Update demo_script.py
Browse files- demo_script.py +154 -154
demo_script.py
CHANGED
@@ -1,154 +1,154 @@
|
|
1 |
-
import csv
|
2 |
-
import os
|
3 |
-
import sys
|
4 |
-
import base64
|
5 |
-
import requests
|
6 |
-
import logging
|
7 |
-
from pathlib import Path
|
8 |
-
|
9 |
-
# ==================== CONFIGURAZIONE ====================
|
10 |
-
|
11 |
-
API_KEY = "ChiaveAPI"
|
12 |
-
MODEL_NAME = "gpt-4o"
|
13 |
-
IMAGE_DIR = Path("images")
|
14 |
-
CSV_PATH = Path("
|
15 |
-
OUTPUT_CSV = Path("risposte_output.csv")
|
16 |
-
|
17 |
-
CATEGORIE = {
|
18 |
-
"AR": "art_recognition",
|
19 |
-
"CR": "chronological_reasoning",
|
20 |
-
"CS": "contextual_summary",
|
21 |
-
"VR": "vision_reading",
|
22 |
-
"VB": "vision_basic",
|
23 |
-
"VL": "vision_logic",
|
24 |
-
"VRS": "vision_reasoning",
|
25 |
-
"IG" : "img_gen"
|
26 |
-
}
|
27 |
-
|
28 |
-
# ==================== LOGGING ====================
|
29 |
-
|
30 |
-
logging.basicConfig(
|
31 |
-
filename='task_log.txt',
|
32 |
-
level=logging.INFO,
|
33 |
-
format='%(asctime)s - %(levelname)s - %(message)s'
|
34 |
-
)
|
35 |
-
|
36 |
-
# ==================== FUNZIONI ====================
|
37 |
-
|
38 |
-
def encode_image(image_path: Path) -> str:
|
39 |
-
with open(image_path, "rb") as img:
|
40 |
-
return base64.b64encode(img.read()).decode("utf-8")
|
41 |
-
|
42 |
-
def validate_image_paths(row):
|
43 |
-
missing = []
|
44 |
-
for col in ['immagine_1_path', 'immagine_2_path']:
|
45 |
-
if row[col].strip():
|
46 |
-
path = IMAGE_DIR / row[col].strip()
|
47 |
-
if not path.exists():
|
48 |
-
missing.append(path)
|
49 |
-
return missing
|
50 |
-
|
51 |
-
def invia_task(prompt, image_paths):
|
52 |
-
content = [{"type": "text", "text": prompt}]
|
53 |
-
for img_path in image_paths:
|
54 |
-
if img_path.exists():
|
55 |
-
b64 = encode_image(img_path)
|
56 |
-
content.append({
|
57 |
-
"type": "image_url",
|
58 |
-
"image_url": {
|
59 |
-
"url": f"data:image/jpeg;base64,{b64}"
|
60 |
-
}
|
61 |
-
})
|
62 |
-
|
63 |
-
payload = {
|
64 |
-
"model": MODEL_NAME,
|
65 |
-
"messages": [
|
66 |
-
{
|
67 |
-
"role": "system",
|
68 |
-
"content": (
|
69 |
-
"Sei uno storico dell’arte specializzato in analisi iconografiche e storico-stilistiche. "
|
70 |
-
"Rispondi sempre in italiano, in stile accademico, formale e neutrale. "
|
71 |
-
"Analizza secondo i criteri storico-artistici e rispondi in modo rigoroso e preciso."
