Spaces:
Sleeping
Sleeping
Upload 2 files
Browse files- app.py +34 -0
- requirements.txt +5 -0
app.py
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import pytesseract
|
| 3 |
+
from PIL import Image
|
| 4 |
+
from docx import Document
|
| 5 |
+
import tempfile
|
| 6 |
+
|
| 7 |
+
# If on Windows, set this path
|
| 8 |
+
pytesseract.pytesseract.tesseract_cmd = r'C:\Program Files\Tesseract-OCR\tesseract.exe'
|
| 9 |
+
|
| 10 |
+
|
| 11 |
+
def image_to_word(image):
|
| 12 |
+
# OCR
|
| 13 |
+
text = pytesseract.image_to_string(Image.open(image))
|
| 14 |
+
|
| 15 |
+
# Create Word document
|
| 16 |
+
doc = Document()
|
| 17 |
+
for line in text.split("\n"):
|
| 18 |
+
doc.add_paragraph(line)
|
| 19 |
+
|
| 20 |
+
# Save temporary docx
|
| 21 |
+
tmp = tempfile.NamedTemporaryFile(delete=False, suffix=".docx")
|
| 22 |
+
doc.save(tmp.name)
|
| 23 |
+
|
| 24 |
+
return tmp.name
|
| 25 |
+
|
| 26 |
+
iface = gr.Interface(
|
| 27 |
+
fn=image_to_word,
|
| 28 |
+
inputs=gr.Image(type="filepath"),
|
| 29 |
+
outputs=gr.File(label="Download Editable Word File"),
|
| 30 |
+
title="Image to Editable Word Converter",
|
| 31 |
+
description="Upload an image of notes or document and get an editable Word file."
|
| 32 |
+
)
|
| 33 |
+
|
| 34 |
+
iface.launch()
|
requirements.txt
ADDED
|
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
gradio
|
| 2 |
+
pytesseract
|
| 3 |
+
pillow
|
| 4 |
+
python-docx
|
| 5 |
+
opencv-python
|