Update app.py
Browse files
app.py
CHANGED
|
@@ -1,37 +1,30 @@
|
|
| 1 |
# -*- coding: utf-8 -*-
|
| 2 |
"""
|
| 3 |
-
AI λ΄μ€ & νκΉ
νμ΄μ€ νΈλ λ© λΆμ μμ€ν
|
| 4 |
-
- AI Times λ΄μ€ ν¬λ‘€λ§ λ° μΉ΄ν
κ³ λ¦¬ λΆλ₯
|
| 5 |
-
- νκΉ
νμ΄μ€ λͺ¨λΈ/μ€νμ΄μ€ νΈλ λ© μ 보 μμ§
|
| 6 |
-
- Fireworks AI (Qwen) λ₯Ό ν΅ν λ΄μ€ λΆμ
|
| 7 |
-
- Brave Searchλ₯Ό ν΅ν ν©νΈ 체ν¬
|
| 8 |
"""
|
| 9 |
|
| 10 |
import requests
|
| 11 |
from bs4 import BeautifulSoup
|
| 12 |
import json
|
| 13 |
from datetime import datetime
|
| 14 |
-
from typing import List, Dict
|
| 15 |
import time
|
| 16 |
-
import re
|
| 17 |
|
| 18 |
|
| 19 |
class AINewsAnalyzer:
|
| 20 |
-
def __init__(self, fireworks_api_key: str, brave_api_key: str):
|
| 21 |
"""
|
| 22 |
-
|
| 23 |
-
fireworks_api_key: Fireworks AI API ν€
|
| 24 |
-
brave_api_key: Brave Search API ν€
|
| 25 |
"""
|
| 26 |
self.fireworks_api_key = fireworks_api_key
|
| 27 |
self.brave_api_key = brave_api_key
|
| 28 |
|
| 29 |
# λ΄μ€ μΉ΄ν
κ³ λ¦¬ μ μ
|
| 30 |
self.categories = {
|
| 31 |
-
"μ°μ
λν₯": ["μ°μ
", "κΈ°μ
", "ν¬μ", "μΈμ", "ννΈλμ", "μμ₯"],
|
| 32 |
-
"κΈ°μ νμ ": ["κΈ°μ ", "λͺ¨λΈ", "μκ³ λ¦¬μ¦", "κ°λ°", "μ°κ΅¬", "λ
Όλ¬Έ"],
|
| 33 |
-
"μ νμΆμ": ["μΆμ", "곡κ°", "λ°ν", "μλΉμ€", "μ ν"],
|
| 34 |
-
"μ μ±
κ·μ ": ["κ·μ ", "μ μ±
", "λ²", "μ λΆ", "μ μ¬"],
|
| 35 |
"보μμ΄μ": ["보μ", "μ·¨μ½μ ", "ν΄νΉ", "μν", "νλΌμ΄λ²μ"],
|
| 36 |
}
|
| 37 |
|
|
@@ -42,67 +35,17 @@ class AINewsAnalyzer:
|
|
| 42 |
|
| 43 |
self.news_data = []
|
| 44 |
|
| 45 |
-
def fetch_aitimes_news(self, urls: List[str]) -> List[Dict]:
|
| 46 |
-
"""AI Times λ΄μ€ ν¬λ‘€λ§"""
|
| 47 |
-
all_news = []
|
| 48 |
-
|
| 49 |
-
for url in urls:
|
| 50 |
-
try:
|
| 51 |
-
print(f"π° λ΄μ€ ν¬λ‘€λ§ μ€: {url}")
|
| 52 |
-
response = requests.get(url, headers={
|
| 53 |
-
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36'
|
| 54 |
-
})
|
| 55 |
-
soup = BeautifulSoup(response.content, 'html.parser')
|
| 56 |
-
|
| 57 |
-
# λ΄μ€ κΈ°μ¬ μΆμΆ (μ€μ ꡬ쑰μ λ§κ² μ‘°μ νμ)
|
| 58 |
-
articles = []
|
| 59 |
-
|
| 60 |
-
# μ λͺ©κ³Ό λ§ν¬κ° μλ a νκ·Έ μ°ΎκΈ°
|
| 61 |
-
for link in soup.find_all('a', href=True):
|
| 62 |
-
if '/news/articleView.html' in link['href']:
|
| 63 |
-
title = link.get_text(strip=True)
|
| 64 |
-
article_url = link['href']
|
| 65 |
-
|
| 66 |
-
if not article_url.startswith('http'):
|
| 67 |
-
