Spaces:
Sleeping
Sleeping
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,204 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from flask import Flask, render_template, request, jsonify
|
2 |
+
import google.generativeai as genai
|
3 |
+
import os
|
4 |
+
from PIL import Image
|
5 |
+
import tempfile
|
6 |
+
|
7 |
+
app = Flask(__name__)
|
8 |
+
|
9 |
+
# Configuration de l'API Gemini
|
10 |
+
token = os.environ.get("TOKEN")
|
11 |
+
genai.configure(api_key=token)
|
12 |
+
|
13 |
+
generation_config = {
|
14 |
+
"temperature": 1,
|
15 |
+
"max_output_tokens": 8192,
|
16 |
+
}
|
17 |
+
|
18 |
+
safety_settings = [
|
19 |
+
{"category": "HARM_CATEGORY_HARASSMENT", "threshold": "BLOCK_NONE"},
|
20 |
+
{"category": "HARM_CATEGORY_HATE_SPEECH", "threshold": "BLOCK_NONE"},
|
21 |
+
{"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", "threshold": "BLOCK_NONE"},
|
22 |
+
{"category": "HARM_CATEGORY_DANGEROUS_CONTENT", "threshold": "BLOCK_NONE"},
|
23 |
+
]
|
24 |
+
|
25 |
+
|
26 |
+
@app.route('/')
|
27 |
+
def generale():
|
28 |
+
return render_template("generale.html")
|
29 |
+
|
30 |
+
|
31 |
+
|
32 |
+
|
33 |
+
def upload_and_process_file(file_path):
|
34 |
+
"""Upload et traite un fichier avec l'API Gemini avec gestion des erreurs améliorée"""
|
35 |
+
max_retries = 3
|
36 |
+
retry_delay = 2 # secondes
|
37 |
+
|
38 |
+
for attempt in range(max_retries):
|
39 |
+
try:
|
40 |
+
print(f"Tentative d'upload {attempt + 1}/{max_retries} pour {file_path}")
|
41 |
+
|
42 |
+
# Vérification du fichier
|
43 |
+
if not os.path.exists(file_path):
|
44 |
+
raise FileNotFoundError(f"Le fichier {file_path} n'existe pas")
|
45 |
+
|
46 |
+
file_size = os.path.getsize(file_path)
|
47 |
+
if file_size == 0:
|
48 |
+
raise ValueError(f"Le fichier {file_path} est vide")
|
49 |
+
|
50 |
+
# Upload du fichier
|
51 |
+
uploaded_file = genai.upload_file(path=file_path)
|
52 |
+
print(f"Upload réussi: {uploaded_file.uri}")
|
53 |
+
|
54 |
+
# Attente du traitement
|
55 |
+
timeout = 300 # 5 minutes
|
56 |
+
start_time = time.time()
|
57 |
+
|
58 |
+
while uploaded_file.state.name == "PROCESSING":
|
59 |
+
if time.time() - start_time > timeout:
|
60 |
+
raise TimeoutError("Timeout pendant le traitement du fichier")
|
61 |
+
|
62 |
+
print(f"En attente du traitement... Temps écoulé: {int(time.time() - start_time)}s")
|
63 |
+
time.sleep(10)
|
64 |
+
uploaded_file = genai.get_file(uploaded_file.name)
|
65 |
+
|
66 |
+
if uploaded_file.state.name == "FAILED":
|
67 |
+
raise ValueError(f"Échec du traitement: {uploaded_file.state.name}")
|
68 |
+
|
69 |
+
print(f"Traitement terminé avec succès: {uploaded_file.uri}")
|
70 |
+
return uploaded_file
|
71 |
+
|
72 |
+
except ssl.SSLError as e:
|
73 |
+
print(f"Erreur SSL lors de l'upload (tentative {attempt + 1}): {e}")
|
74 |
+
if attempt < max_retries - 1:
|
75 |
+
time.sleep(retry_delay * (attempt + 1))
|
76 |
+
else:
|
77 |
+
raise
|
78 |
+
|
79 |
+
except Exception as e:
|
80 |
+
print(f"Erreur lors de l'upload (tentative {attempt + 1}): {e}")
|
81 |
+
if attempt < max_retries - 1:
|
82 |
+
time.sleep(retry_delay * (attempt + 1))
|
83 |
+
else:
|
84 |
+
raise
|
85 |
+
|
86 |
+
|
87 |
+
|
88 |
+
def is_youtube_url(url):
|
89 |
+
"""Vérifie si l'URL est une URL YouTube"""
|
90 |
+
parsed = urlparse(url)
|
91 |
+
return any(domain in parsed.netloc for domain in ['youtube.com', 'youtu.be'])
|
92 |
+
|
93 |
+
def download_youtube_video(url):
|
94 |
+
"""Télécharge une vidéo YouTube avec gestion des erreurs améliorée"""
|
95 |
+
try:
|
96 |
+
with tempfile.TemporaryDirectory() as temp_dir:
|
97 |
+
ydl_opts = {
|
98 |
+
'format': 'best[filesize<50M]', # Limite la taille du fichier
|
99 |
+
'outtmpl': os.path.join(temp_dir, '%(title)s.%(ext)s'),
|
100 |
+
'quiet': True,
|
101 |
+
'no_warnings': True,
|
102 |
+
'extract_flat': False
|
103 |
+
}
|
104 |
+
|
105 |
+
with yt_dlp.YoutubeDL(ydl_opts) as ydl:
|
106 |
+
print(f"Téléchargement de la vidéo: {url}")
|
107 |
+
info = ydl.extract_info(url, download=True)