|
72 |
-
)
|
73 |
-
},
|
74 |
-
{"role": "user", "content": content}
|
75 |
-
],
|
76 |
-
"max_tokens": 1000
|
77 |
-
}
|
78 |
-
|
79 |
-
response = requests.post(
|
80 |
-
"https://api.openai.com/v1/chat/completions",
|
81 |
-
headers={"Authorization": f"Bearer {API_KEY}", "Content-Type": "application/json"},
|
82 |
-
json=payload
|
83 |
-
)
|
84 |
-
|
85 |
-
if response.status_code == 200:
|
86 |
-
return response.json()["choices"][0]["message"]["content"]
|
87 |
-
else:
|
88 |
-
raise Exception(f"Errore API: {response.status_code} - {response.text}")
|
89 |
-
|
90 |
-
# ==================== MAIN ====================
|
91 |
-
|
92 |
-
def main():
|
93 |
-
if len(sys.argv) < 2:
|
94 |
-
print("Uso: python test_01.py AR CR ...")
|
95 |
-
sys.exit(1)
|
96 |
-
|
97 |
-
categorie_input = sys.argv[1:]
|
98 |
-
categorie_mapped = {CATEGORIE[c] for c in categorie_input if c in CATEGORIE}
|
99 |
-
|
100 |
-
if not categorie_mapped:
|
101 |
-
print("Nessuna categoria valida fornita.")
|
102 |
-
sys.exit(1)
|
103 |
-
|
104 |
-
results = []
|
105 |
-
|
106 |
-
with open(CSV_PATH, encoding='utf-8') as f:
|
107 |
-
reader = csv.DictReader(f)
|
108 |
-
for idx, row in enumerate(reader, 1):
|
109 |
-
if row["categoria"] not in categorie_mapped:
|
110 |
-
continue
|
111 |
-
|
112 |
-
image_paths = [IMAGE_DIR / row["immagine_1_path"].strip()]
|
113 |
-
if row["immagine_2_path"].strip():
|
114 |
-
image_paths.append(IMAGE_DIR / row["immagine_2_path"].strip())
|
115 |
-
|
116 |
-
missing = validate_image_paths(row)
|
117 |
-
if missing:
|
118 |
-
msg = f"[TASK {idx}] Immagini mancanti: {[str(m) for m in missing]}"
|
119 |
-
logging.warning(msg)
|
120 |
-
print(msg)
|
121 |
-
continue
|
122 |
-
|
123 |
-
# Costruzione del prompt
|
124 |
-
base_prompt = row["prompt"].strip() + " " + row["instructions"].strip()
|
125 |
-
opzioni = [row.get("opzione_1", ""), row.get("opzione_2", ""), row.get("opzione_3", "")]
|
126 |
-
opzioni = [opt.strip() for opt in opzioni if opt.strip()]
|
127 |
-
if opzioni:
|
128 |
-
base_prompt += "\n\nOpzioni disponibili:\n" + "\n".join(f"- {opt}" for opt in opzioni)
|
129 |
-
|
130 |
-
try:
|
131 |
-
result = invia_task(base_prompt, image_paths)
|
132 |
-
logging.info(f"[TASK {idx}] Completato con successo.")
|
133 |
-
results.append({
|
134 |
-
"task_id": idx,
|
135 |
-
"risposta": result,
|
136 |
-
"prompt_inviato": base_prompt
|
137 |
-
})
|
138 |
-
except Exception as e:
|
139 |
-
logging.error(f"[TASK {idx}] Errore durante l'invio: {e}")
|
140 |
-
print(f"[TASK {idx}] Errore - vedi log")
|
141 |
-
|
142 |
-
# Scrittura risultati su CSV
|
143 |
-
if results:
|
144 |
-
with open(OUTPUT_CSV, mode='w', encoding='utf-8', newline='') as out_csv:
|
145 |
-
writer = csv.DictWriter(out_csv, fieldnames=["task_id", "risposta", "prompt_inviato"])
|
146 |
-
writer.writeheader()
|
147 |
-
for r in results:
|
148 |
-
writer.writerow(r)
|
149 |
-
|
150 |
-
print(f"Esecuzione completata. Risposte salvate in {OUTPUT_CSV}.")
