File size: 10,105 Bytes
4deb54c aef7e33 4deb54c aef7e33 4deb54c d479d0f 4deb54c 1f593cd 4deb54c aef7e33 4deb54c aef7e33 4deb54c aef7e33 4deb54c aef7e33 4deb54c aef7e33 4deb54c aef7e33 4deb54c aef7e33 4deb54c aef7e33 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 |
from flask import Flask, render_template,request, redirect,url_for, jsonify , session
from helper_functions import predict_class , prepare_text , inference , predict , align_predictions_with_sentences , load_models
import fitz # PyMuPDF
import os, shutil
import torch
import tempfile
from pydub import AudioSegment
import logging
app = Flask(__name__)
app.config['UPLOAD_FOLDER'] = 'static/uploads'
# Global variables for models
global_model = None
global_neptune = None
global_tokenizer = None
global_pipe = None
def init_app():
global global_model, global_neptune, global_pipe
print("Loading models...")
global_model, global_neptune, global_pipe = load_models()
print("Models loaded successfully!")
init_app()
@app.route("/")
def home():
predict_class = ""
class_probabilities = dict()
chart_data = dict()
return render_template('pdf.html', class_probabilities= class_probabilities, predicted_class=predict_class,chart_data = chart_data)
@app.route('/pdf')
def pdf():
predict_class = ""
class_probabilities = dict()
chart_data = dict()
return render_template('pdf.html', class_probabilities= class_probabilities, predicted_class=predict_class,chart_data = chart_data)
@app.route('/pdf/upload' , methods = ['POST'])
def treatment():
global global_model, global_tokenizer
if request.method == 'POST' :
# Récupérer le fichier PDF de la requête
file = request.files['file']
filename = file.filename
# Enregistrer le fichier dans le répertoire de téléchargement
filepath = app.config['UPLOAD_FOLDER'] + "/" + filename
file.save(filepath)
# Ouvrir le fichier PDF
pdf_document = fitz.open(filepath)
# Initialiser une variable pour stocker le texte extrait
extracted_text = ""
# Boucler à travers chaque page pour extraire le texte
for page_num in range(len(pdf_document)):
# Récupérer l'objet de la page
page = pdf_document.load_page(page_num)
# Extraire le texte de la page
page_text = page.get_text()
# Ajouter le texte de la page à la variable d'extraction
extracted_text += f"\nPage {page_num + 1}:\n{page_text}"
# Fermer le fichier PDF
pdf_document.close()
# Prepare data for the chart
predicted_class , class_probabilities = predict_class([extracted_text] , global_model)
chart_data = {
'datasets': [{
'data': list(class_probabilities.values()),
'backgroundColor': [color[2] for color in class_probabilities.keys()],
'borderColor': [color[2] for color in class_probabilities.keys()]
}],
'labels': [label[0] for label in class_probabilities.keys()]
}
print(predict_class)
print(chart_data)
# clear the uploads folder
for filename in os.listdir(app.config['UPLOAD_FOLDER']):
file_path = os.path.join(app.config['UPLOAD_FOLDER'], filename)
try:
if os.path.isfile(file_path) or os.path.islink(file_path):
os.unlink(file_path)
elif os.path.isdir(file_path):
shutil.rmtree(file_path)
except Exception as e:
print('Failed to delete %s. Reason: %s' % (file_path, e))
return render_template('pdf.html',extracted_text = extracted_text, class_probabilities=class_probabilities, predicted_class=predicted_class, chart_data = chart_data)
return render_template('pdf.html')
## Sentence
@app.route('/sentence' , methods = ['GET' , 'POST'])
def sentence():
global global_model, global_tokenizer
if request.method == 'POST':
# Get the form data
text = [request.form['text']]
predicted_class , class_probabilities = predict_class(text , global_model)
# Prepare data for the chart
chart_data = {
'datasets': [{
'data': list(class_probabilities.values()),
'backgroundColor': [color[2 ] for color in class_probabilities.keys()],
'borderColor': [color[2] for color in class_probabilities.keys()]
}],
'labels': [label[0] for label in class_probabilities.keys()]
}
print(chart_data)
# clear the uploads folder
for filename in os.listdir(app.config['UPLOAD_FOLDER']):
file_path = os.path.join(app.config['UPLOAD_FOLDER'], filename)
try:
if os.path.isfile(file_path) or os.path.islink(file_path):
