Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
|
@@ -319,28 +319,34 @@ def get_korea_models():
|
|
| 319 |
print(f"Error fetching Korea models: {str(e)}")
|
| 320 |
return []
|
| 321 |
|
| 322 |
-
def get_all_models(limit=
|
| 323 |
"""모든 모델과 Korea 관련 모델 가져오기"""
|
| 324 |
all_models = []
|
|
|
|
| 325 |
|
| 326 |
-
#
|
| 327 |
-
|
| 328 |
-
|
| 329 |
-
|
| 330 |
-
|
| 331 |
-
|
| 332 |
-
|
| 333 |
-
|
| 334 |
-
|
| 335 |
-
|
| 336 |
-
|
| 337 |
-
|
| 338 |
-
|
| 339 |
-
|
| 340 |
-
|
| 341 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 342 |
|
| 343 |
-
#
|
| 344 |
korea_params = {
|
| 345 |
"search": "korea",
|
| 346 |
"full": "True",
|
|
@@ -365,30 +371,6 @@ def get_all_models(limit=1000):
|
|
| 365 |
all_models.append(model)
|
| 366 |
existing_ids.add(model.get('id', ''))
|
| 367 |
|
| 368 |
-
# 3. Korean 검색 결과 가져오기
|
| 369 |
-
korean_params = {
|
| 370 |
-
"search": "korean",
|
| 371 |
-
"full": "True",
|
| 372 |
-
"config": "True",
|
| 373 |
-
"limit": limit
|
| 374 |
-
}
|
| 375 |
-
|
| 376 |
-
korean_response = requests.get(
|
| 377 |
-
"https://huggingface.co/api/models",
|
| 378 |
-
headers={'Accept': 'application/json'},
|
| 379 |
-
params=korean_params
|
| 380 |
-
)
|
| 381 |
-
|
| 382 |
-
if korean_response.status_code == 200:
|
| 383 |
-
korean_models = korean_response.json()
|
| 384 |
-
print(f"Fetched {len(korean_models)} Korean-related models")
|
| 385 |
-
|
| 386 |
-
# 중복 제거하면서 Korean 모델 추가
|
| 387 |
-
for model in korean_models:
|
| 388 |
-
if model.get('id', '') not in existing_ids:
|
| 389 |
-
all_models.append(model)
|
| 390 |
-
existing_ids.add(model.get('id', ''))
|
| 391 |
-
|
| 392 |
print(f"Total unique models: {len(all_models)}")
|
| 393 |
return all_models[:limit]
|
| 394 |
|
|
@@ -407,8 +389,8 @@ def get_models_data(progress=gr.Progress()):
|
|
| 407 |
empty_df = pd.DataFrame(columns=['Global Rank', 'Model ID', 'Title', 'Downloads', 'Likes', 'Korea Search', 'URL'])
|
| 408 |
return fig, error_html, empty_df
|
| 409 |
|
| 410 |
-
# 일반 모델과 Korea 관련 모델 모두 가져오기
|
| 411 |
-
all_global_models = get_all_models(limit=
|
| 412 |
korea_models = get_korea_models()
|
| 413 |
|
| 414 |
# 모든 모델 통합 (중복 제거)
|
|
@@ -439,7 +421,7 @@ def get_models_data(progress=gr.Progress()):
|
|
| 439 |
filtered_models.append({
|
| 440 |
'id': model_id,
|
| 441 |
'global_rank': next((idx for idx, m in enumerate(all_models, 1)
|
| 442 |
-
if m.get('id', '').strip() == model_id.strip()), 'Not in top
|
| 443 |
'downloads': model_data.get('downloads', 0),
|
| 444 |
'likes': model_data.get('likes', 0),
|
| 445 |
'title': model_data.get('title', 'No Title'),
|
|
@@ -448,7 +430,7 @@ def get_models_data(progress=gr.Progress()):
|
|
| 448 |
else:
|
| 449 |
