Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -3,18 +3,26 @@ from doctr.io import DocumentFile
|
|
| 3 |
from doctr.models import ocr_predictor
|
| 4 |
import gradio as gr
|
| 5 |
|
|
|
|
| 6 |
os.environ['USE_TORCH'] = '1'
|
| 7 |
|
| 8 |
-
#
|
| 9 |
predictor = ocr_predictor(pretrained=True)
|
| 10 |
|
|
|
|
| 11 |
title = "English OCR"
|
| 12 |
-
description = """Upload an image to extract English text!"""
|
| 13 |
|
|
|
|
| 14 |
def ocr(img):
|
|
|
|
| 15 |
img.save("out.jpg")
|
| 16 |
doc = DocumentFile.from_images("out.jpg")
|
|
|
|
|
|
|
| 17 |
output = predictor(doc)
|
|
|
|
|
|
|
| 18 |
res = ""
|
| 19 |
for page in output.pages:
|
| 20 |
for block in page.blocks:
|
|
@@ -23,20 +31,25 @@ def ocr(img):
|
|
| 23 |
res += word.value + " "
|
| 24 |
res += "\n"
|
| 25 |
res += "\n"
|
| 26 |
-
|
| 27 |
-
|
|
|
|
|
|
|
| 28 |
f.write(res)
|
| 29 |
-
|
|
|
|
| 30 |
|
|
|
|
| 31 |
demo = gr.Interface(
|
| 32 |
fn=ocr,
|
| 33 |
-
inputs=[gr.Image(type="pil")],
|
| 34 |
outputs=[
|
| 35 |
gr.Textbox(lines=20, label="Full Text"),
|
| 36 |
gr.File(label="Download OCR Results")
|
| 37 |
],
|
| 38 |
title=title,
|
| 39 |
-
description=description
|
| 40 |
)
|
| 41 |
|
|
|
|
| 42 |
demo.launch(debug=True)
|
|
|
|
| 3 |
from doctr.models import ocr_predictor
|
| 4 |
import gradio as gr
|
| 5 |
|
| 6 |
+
# เปิดใช้ PyTorch GPU/CPU
|
| 7 |
os.environ['USE_TORCH'] = '1'
|
| 8 |
|
| 9 |
+
# โหลดโมเดล OCR ภาษาอังกฤษ (pretrained)
|
| 10 |
predictor = ocr_predictor(pretrained=True)
|
| 11 |
|
| 12 |
+
# Title/Description สำหรับ Gradio
|
| 13 |
title = "English OCR"
|
| 14 |
+
description = """Upload an image to extract English text from it!"""
|
| 15 |
|
| 16 |
+
# ฟังก์ชัน OCR
|
| 17 |
def ocr(img):
|
| 18 |
+
# บันทึกไฟล์ชั่วคราว
|
| 19 |
img.save("out.jpg")
|
| 20 |
doc = DocumentFile.from_images("out.jpg")
|
| 21 |
+
|
| 22 |
+
# ทำ OCR
|
| 23 |
output = predictor(doc)
|
| 24 |
+
|
| 25 |
+
# รวมผลลัพธ์เป็น string
|
| 26 |
res = ""
|
| 27 |
for page in output.pages:
|
| 28 |
for block in page.blocks:
|
|
|
|
| 31 |
res += word.value + " "
|
| 32 |
res += "\n"
|
| 33 |
res += "\n"
|
| 34 |
+
|
| 35 |
+
# บันทึกผลเป็นไฟล์ txt
|
| 36 |
+
output_name = "RESULT_OCR.txt"
|
| 37 |
+
with open(output_name, "w", encoding="utf-8", errors="ignore") as f:
|
| 38 |
f.write(res)
|
| 39 |
+
|
| 40 |
+
return res, output_name
|
| 41 |
|
| 42 |
+
# สร้าง Gradio Interface
|
| 43 |
demo = gr.Interface(
|
| 44 |
fn=ocr,
|
| 45 |
+
inputs=[gr.Image(type="pil")], # รับภาพจากผู้ใช้
|
| 46 |
outputs=[
|
| 47 |
gr.Textbox(lines=20, label="Full Text"),
|
| 48 |
gr.File(label="Download OCR Results")
|
| 49 |
],
|
| 50 |
title=title,
|
| 51 |
+
description=description
|
| 52 |
)
|
| 53 |
|
| 54 |
+
# Launch
|
| 55 |
demo.launch(debug=True)
|