os.unlink(file_path)
elif os.path.isdir(file_path):
shutil.rmtree(file_path)
except Exception as e:
print('Failed to delete %s. Reason: %s' % (file_path, e))
return render_template('response_sentence.html', text=text, class_probabilities=class_probabilities, predicted_class=predicted_class,chart_data = chart_data)
# Render the initial form page
return render_template('sentence.html')
## Voice
@app.route("/voice_backup")
def slu_backup():
input_file = "static/uploads/2022.jep-architectures-neuronales.pdf"
# Ouvrir le fichier PDF
pdf_document = fitz.open(input_file)
# Initialiser une variable pour stocker le texte extrait
extracted_text = ""
# Boucler à travers chaque page pour extraire le texte
for page_num in range(len(pdf_document)):
# Récupérer l'objet de la page
page = pdf_document.load_page(page_num)
# Extraire le texte de la page
page_text = page.get_text()
# Ajouter le texte de la page à la variable d'extraction
extracted_text += f"\nPage {page_num + 1}:\n{page_text}"
# Fermer le fichier PDF
pdf_document.close()
# Prepare data for the chart
inference_batch, sentences = inference(extracted_text)
predictions = predict(inference_batch)
sentences_prediction = align_predictions_with_sentences(sentences, predictions)
predicted_class , class_probabilities = predict_class([extracted_text] , global_model)
chart_data = {
'datasets': [{
'data': list(class_probabilities.values()),
'backgroundColor': [color[2 ] for color in class_probabilities.keys()],
'borderColor': [color[2] for color in class_probabilities.keys()]
}],
'labels': [label[0] for label in class_probabilities.keys()]
}
print(class_probabilities)
print(chart_data)
print(sentences_prediction)
return render_template('voice_backup.html',extracted_text = extracted_text, class_probabilities=class_probabilities, predicted_class=predicted_class, chart_data = chart_data, sentences_prediction = sentences_prediction)
logging.basicConfig(level=logging.DEBUG)
@app.route("/voice", methods=['GET', 'POST'])
def slu():
global global_neptune, global_pipe, global_model
if request.method == 'POST':
logging.debug("Received POST request")
audio_file = request.files.get('audio')
if audio_file:
logging.debug(f"Received audio file: {audio_file.filename}")
# Save audio data to a temporary file
with tempfile.NamedTemporaryFile(delete=False, suffix='.wav') as temp_audio:
audio_file.save(temp_audio)
temp_audio_path = temp_audio.name
logging.debug(f"Saved audio to temporary file: {temp_audio_path}")
try:
# Transcribe audio using Whisper
result = global_pipe(temp_audio_path)
extracted_text = result["text"]
logging.debug(f"Transcribed text: {extracted_text}")
# Process the transcribed text
inference_batch, sentences = inference(extracted_text)
predictions = predict(inference_batch, global_neptune)
sentences_prediction = align_predictions_with_sentences(sentences, predictions)
predicted_class, class_probabilities = predict_class([extracted_text], global_model)
chart_data = {
'datasets': [{
'data': list(class_probabilities.values()),
'backgroundColor': [color[2] for color in class_probabilities.keys()],
'borderColor': [color[2] for color in class_probabilities.keys()]
}],
'labels': [label[0] for label in class_probabilities.keys()]
}
response_data = {
'extracted_text': extracted_text,
'class_probabilities' : class_probabilities,
'predicted_class': predicted_class,
'chart_data': chart_data,
'sentences_prediction': sentences_prediction
}
logging.debug(f"Prepared response data: {response_data}")
return render_template('voice.html',
class_probabilities= class_probabilities,
predicted_class= predicted_class,
chart_data= chart_data,
sentences_prediction=sentences_prediction)
except Exception as e:
logging.error(f"Error processing audio: {str(e)}")
return jsonify({'error': str(e)}), 500
finally:
# Remove temporary file
os.unlink(temp_audio_path)
else:
logging.error("No audio file received")
return jsonify({'error': 'No audio file received'}), 400
# For GET request
logging.debug("Received GET request")
return render_template('voice.html',
class_probabilities={},
predicted_class=[""],
chart_data={},
sentences_prediction={})
if __name__ == '__main__':
app.run(debug=True) |