filtered_models.append({
|
| 450 |
'id': model_id,
|
| 451 |
-
'global_rank': 'Not in top
|
| 452 |
'downloads': 0,
|
| 453 |
'likes': 0,
|
| 454 |
'title': 'No Title',
|
|
@@ -470,7 +452,7 @@ def get_models_data(progress=gr.Progress()):
|
|
| 470 |
|
| 471 |
fig.add_trace(go.Bar(
|
| 472 |
x=ids,
|
| 473 |
-
y=[
|
| 474 |
text=[f"Rank: #{r}<br>Downloads: {format(m['downloads'], ',')}<br>Likes: {format(m['likes'], ',')}"
|
| 475 |
for r, m in zip(ranks, valid_models)],
|
| 476 |
textposition='auto',
|
|
@@ -480,27 +462,29 @@ def get_models_data(progress=gr.Progress()):
|
|
| 480 |
))
|
| 481 |
|
| 482 |
fig.update_layout(
|
| 483 |
-
title="HuggingFace Models Global Rankings",
|
| 484 |
xaxis_title="Model ID",
|
| 485 |
yaxis_title="Global Rank",
|
| 486 |
yaxis=dict(
|
| 487 |
-
ticktext=[f"#{i}" for i in range(1,
|
| 488 |
-
tickvals=[
|
| 489 |
-
range=[0,
|
| 490 |
),
|
| 491 |
height=800,
|
| 492 |
-
showlegend=False
|
|
|
|
|
|
|
| 493 |
)
|
| 494 |
|
| 495 |
# HTML 카드 생성
|
| 496 |
html_content = """
|
| 497 |
<div style='padding: 20px; background: #f5f5f5;'>
|
| 498 |
-
<h2 style='color: #2c3e50;'>Models Rankings</h2>
|
| 499 |
<div style='display: grid; grid-template-columns: repeat(auto-fill, minmax(300px, 1fr)); gap: 20px;'>
|
| 500 |
"""
|
| 501 |
|
| 502 |
for model in filtered_models:
|
| 503 |
-
rank_display = f"Global Rank #{model['global_rank']}" if isinstance(model['global_rank'], (int, float)) else "Not in top
|
| 504 |
korea_badge = "🇰🇷 Korea Search Result" if model['is_korea'] else ""
|
| 505 |
|
| 506 |
html_content += f"""
|
|
|
|
| 319 |
print(f"Error fetching Korea models: {str(e)}")
|
| 320 |
return []
|
| 321 |
|
| 322 |
+
def get_all_models(limit=3000):
|
| 323 |
"""모든 모델과 Korea 관련 모델 가져오기"""
|
| 324 |
all_models = []
|
| 325 |
+
page_size = 1000 # API의 한 번 요청당 최대 크기
|
| 326 |
|
| 327 |
+
# 여러 페이지에 걸쳐 데이터 수집
|
| 328 |
+
for offset in range(0, limit, page_size):
|
| 329 |
+
params = {
|
| 330 |
+
'limit': min(page_size, limit - offset),
|
| 331 |
+
'full': 'True',
|
| 332 |
+
'config': 'True',
|
| 333 |
+
'offset': offset
|
| 334 |
+
}
|
| 335 |
+
|
| 336 |
+
response = requests.get(
|
| 337 |
+
"https://huggingface.co/api/models",
|
| 338 |
+
headers={'Accept': 'application/json'},
|
| 339 |
+
params=params
|
| 340 |
+
)
|
| 341 |
+
|
| 342 |
+
if response.status_code == 200:
|
| 343 |
+
all_models.extend(response.json())
|
| 344 |
+
print(f"Fetched models {offset+1} to {offset+len(response.json())}")
|
| 345 |
+
else:
|
| 346 |
+
print(f"Failed to fetch models at offset {offset}: {response.status_code}")
|
| 347 |
+
break
|
| 348 |
|
| 349 |
+
# Korea 검색 결과도 동일하게 확장
|
| 350 |
korea_params = {
|
| 351 |
"search": "korea",
|