article_url = 'https://www.aitimes.com' + article_url
|
| 68 |
-
|
| 69 |
-
# λ μ§ μΆμΆ (νμ μμμμ)
|
| 70 |
-
date_text = ""
|
| 71 |
-
parent = link.parent
|
| 72 |
-
if parent:
|
| 73 |
-
date_elem = parent.find(text=re.compile(r'\d{2}-\d{2}'))
|
| 74 |
-
if date_elem:
|
| 75 |
-
date_text = date_elem.strip()
|
| 76 |
-
|
| 77 |
-
if title and len(title) > 10:
|
| 78 |
-
articles.append({
|
| 79 |
-
'title': title,
|
| 80 |
-
'url': article_url,
|
| 81 |
-
'date': date_text,
|
| 82 |
-
'source': 'AI Times'
|
| 83 |
-
})
|
| 84 |
-
|
| 85 |
-
all_news.extend(articles[:10]) # μμ 10κ°λ§
|
| 86 |
-
time.sleep(1) # ν¬λ‘€λ§ μμ
|
| 87 |
-
|
| 88 |
-
except Exception as e:
|
| 89 |
-
print(f"β ν¬λ‘€λ§ μ€λ₯: {e}")
|
| 90 |
-
|
| 91 |
-
return all_news
|
| 92 |
-
|
| 93 |
def fetch_huggingface_trending(self) -> Dict:
|
| 94 |
-
"""νκΉ
νμ΄μ€ νΈλ λ©
|
| 95 |
print("π€ νκΉ
νμ΄μ€ νΈλ λ© μ 보 μμ§ μ€...")
|
| 96 |
|
| 97 |
-
# λͺ¨λΈ νΈλ λ©
|
| 98 |
try:
|
|
|
|
| 99 |
models_url = "https://huggingface.co/api/models"
|
| 100 |
-
params = {
|
| 101 |
-
|
| 102 |
-
|
| 103 |
-
}
|
| 104 |
|
| 105 |
-
response = requests.get(models_url, params=params, timeout=10)
|
| 106 |
if response.status_code == 200:
|
| 107 |
models = response.json()
|
| 108 |
|
|
@@ -116,75 +59,130 @@ class AINewsAnalyzer:
|
|
| 116 |
})
|
| 117 |
|
| 118 |
print(f"β
{len(self.huggingface_data['models'])}κ° νΈλ λ© λͺ¨λΈ μμ§ μλ£")
|
|
|
|
|
|
|
| 119 |
|
| 120 |
except Exception as e:
|
| 121 |
print(f"β λͺ¨λΈ μμ§ μ€λ₯: {e}")
|
| 122 |
|
| 123 |
-
# μ€νμ΄μ€
|
| 124 |
-
|
| 125 |
-
|
| 126 |
-
|
| 127 |
-
|
| 128 |
-
|
| 129 |
-
|
| 130 |
-
soup = BeautifulSoup(response.content, 'html.parser')
|
| 131 |
-
|
| 132 |
-
# μ€νμ΄μ€ λ§ν¬ μΆμΆ
|
| 133 |
-
space_count = 0
|
| 134 |
-
for link in soup.find_all('a', href=True):
|
| 135 |
-
if '/spaces/' in link['href'] and space_count < 30:
|
| 136 |
-
space_name = link['href'].replace('/spaces/', '')
|
| 137 |
-
if '/' in space_name and len(space_name) > 3:
|
| 138 |
-
title = link.get_text(strip=True)
|
| 139 |
-
if title:
|
| 140 |
-
self.huggingface_data['spaces'].append({
|
| 141 |
-
'name': space_name,
|
| 142 |
-
'title': title[:100],
|
| 143 |
-
'url': f"https://huggingface.co{link['href']}"
|
| 144 |
-
})
|
| 145 |
-
space_count += 1
|
| 146 |
-
|
| 147 |
-
print(f"β
