |
+
fn=call_ocr,
|
63 |
+
inputs=gr.Image(type="filepath", label="Upload Image"),
|
64 |
+
outputs=[gr.Image(label="Annotated Image"), gr.Textbox(label="Detected Labels")],
|
65 |
+
title="Optical Character Recognition (OCR)"
|
66 |
+
)
|
67 |
+
|
68 |
+
if __name__ == "__main__":
|
69 |
+
iface.launch(debug=True)
|
requirements.txt
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
gradio
|
2 |
+
groq
|
3 |
+
Pillow
|
4 |
+
requests
|