|
151 |
-
|
152 |
-
if __name__ == "__main__":
|
153 |
-
main()
|
154 |
-
|
|
|
1 |
+
import csv
|
2 |
+
import os
|
3 |
+
import sys
|
4 |
+
import base64
|
5 |
+
import requests
|
6 |
+
import logging
|
7 |
+
from pathlib import Path
|
8 |
+
|
9 |
+
# ==================== CONFIGURAZIONE ====================
|
10 |
+
|
11 |
+
API_KEY = "ChiaveAPI"
|
12 |
+
MODEL_NAME = "gpt-4o"
|
13 |
+
IMAGE_DIR = Path("images")
|
14 |
+
CSV_PATH = Path("ArtVision-0725.csv")
|
15 |
+
OUTPUT_CSV = Path("risposte_output.csv")
|
16 |
+
|
17 |
+
CATEGORIE = {
|
18 |
+
"AR": "art_recognition",
|
19 |
+
"CR": "chronological_reasoning",
|
20 |
+
"CS": "contextual_summary",
|
21 |
+
"VR": "vision_reading",
|
22 |
+
"VB": "vision_basic",
|
23 |
+
"VL": "vision_logic",
|
24 |
+
"VRS": "vision_reasoning",
|
25 |
+
"IG" : "img_gen"
|
26 |
+
}
|
27 |
+
|
28 |
+
# ==================== LOGGING ====================
|
29 |
+
|
30 |
+
logging.basicConfig(
|
31 |
+
filename='task_log.txt',
|
32 |
+
level=logging.INFO,
|
33 |
+
format='%(asctime)s - %(levelname)s - %(message)s'
|
34 |
+
)
|
35 |
+
|
36 |
+
# ==================== FUNZIONI ====================
|
37 |
+
|
38 |
+
def encode_image(image_path: Path) -> str:
|
39 |
+
with open(image_path, "rb") as img:
|
40 |
+
return base64.b64encode(img.read()).decode("utf-8")
|
41 |
+
|
42 |
+
def validate_image_paths(row):
|
43 |
+
missing = []
|
44 |
+
for col in ['immagine_1_path', 'immagine_2_path']:
|
45 |
+
if row[col].strip():
|
46 |
+
path = IMAGE_DIR / row[col].strip()
|
47 |
+
if not path.exists():
|
48 |
+
missing.append(path)
|
49 |
+
return missing
|
50 |
+
|
51 |
+
def invia_task(prompt, image_paths):
|
52 |
+
content = [{"type": "text", "text": prompt}]
|
53 |
+
for img_path in image_paths:
|
54 |
+
if img_path.exists():
|
55 |
+
b64 = encode_image(img_path)
|
56 |
+
content.append({
|
57 |
+
"type": "image_url",
|
58 |
+
"image_url": {
|
59 |
+
"url": f"data:image/jpeg;base64,{b64}"
|
60 |
+
}
|
61 |
+
})
|
62 |
+
|
63 |
+
payload = {
|
64 |
+
"model": MODEL_NAME,
|
65 |
+
"messages": [
|
66 |
+
{
|
67 |
+
"role": "system",
|
68 |
+
"content": (
|
69 |
+
"Sei uno storico dell’arte specializzato in analisi iconografiche e storico-stilistiche. "
|
70 |
+
"Rispondi sempre in italiano, in stile accademico, formale e neutrale. "
|
71 |
+
"Analizza secondo i criteri storico-artistici e rispondi in modo rigoroso e preciso."