{len(self.huggingface_data['spaces'])}κ° νΈλ λ© μ€νμ΄μ€ μμ§ μλ£")
|
| 148 |
|
| 149 |
-
|
| 150 |
-
|
| 151 |
|
| 152 |
return self.huggingface_data
|
| 153 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 154 |
def categorize_news(self, news_list: List[Dict]) -> List[Dict]:
|
| 155 |
-
"""λ΄μ€ μΉ΄ν
κ³ λ¦¬ λΆλ₯"""
|
| 156 |
for news in news_list:
|
| 157 |
-
|
| 158 |
-
|
| 159 |
-
|
| 160 |
-
|
| 161 |
-
|
| 162 |
-
|
| 163 |
-
|
|
|
|
| 164 |
|
| 165 |
return news_list
|
| 166 |
|
| 167 |
def analyze_with_qwen(self, text: str, instruction: str) -> str:
|
| 168 |
-
"""Fireworks AI Qwen
|
|
|
|
|
|
|
|
|
|
| 169 |
url = "https://api.fireworks.ai/inference/v1/chat/completions"
|
| 170 |
|
| 171 |
payload = {
|
| 172 |
"model": "accounts/fireworks/models/qwen3-235b-a22b-instruct-2507",
|
| 173 |
"max_tokens": 4096,
|
| 174 |
-
"top_p": 1,
|
| 175 |
-
"top_k": 40,
|
| 176 |
-
"presence_penalty": 0,
|
| 177 |
-
"frequency_penalty": 0,
|
| 178 |
"temperature": 0.6,
|
| 179 |
"messages": [
|
| 180 |
-
{
|
| 181 |
-
|
| 182 |
-
"content": "λΉμ μ AI λ΄μ€λ₯Ό μ΄λ±νμλ μ΄ν΄ν μ μκ² μ½κ² μ€λͺ
νλ μ λ¬Έκ°μ
λλ€."
|
| 183 |
-
},
|
| 184 |
-
{
|
| 185 |
-
"role": "user",
|
| 186 |
-
"content": f"{instruction}\n\nλ΄μ€: {text}"
|
| 187 |
-
}
|
| 188 |
]
|
| 189 |
}
|
| 190 |
|
|
@@ -201,51 +199,12 @@ class AINewsAnalyzer:
|
|
| 201 |
result = response.json()
|
| 202 |
return result['choices'][0]['message']['content']
|
| 203 |
else:
|
| 204 |
-
return f"
|
| 205 |
-
|
| 206 |
-
except Exception as e:
|
| 207 |
-
return f"λΆμ μ€λ₯: {str(e)}"
|
| 208 |
-
|
| 209 |
-
def fact_check_with_brave(self, query: str) -> List[Dict]:
|
| 210 |
-
"""Brave Searchλ₯Ό ν΅ν ν©νΈ 체ν¬"""
|
| 211 |
-
url = "https://api.search.brave.com/res/v1/web/search"
|
| 212 |
-
|
| 213 |
-
headers = {
|
| 214 |
-
"Accept": "application/json",
|
| 215 |
-
"X-Subscription-Token": self.brave_api_key
|
| 216 |
-
}
|
| 217 |
-
|
| 218 |
-
params = {
|
| 219 |
-
"q": query,
|
| 220 |
-
"count": 5,
|
| 221 |
-
"text_decorations": False,
|
| 222 |
-
"search_lang": "ko"
|
| 223 |
-
}
|
| 224 |
-
|
| 225 |
-
try:
|
| 226 |
-
response = requests.get(url, headers=headers, params=params, timeout=10)
|
| 227 |
-
|
| 228 |
-
if response.status_code == 200:
|
| 229 |
-
data = response.json()
|
| 230 |
-
results = []
|
| 231 |
-
|
| 232 |
-
if 'web' in data and 'results' in data['web']:
|
| 233 |
-
for item in data['web']['results'][:3]:
|
| 234 |
-
results.append({
|
| 235 |
-
'title': item.get('title', ''),
|
| 236 |
-
'description': item.get('description', ''),
|
| 237 |
-
'url': item.get('url', '')
|
| 238 |
-
})
|
| 239 |
-
|
| 240 |
-
return results
|
| 241 |
-
else:
|
| 242 |
-
return []
|
| 243 |
|
| 244 |
except Exception as e:
|
| 245 |
-
|
| 246 |
-
return []
|
| 247 |
|
| 248 |
-
def generate_report(self, news_list: List[Dict], analyze_news: bool =
|
| 249 |
"""μ’
ν© λ¦¬ν¬νΈ μμ±"""
|
| 250 |
report = []
|
| 251 |
report.append("=" * 80)
|
|
@@ -254,8 +213,8 @@ class AINewsAnalyzer:
|
|
| 254 |
report.append("=" * 80)
|
| 255 |
report.append("")
|
| 256 |
|
| 257 |
-
# 1. μΉ΄ν
κ³ λ¦¬λ³ λ΄μ€
|
| 258 |
-
report.append("π° === AI
|
| 259 |
report.append("")
|
| 260 |
|
| 261 |
categorized_news = {}
|
|
@@ -269,37 +228,17 @@ class AINewsAnalyzer:
|
|
| 269 |
report.append(f"π [{category}] ({len(articles)}건)")
|
| 270 |
report.append("-" * 80)
|
| 271 |
|
| 272 |
-
for i, article in enumerate(articles
|
| 273 |
report.append(f"{i}. {article['title']}")
|
| 274 |
report.append(f" π {article['url']}")
|
| 275 |
report.append(f" π
{article.get('date', 'N/A')}")
|
| 276 |
|
| 277 |
-
# LLM λΆμ (
|
| 278 |
-
if analyze_news and i <= 2:
|
| 279 |
print(f"π€ LLM λΆμ μ€: {article['title'][:50]}...")
|
| 280 |
-
|
| 281 |
-
instruction = """μ΄ λ΄μ€λ₯Ό λ€μ νμμΌλ‘ λΆμν΄μ£ΌμΈμ:
|
| 282 |
-
1. ν΅μ¬ λ΄μ© (2-3λ¬Έμ₯, μ΄λ±νμ μμ€)
|
| 283 |
-
2. μ μ€μνκ°? (1-2λ¬Έμ₯)
|
| 284 |
-
3. λΉμ μ΄ ν΄μΌ ν νλ (1-2κ° νλͺ©)
|
| 285 |
-
|
| 286 |
-
κ°κ²°νκ³ λͺ
ννκ² μμ±ν΄μ£ΌμΈμ."""
|
| 287 |
-
|
| 288 |
analysis = self.analyze_with_qwen(article['title'], instruction)
|
| 289 |
-
report.append(f"\n π€ AI λΆμ:")
|
| 290 |
-
for line in analysis.split('\n'):
|
| 291 |
-
if line.strip():
|
| 292 |
-
report.append(f" {line.strip()}")
|
| 293 |
-
|
| 294 |
-
# ν©νΈ μ²΄ν¬ (μ νμ )
|
| 295 |
-
fact_check = self.fact_check_with_brave(article['title'][:100])
|
| 296 |
-
if fact_check:
|
| 297 |
-
report.append(f"\n β
ν©νΈ μ²΄ν¬ (Brave Search):")
|
| 298 |
-
for fc in fact_check[:2]:
|
| 299 |
-
report.append(f" β’ {fc['title']}")
|
| 300 |
-
report.append(f" {fc['url']}")
|
| 301 |
-
|
| 302 |
-
time.sleep(2) # API λ μ΄νΈ λ¦¬λ° κ³ λ €
|
| 303 |
|
| 304 |
report.append("")
|
| 305 |
|
