Spaces:
Paused
Paused
Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,113 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import base64
|
| 2 |
+
import io
|
| 3 |
+
import os
|
| 4 |
+
import zipfile
|
| 5 |
+
from dash import Dash, dcc, html, Input, Output, State
|
| 6 |
+
import dash_bootstrap_components as dbc
|
| 7 |
+
from docx import Document
|
| 8 |
+
import markdown
|
| 9 |
+
import threading
|
| 10 |
+
import time
|
| 11 |
+
|
| 12 |
+
app = Dash(__name__, external_stylesheets=[dbc.themes.BOOTSTRAP])
|
| 13 |
+
|
| 14 |
+
app.layout = dbc.Container([
|
| 15 |
+
html.H1("Auto-Wiki", className="my-4"),
|
| 16 |
+
dcc.Upload(
|
| 17 |
+
id='upload-data',
|
| 18 |
+
children=html.Div([
|
| 19 |
+
'Drag and Drop or ',
|
| 20 |
+
html.A('Select Files')
|
| 21 |
+
]),
|
| 22 |
+
style={
|
| 23 |
+
'width': '100%',
|
| 24 |
+
'height': '60px',
|
| 25 |
+
'lineHeight': '60px',
|
| 26 |
+
'borderWidth': '1px',
|
| 27 |
+
'borderStyle': 'dashed',
|
| 28 |
+
'borderRadius': '5px',
|
| 29 |
+
'textAlign': 'center',
|
| 30 |
+
'margin': '10px'
|
| 31 |
+
},
|
| 32 |
+
multiple=True
|
| 33 |
+
),
|
| 34 |
+
html.Div(id='upload-output'),
|
| 35 |
+
dbc.Progress(id="upload-progress", label="Upload Progress", style={"visibility": "hidden"}),
|
| 36 |
+
dbc.Progress(id="conversion-progress", label="Conversion Progress", style={"visibility": "hidden"}),
|
| 37 |
+
dbc.Button("Convert and Download", id="convert-button", color="primary", className="mt-3", disabled=True),
|
| 38 |
+
dcc.Download(id="download-zip")
|
| 39 |
+
])
|
| 40 |
+
|
| 41 |
+
def process_docx(contents, filename):
|
| 42 |
+
content_type, content_string = contents.split(',')
|
| 43 |
+
decoded = base64.b64decode(content_string)
|
| 44 |
+
doc = Document(io.BytesIO(decoded))
|
| 45 |
+
full_text = []
|
| 46 |
+
for para in doc.paragraphs:
|
| 47 |
+
full_text.append(para.text)
|
| 48 |
+
return '\n\n'.join(full_text)
|
| 49 |
+
|
| 50 |
+
@app.callback(
|
| 51 |
+
[Output('upload-output', 'children'),
|
| 52 |
+
Output('convert-button', 'disabled'),
|
| 53 |
+
Output('upload-progress', 'value'),
|
| 54 |
+
Output('upload-progress', 'style'),
|
| 55 |
+
Output('conversion-progress', 'value'),
|
| 56 |
+
Output('conversion-progress', 'style'),
|
| 57 |
+
Output('download-zip', 'data')],
|
| 58 |
+
[Input('upload-data', 'contents'),
|
| 59 |
+
Input('upload-data', 'filename'),
|
| 60 |
+
Input('convert-button', 'n_clicks')],
|
| 61 |
+
[State('upload-data', 'contents'),
|
| 62 |
+
State('upload-data', 'filename')]
|
| 63 |
+
)
|
| 64 |
+
def update_output(list_of_contents, list_of_names, n_clicks, contents, filenames):
|
| 65 |
+
ctx = dash.callback_context
|
| 66 |
+
if not ctx.triggered:
|
| 67 |
+
return dash.no_update
|
| 68 |
+
|
| 69 |
+
if ctx.triggered[0]['prop_id'] == 'upload-data.contents':
|
| 70 |
+
if list_of_contents is not None:
|
| 71 |
+
children = [
|
| 72 |
+
html.Div([
|
| 73 |
+
html.H5(f"File: {name}"),
|
| 74 |
+
html.Hr()
|
| 75 |
+
]) for name in list_of_names
|
| 76 |
+
]
|
| 77 |
+
return children, False, 100, {"visibility": "visible"}, 0, {"visibility": "hidden"}, None
|
| 78 |
+
return dash.no_update
|
| 79 |
+
|
| 80 |
+
if ctx.triggered[0]['prop_id'] == 'convert-button.n_clicks':
|
| 81 |
+
if n_clicks is None:
|
| 82 |
+
return dash.no_update
|
| 83 |
+
|
| 84 |
+
if not contents:
|
| 85 |
+
return dash.no_update
|
| 86 |
+
|
| 87 |
+
def process_files():
|
| 88 |
+
processed_files = []
|
| 89 |
+
for i, (c, n) in enumerate(zip(contents, filenames)):
|
| 90 |
+
text = process_docx(c, n)
|
| 91 |
+
md = markdown.markdown(text)
|
| 92 |
+
processed_files.append((n.replace('.docx', '.md'), md))
|
| 93 |
+
time.sleep(0.1) # Simulate processing time
|
| 94 |
+
app.callback_context.response.set_data(f'{{"progress": {(i+1)/len(contents)*100}}}')
|
| 95 |
+
|
| 96 |
+
zip_buffer = io.BytesIO()
|
| 97 |
+
with zipfile.ZipFile(zip_buffer, 'w', zipfile.ZIP_DEFLATED) as zip_file:
|
| 98 |
+
for name, content in processed_files:
|
| 99 |
+
zip_file.writestr(name, content)
|
| 100 |
+
|
| 101 |
+
return zip_buffer.getvalue()
|
| 102 |
+
|
| 103 |
+
thread = threading.Thread(target=process_files)
|
| 104 |
+
thread.start()
|
| 105 |
+
|
| 106 |
+
return dash.no_update, True, 100, {"visibility": "visible"}, 0, {"visibility": "visible"}, dcc.send_bytes(process_files(), "converted_files.zip")
|
| 107 |
+
|
| 108 |
+
return dash.no_update
|
| 109 |
+
|
| 110 |
+
if __name__ == '__main__':
|
| 111 |
+
print("Starting the Dash application...")
|
| 112 |
+
app.run(debug=True, host='0.0.0.0', port=7860)
|
| 113 |
+
print("Dash application has finished running.")
|