| 352 |
"full": "True",
|
|
|
|
| 371 |
all_models.append(model)
|
| 372 |
existing_ids.add(model.get('id', ''))
|
| 373 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 374 |
print(f"Total unique models: {len(all_models)}")
|
| 375 |
return all_models[:limit]
|
| 376 |
|
|
|
|
| 389 |
empty_df = pd.DataFrame(columns=['Global Rank', 'Model ID', 'Title', 'Downloads', 'Likes', 'Korea Search', 'URL'])
|
| 390 |
return fig, error_html, empty_df
|
| 391 |
|
| 392 |
+
# 일반 모델과 Korea 관련 모델 모두 가져오기 (3000위까지)
|
| 393 |
+
all_global_models = get_all_models(limit=3000)
|
| 394 |
korea_models = get_korea_models()
|
| 395 |
|
| 396 |
# 모든 모델 통합 (중복 제거)
|
|
|
|
| 421 |
filtered_models.append({
|
| 422 |
'id': model_id,
|
| 423 |
'global_rank': next((idx for idx, m in enumerate(all_models, 1)
|
| 424 |
+
if m.get('id', '').strip() == model_id.strip()), 'Not in top 3000'),
|
| 425 |
'downloads': model_data.get('downloads', 0),
|
| 426 |
'likes': model_data.get('likes', 0),
|
| 427 |
'title': model_data.get('title', 'No Title'),
|
|
|
|
| 430 |
else:
|
| 431 |
filtered_models.append({
|
| 432 |
'id': model_id,
|
| 433 |
+
'global_rank': 'Not in top 3000',
|
| 434 |
'downloads': 0,
|
| 435 |
'likes': 0,
|
| 436 |
'title': 'No Title',
|
|
|
|
| 452 |
|
| 453 |
fig.add_trace(go.Bar(
|
| 454 |
x=ids,
|
| 455 |
+
y=[3001 - r for r in ranks], # Y축 범위 3000까지 확장
|
| 456 |
text=[f"Rank: #{r}<br>Downloads: {format(m['downloads'], ',')}<br>Likes: {format(m['likes'], ',')}"
|
| 457 |
for r, m in zip(ranks, valid_models)],
|
| 458 |
textposition='auto',
|
|
|
|
| 462 |
))
|
| 463 |
|
| 464 |
fig.update_layout(
|
| 465 |
+
title="HuggingFace Models Global Rankings (Up to #3000)",
|
| 466 |
xaxis_title="Model ID",
|
| 467 |
yaxis_title="Global Rank",
|
| 468 |
yaxis=dict(
|
| 469 |
+
ticktext=[f"#{i}" for i in range(1, 3001, 100)],
|
| 470 |
+
tickvals=[3001 - i for i in range(1, 3001, 100)],
|
| 471 |
+
range=[0, 3000]
|
| 472 |
),
|
| 473 |
height=800,
|
| 474 |
+
showlegend=False,
|
| 475 |
+
template='plotly_white',
|
| 476 |
+
xaxis_tickangle=-45
|
| 477 |
)
|
| 478 |
|
| 479 |
# HTML 카드 생성
|
| 480 |
html_content = """
|
| 481 |
<div style='padding: 20px; background: #f5f5f5;'>
|
| 482 |
+
<h2 style='color: #2c3e50;'>Models Rankings (Up to #3000)</h2>
|
| 483 |
<div style='display: grid; grid-template-columns: repeat(auto-fill, minmax(300px, 1fr)); gap: 20px;'>
|
| 484 |
"""
|
| 485 |
|
| 486 |
for model in filtered_models:
|
| 487 |
+
rank_display = f"Global Rank #{model['global_rank']}" if isinstance(model['global_rank'], (int, float)) else "Not in top 3000"
|
| 488 |
korea_badge = "🇰🇷 Korea Search Result" if model['is_korea'] else ""
|
| 489 |
|
| 490 |
html_content += f"""
|