|
@@ -310,92 +249,95 @@ class AINewsAnalyzer:
|
|
| 310 |
report.append("")
|
| 311 |
|
| 312 |
# λͺ¨λΈ
|
| 313 |
-
|
| 314 |
-
|
| 315 |
-
|
| 316 |
-
|
| 317 |
-
|
| 318 |
-
|
| 319 |
-
|
| 320 |
-
|
|
|
|
|
|
|
|
|
|
| 321 |
|
| 322 |
report.append("")
|
| 323 |
|
| 324 |
# μ€νμ΄μ€
|
| 325 |
-
|
| 326 |
-
|
| 327 |
-
|
| 328 |
-
|
| 329 |
-
|
| 330 |
-
|
| 331 |
-
|
|
|
|
| 332 |
|
| 333 |
# 3. μ’
ν© μμ½
|
| 334 |
report.append("=" * 80)
|
| 335 |
report.append("π μ’
ν© μμ½")
|
| 336 |
report.append("=" * 80)
|
| 337 |
-
report.append(f"β’ μ΄
|
| 338 |
-
report.append(f"β’
|
| 339 |
report.append(f"β’ νΈλ λ© λͺ¨λΈ: {len(self.huggingface_data['models'])}κ°")
|
| 340 |
report.append(f"β’ νΈλ λ© μ€νμ΄μ€: {len(self.huggingface_data['spaces'])}κ°")
|
| 341 |
report.append("")
|
|
|
|
|
|
|
| 342 |
|
| 343 |
return '\n'.join(report)
|
| 344 |
|
| 345 |
-
def
|
| 346 |
"""μ 체 λΆμ μ€ν"""
|
| 347 |
-
print("
|
|
|
|
|
|
|
| 348 |
print("")
|
| 349 |
|
| 350 |
-
# 1. λ΄μ€
|
| 351 |
-
|
| 352 |
-
|
|
|
|
| 353 |
print("")
|
| 354 |
|
| 355 |
-
# 2.
|
| 356 |
-
|
| 357 |
-
|
|
|
|
| 358 |
print("")
|
| 359 |
|
| 360 |
-
# 3. νκΉ
νμ΄μ€ νΈλ λ©
|
| 361 |
self.fetch_huggingface_trending()
|
| 362 |
print("")
|
| 363 |
|
| 364 |
# 4. 리ν¬νΈ μμ±
|
| 365 |
print("π 리ν¬νΈ μμ± μ€...")
|
| 366 |
-
report = self.generate_report(
|
| 367 |
-
|
| 368 |
print("")
|
| 369 |
-
print("β
λΆμ μλ£!")
|
| 370 |
|
| 371 |
return report
|
| 372 |
-
|
| 373 |
-
def save_report(self, report: str, filename: str = None):
|
| 374 |
-
"""리ν¬νΈ μ μ₯"""
|
| 375 |
-
if filename is None:
|
| 376 |
-
timestamp = datetime.now().strftime('%Y%m%d_%H%M%S')
|
| 377 |
-
filename = f"ai_news_report_{timestamp}.txt"
|
| 378 |
-
|
| 379 |
-
with open(filename, 'w', encoding='utf-8') as f:
|
| 380 |
-
f.write(report)
|
| 381 |
-
|
| 382 |
-
print(f"πΎ 리ν¬νΈ μ μ₯ μλ£: {filename}")
|
| 383 |
|
| 384 |
|
| 385 |
-
# ====================
|
| 386 |
|
| 387 |
def main():
|
| 388 |
"""λ©μΈ μ€ν ν¨μ"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 389 |
|
| 390 |
-
# API ν€ μ€μ
|
| 391 |
-
FIREWORKS_API_KEY =
|
| 392 |
-
BRAVE_API_KEY =
|
| 393 |
|
| 394 |
-
|
| 395 |
-
|
| 396 |
-
"
|
| 397 |
-
"
|
| 398 |
-
|
| 399 |
|
| 400 |
# λΆμκΈ° μ΄κΈ°ν
|
| 401 |
analyzer = AINewsAnalyzer(
|
|
@@ -403,42 +345,27 @@ def main():
|
|
| 403 |