|
108 |
+
video_path = os.path.join(temp_dir, ydl.prepare_filename(info))
|
109 |
+
|
110 |
+
if not os.path.exists(video_path):
|
111 |
+
raise FileNotFoundError(f"La vidéo n'a pas été téléchargée correctement: {video_path}")
|
112 |
+
|
113 |
+
# Copie vers un fichier temporaire permanent
|
114 |
+
temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=os.path.splitext(video_path)[1])
|
115 |
+
with open(video_path, 'rb') as f:
|
116 |
+
temp_file.write(f.read())
|
117 |
+
|
118 |
+
print(f"Vidéo téléchargée avec succès: {temp_file.name}")
|
119 |
+
return temp_file.name
|
120 |
+
|
121 |
+
except Exception as e:
|
122 |
+
print(f"Erreur lors du téléchargement de la vidéo: {e}")
|
123 |
+
return None
|
124 |
+
|
125 |
+
def telecharger_pdf(url):
|
126 |
+
"""Télécharge un PDF et retourne le chemin du fichier"""
|
127 |
+
try:
|
128 |
+
response = requests.get(url)
|
129 |
+
response.raise_for_status()
|
130 |
+
|
131 |
+
with tempfile.NamedTemporaryFile(delete=False, suffix='.pdf') as temp_file:
|
132 |
+
temp_file.write(response.content)
|
133 |
+
return temp_file.name
|
134 |
+
except Exception as e:
|
135 |
+
print(f"Erreur lors du téléchargement du PDF : {e}")
|
136 |
+
return None
|
137 |
+
|
138 |
+
def allowed_file(filename):
|
139 |
+
"""Vérifie si l'extension du fichier est autorisée"""
|
140 |
+
ALLOWED_EXTENSIONS = {'pdf', 'png', 'jpg', 'jpeg', 'gif', 'mp4', 'mov', 'avi'}
|
141 |
+
return '.' in filename and filename.rsplit('.', 1)[1].lower() in ALLOWED_EXTENSIONS
|
142 |
+
|
143 |
+
@app.route('/submit', methods=['POST'])
|
144 |
+
def submit_question():
|
145 |
+
question = request.form.get('question')
|
146 |
+
urls = request.form.getlist('urls')
|
147 |
+
files = request.files.getlist('files')
|
148 |
+
|
149 |
+
print("URLs reçues:", urls)
|
150 |
+
content = [question]
|
151 |
+
temp_files = []
|
152 |
+
|
153 |
+
try:
|
154 |
+
# Traitement des fichiers uploadés
|
155 |
+
for file in files:
|
156 |
+
if file and allowed_file(file.filename):
|
157 |
+
with tempfile.NamedTemporaryFile(delete=False, suffix=os.path.splitext(file.filename)[1]) as temp_file:
|
158 |
+
file.save(temp_file.name)
|
159 |
+
temp_files.append(temp_file.name)
|
160 |
+
if file.filename.lower().endswith(('.png', '.jpg', '.jpeg', '.gif')):
|
161 |
+
content.append(PIL.Image.open(temp_file.name))
|
162 |
+
else:
|
163 |
+
uploaded_file = upload_and_process_file(temp_file.name)
|
164 |
+
content.append(uploaded_file)
|
165 |
+
|
166 |
+
# Traitement des URLs
|
167 |
+
for url in urls:
|
168 |
+
print(f"Traitement de l'URL : {url}")
|
169 |
+
if is_youtube_url(url):
|
170 |
+
video_path = download_youtube_video(url)
|
171 |
+
if video_path:
|
172 |
+
temp_files.append(video_path)
|
173 |
+
uploaded_file = upload_and_process_file(video_path)
|
174 |
+
content.append(uploaded_file)
|
175 |
+
elif url.lower().endswith('.pdf'):
|
176 |
+
pdf_path = telecharger_pdf(url)
|
177 |
+
if pdf_path:
|
178 |
+
temp_files.append(pdf_path)
|
179 |
+
uploaded_file = upload_and_process_file(pdf_path)
|
180 |
+
content.append(uploaded_file)
|
181 |
+
|
182 |
+
|
183 |
+
|
184 |
+
|
185 |
+
# Génération de contenu avec Gemini
|
186 |
+
model = genai.GenerativeModel(
|
187 |
+
model_name="models/gemini-1.5-flash-002",
|
188 |
+
safety_settings=safety_settings,
|
189 |
+
system_instruction="Tu es un assistant intelligent. ton but est d'assister au mieux que tu peux. tu as été créé par Aenir et tu t'appelles Mariam."
|
190 |
+
)
|
191 |
+
response = model.generate_content(content, request_options={"timeout": 600})
|
192 |
+
return jsonify({"response": response.text})
|
193 |
+
|
194 |
+
except Exception as e:
|
195 |
+
return jsonify({"error": str(e)}), 500
|
196 |
+
finally:
|
197 |
+
# Nettoyage des fichiers temporaires
|
198 |
+
for temp_file in temp_files:
|
199 |
+
try:
|
200 |
+
os.unlink(temp_file)
|
201 |
+
except Exception as e:
|
202 |
+
print(f"Erreur lors de la suppression du fichier temporaire {temp_file}: {e}")
|
203 |
+
|
204 |
+
|