|
72 |
+
)
|
73 |
+
},
|
74 |
+
{"role": "user", "content": content}
|
75 |
+
],
|
76 |
+
"max_tokens": 1000
|
77 |
+
}
|
78 |
+
|
79 |
+
response = requests.post(
|
80 |
+
"https://api.openai.com/v1/chat/completions",
|
81 |
+
headers={"Authorization": f"Bearer {API_KEY}", "Content-Type": "application/json"},
|
82 |
+
json=payload
|
83 |
+
)
|
84 |
+
|
85 |
+
if response.status_code == 200:
|
86 |
+
return response.json()["choices"][0]["message"]["content"]
|
87 |
+
else:
|
88 |
+
raise Exception(f"Errore API: {response.status_code} - {response.text}")
|
89 |
+
|
90 |
+
# ==================== MAIN ====================
|
91 |
+
|
92 |
+
def main():
|
93 |
+
if len(sys.argv) < 2:
|
94 |
+
print("Uso: python test_01.py AR CR ...")
|
95 |
+
sys.exit(1)
|
96 |
+
|
97 |
+
categorie_input = sys.argv[1:]
|
98 |
+
categorie_mapped = {CATEGORIE[c] for c in categorie_input if c in CATEGORIE}
|
99 |
+
|
100 |
+
if not categorie_mapped:
|
101 |
+
print("Nessuna categoria valida fornita.")
|
102 |
+
sys.exit(1)
|
103 |
+
|
104 |
+
results = []
|
105 |
+
|
106 |
+
with open(CSV_PATH, encoding='utf-8') as f:
|
107 |
+
reader = csv.DictReader(f)
|
108 |
+
for idx, row in enumerate(reader, 1):
|
109 |
+
if row["categoria"] not in categorie_mapped:
|
110 |
+
continue
|
111 |
+
|
112 |
+
image_paths = [IMAGE_DIR / row["immagine_1_path"].strip()]
|
113 |
+
if row["immagine_2_path"].strip():
|
114 |
+
image_paths.append(IMAGE_DIR / row["immagine_2_path"].strip())
|
115 |
+
|
116 |
+
missing = validate_image_paths(row)
|
117 |
+
if missing:
|
118 |
+
msg = f"[TASK {idx}] Immagini mancanti: {[str(m) for m in missing]}"
|
119 |
+
logging.warning(msg)
|
120 |
+
print(msg)
|
121 |
+
continue
|
122 |
+
|
123 |
+
# Costruzione del prompt
|
124 |
+
base_prompt = row["prompt"].strip() + " " + row["instructions"].strip()
|
125 |
+
opzioni = [row.get("opzione_1", ""), row.get("opzione_2", ""), row.get("opzione_3", "")]
|
126 |
+
opzioni = [opt.strip() for opt in opzioni if opt.strip()]
|
127 |
+
if opzioni:
|
128 |
+
base_prompt += "\n\nOpzioni disponibili:\n" + "\n".join(f"- {opt}" for opt in opzioni)
|
129 |
+
|
130 |
+
try:
|
131 |
+
result = invia_task(base_prompt, image_paths)
|
132 |
+
logging.info(f"[TASK {idx}] Completato con successo.")
|
133 |
+
results.append({
|
134 |
+
"task_id": idx,
|
135 |
+
"risposta": result,
|
136 |
+
"prompt_inviato": base_prompt
|
137 |
+
})
|
138 |
+
except Exception as e:
|
139 |
+
logging.error(f"[TASK {idx}] Errore durante l'invio: {e}")
|
140 |
+
print(f"[TASK {idx}] Errore - vedi log")
|
141 |
+
|
142 |
+
# Scrittura risultati su CSV
|
143 |
+
if results:
|
144 |
+
with open(OUTPUT_CSV, mode='w', encoding='utf-8', newline='') as out_csv:
|
145 |
+
writer = csv.DictWriter(out_csv, fieldnames=["task_id", "risposta", "prompt_inviato"])
|
146 |
+
writer.writeheader()
|
147 |
+
for r in results:
|
148 |
+
writer.writerow(r)
|
149 |
+
|
150 |
+
print(f"Esecuzione completata. Risposte salvate in {OUTPUT_CSV}.")
|
151 |
+
|
152 |
+
if __name__ == "__main__":
|
153 |
+
main()
|
154 |
+
|