Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -2,7 +2,7 @@ import gradio as gr
|
|
2 |
from database import save_to_db
|
3 |
from preprocessing import read_file
|
4 |
|
5 |
-
def process_file(
|
6 |
"""
|
7 |
Processes an uploaded file, extracts its text content, and saves it to the database.
|
8 |
|
@@ -36,24 +36,25 @@ def process_file(file, topic):
|
|
36 |
>>> process_file(unsupported_file, ["Science"])
|
37 |
'Error processing file: Unsupported file format. Only .docx, .txt, and .pdf are allowed.'
|
38 |
"""
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
|
|
51 |
|
52 |
# Define Gradio interface
|
53 |
with gr.Blocks() as demo:
|
54 |
gr.Markdown("# Dataset Upload Interface")
|
55 |
with gr.Row():
|
56 |
-
file_input = gr.File(label="Upload File (.docx or .txt or .pdf)")
|
57 |
topic_input = gr.Textbox(label="Topics (comma-separated)", placeholder="e.g., science, technology, law, medicin")
|
58 |
submit_button = gr.Button("Upload and Process")
|
59 |
output_text = gr.Textbox(label="Status")
|
|
|
2 |
from database import save_to_db
|
3 |
from preprocessing import read_file
|
4 |
|
5 |
+
def process_file(files, topic):
|
6 |
"""
|
7 |
Processes an uploaded file, extracts its text content, and saves it to the database.
|
8 |
|
|
|
36 |
>>> process_file(unsupported_file, ["Science"])
|
37 |
'Error processing file: Unsupported file format. Only .docx, .txt, and .pdf are allowed.'
|
38 |
"""
|
39 |
+
for file in files:
|
40 |
+
try:
|
41 |
+
# Read the file content
|
42 |
+
file_path = file.name
|
43 |
+
text = read_file(file_path)
|
44 |
+
|
45 |
+
# Spl
|
46 |
+
# Save chunks to database
|
47 |
+
save_to_db([text], topic)
|
48 |
+
|
49 |
+
return f"File {file_path} processed successfully! file saved to the database."
|
50 |
+
except Exception as e:
|
51 |
+
return f"Error processing for file {file_path}: {str(e)}"
|
52 |
|
53 |
# Define Gradio interface
|
54 |
with gr.Blocks() as demo:
|
55 |
gr.Markdown("# Dataset Upload Interface")
|
56 |
with gr.Row():
|
57 |
+
file_input = gr.File(label="Upload File (.docx or .txt or .pdf)", file_count="multiple")
|
58 |
topic_input = gr.Textbox(label="Topics (comma-separated)", placeholder="e.g., science, technology, law, medicin")
|
59 |
submit_button = gr.Button("Upload and Process")
|
60 |
output_text = gr.Textbox(label="Status")
|