Spaces:
Build error
Build error
Update app.py
Browse files
app.py
CHANGED
|
@@ -418,70 +418,36 @@ def anonymization_with_fn(selected_sentences, query):
|
|
| 418 |
identified_words_output_df: gr.update(value=identified_df, visible=False),
|
| 419 |
}
|
| 420 |
|
|
|
|
|
|
|
| 421 |
|
| 422 |
-
|
|
|
|
|
|
|
|
|
|
| 423 |
|
| 424 |
-
|
|
|
|
|
|
|
|
|
|
| 425 |
|
| 426 |
-
if not (KEYS_DIR / f"{USER_ID}/evaluation_key").is_file():
|
| 427 |
-
error_message = "Error ❌: Please generate the key first!"
|
| 428 |
-
return {chatgpt_response_anonymized: gr.update(value=error_message)}
|
| 429 |
-
|
| 430 |
-
if not (CLIENT_DIR / f"{USER_ID}_encrypted_output").is_file():
|
| 431 |
-
error_message = "Error ❌: Please encrypt your query first!"
|
| 432 |
-
return {chatgpt_response_anonymized: gr.update(value=error_message)}
|
| 433 |
-
|
| 434 |
-
context_prompt = read_txt(PROMPT_PATH)
|
| 435 |
-
|
| 436 |
-
# Prepare prompt
|
| 437 |
-
query = (
|
| 438 |
-
"Document content:\n```\n"
|
| 439 |
-
+ anonymized_document
|
| 440 |
-
+ "\n\n```"
|
| 441 |
-
+ "Query:\n```\n"
|
| 442 |
-
+ anonymized_query
|
| 443 |
-
+ "\n```"
|
| 444 |
-
)
|
| 445 |
-
print(f'Prompt of CHATGPT:\n{query}')
|
| 446 |
-
|
| 447 |
-
completion = client.chat.completions.create(
|
| 448 |
-
model="gpt-4-1106-preview", # Replace with "gpt-4" if available
|
| 449 |
-
messages=[
|
| 450 |
-
{"role": "system", "content": context_prompt},
|
| 451 |
-
{"role": "user", "content": query},
|
| 452 |
-
],
|
| 453 |
-
)
|
| 454 |
-
anonymized_response = completion.choices[0].message.content
|
| 455 |
-
uuid_map = read_json(MAPPING_UUID_PATH)
|
| 456 |
-
|
| 457 |
-
inverse_uuid_map = {
|
| 458 |
-
v: k for k, v in uuid_map.items()
|
| 459 |
-
} # TODO load the inverse mapping from disk for efficiency
|
| 460 |
-
|
| 461 |
-
# Pattern to identify words and non-words (including punctuation, spaces, etc.)
|
| 462 |
-
tokens = re.findall(r"(\b[\w\.\/\-@]+\b|[\s,.!?;:'\"-]+)", anonymized_response)
|
| 463 |
-
processed_tokens = []
|
| 464 |
-
|
| 465 |
-
for token in tokens:
|
| 466 |
-
# Directly append non-word tokens or whitespace to processed_tokens
|
| 467 |
-
if not token.strip() or not re.match(r"\w+", token):
|
| 468 |
-
processed_tokens.append(token)
|
| 469 |
-
continue
|
| 470 |
-
|
| 471 |
-
if token in inverse_uuid_map:
|
| 472 |
-
processed_tokens.append(inverse_uuid_map[token])
|
| 473 |
-
else:
|
| 474 |
-
processed_tokens.append(token)
|
| 475 |
-
deanonymized_response = "".join(processed_tokens)
|
| 476 |
-
|
| 477 |
-
return {chatgpt_response_anonymized: gr.update(value=anonymized_response),
|
| 478 |
-
chatgpt_response_deanonymized: gr.update(value=deanonymized_response)}
|
| 479 |
|
| 480 |
|
| 481 |
demo = gr.Blocks(css=".markdown-body { font-size: 18px; }")
|
| 482 |
|
| 483 |
with demo:
|
| 484 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 485 |
gr.Markdown(
|
| 486 |
"""
|
| 487 |
<p align="center">
|
|
|
|
| 418 |
identified_words_output_df: gr.update(value=identified_df, visible=False),
|
| 419 |
}
|
| 420 |
|
| 421 |
+
# Define the folder path containing audio files
|
| 422 |
+
AUDIO_FOLDER_PATH = "./files/"
|
| 423 |
|
| 424 |
+
# Function to list available audio files in the folder
|
| 425 |
+
def get_audio_files():
|
| 426 |
+
files = [f for f in os.listdir(AUDIO_FOLDER_PATH) if f.endswith(('.wav', '.mp3'))]
|
| 427 |
+
return files
|
| 428 |
|
| 429 |
+
# Step 1: Load and display audio file
|
| 430 |
+
def load_audio_file(selected_audio):
|
| 431 |
+
file_path = os.path.join(AUDIO_FOLDER_PATH, selected_audio)
|
| 432 |
+
return file_path
|
| 433 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 434 |
|
| 435 |
|
| 436 |
demo = gr.Blocks(css=".markdown-body { font-size: 18px; }")
|
| 437 |
|
| 438 |
with demo:
|
| 439 |
|
| 440 |
+
# Step 1: Add an audio file
|
| 441 |
+
gr.Markdown("## Step 1: Add an Audio File")
|
| 442 |
+
audio_files = get_audio_files()
|
| 443 |
+
|
| 444 |
+
with gr.Row():
|
| 445 |
+
audio_file_dropdown = gr.Dropdown(audio_files, label="Select an Audio File", interactive=True)
|
| 446 |
+
audio_output = gr.Audio(label="Selected Audio", type="filepath")
|
| 447 |
+
|
| 448 |
+
# When an audio file is selected, it will display the file path
|
| 449 |
+
audio_file_dropdown.change(fn=load_audio_file, inputs=[audio_file_dropdown], outputs=[audio_output])
|
| 450 |
+
|
| 451 |
gr.Markdown(
|
| 452 |
"""
|
| 453 |
<p align="center">
|