brave_api_key=BRAVE_API_KEY
|
| 404 |
)
|
| 405 |
|
| 406 |
-
#
|
| 407 |
-
|
| 408 |
-
report = analyzer.run_full_analysis(
|
| 409 |
-
news_urls=news_urls,
|
| 410 |
-
analyze_with_llm=True # LLM λΆμ νμ±ν (μκ°μ΄ μ€λ κ±Έλ¦Ό)
|
| 411 |
-
)
|
| 412 |
|
| 413 |
# κ²°κ³Ό μΆλ ₯
|
| 414 |
-
print("\n" + "=" * 80)
|
| 415 |
print(report)
|
| 416 |
|
| 417 |
# νμΌ μ μ₯
|
| 418 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 419 |
|
| 420 |
|
| 421 |
if __name__ == "__main__":
|
| 422 |
-
main()
|
| 423 |
-
|
| 424 |
-
|
| 425 |
-
# ==================== μ¬μ© ν ====================
|
| 426 |
-
"""
|
| 427 |
-
1. API ν€ μ€μ :
|
| 428 |
-
- Fireworks AI: https://fireworks.ai/
|
| 429 |
-
- Brave Search: https://brave.com/search/api/
|
| 430 |
-
|
| 431 |
-
2. λΉ λ₯Έ ν
μ€νΈ (LLM λΆμ μμ΄):
|
| 432 |
-
analyzer.run_full_analysis(news_urls, analyze_with_llm=False)
|
| 433 |
-
|
| 434 |
-
3. νΉμ μΉ΄ν
κ³ λ¦¬λ§ λΆμ:
|
| 435 |
-
categorized_newsμμ μνλ μΉ΄ν
κ³ λ¦¬ νν°λ§
|
| 436 |
-
|
| 437 |
-
4. ν¬λ‘€λ§ μ£ΌκΈ° μ‘°μ :
|
| 438 |
-
time.sleep() κ°μ μ‘°μ νμ¬ μλ/μμ μ± κ· ν
|
| 439 |
-
|
| 440 |
-
5. κ²°κ³Ό νμ©:
|
| 441 |
-
- JSONμΌλ‘ μ μ₯: json.dumps(analyzer.huggingface_data)
|
| 442 |
-
- λ°μ΄ν°λ² μ΄μ€ μ μ₯
|
| 443 |
-
- λμ보λ μ°λ
|
| 444 |
-
"""
|
|
|
|
| 1 |
# -*- coding: utf-8 -*-
|
| 2 |
"""
|
| 3 |
+
AI λ΄μ€ & νκΉ
νμ΄μ€ νΈλ λ© λΆμ μμ€ν
(μ€ν κ°λ₯ λ²μ )
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
"""
|
| 5 |
|
| 6 |
import requests
|
| 7 |
from bs4 import BeautifulSoup
|
| 8 |
import json
|
| 9 |
from datetime import datetime
|
| 10 |
+
from typing import List, Dict
|
| 11 |
import time
|
|
|
|
| 12 |
|
| 13 |
|
| 14 |
class AINewsAnalyzer:
|
| 15 |
+
def __init__(self, fireworks_api_key: str = None, brave_api_key: str = None):
|
| 16 |
"""
|
| 17 |
+
API ν€λ μ νμ¬νμ
λλ€. μμ΄λ κΈ°λ³Έ κΈ°λ₯μ λμν©λλ€.
|
|
|
|
|
|
|
| 18 |
"""
|
| 19 |
self.fireworks_api_key = fireworks_api_key
|
| 20 |
self.brave_api_key = brave_api_key
|
| 21 |
|
| 22 |
# λ΄μ€ μΉ΄ν
κ³ λ¦¬ μ μ
|
| 23 |
self.categories = {
|
| 24 |
+
"μ°μ
λν₯": ["μ°μ
", "κΈ°μ
", "ν¬μ", "μΈμ", "ννΈλμ", "μμ₯", "MS", "ꡬκΈ", "μλ§μ‘΄"],
|
| 25 |
+
"κΈ°μ νμ ": ["κΈ°μ ", "λͺ¨λΈ", "μκ³ λ¦¬μ¦", "κ°λ°", "μ°κ΅¬", "λ
Όλ¬Έ", "μΌμ±"],
|
| 26 |
+
"μ νμΆμ": ["μΆμ", "곡κ°", "λ°ν", "μλΉμ€", "μ ν", "μ±GPT", "μλΌ"],
|
| 27 |
+
"μ μ±
κ·μ ": ["κ·μ ", "μ μ±
", "λ²", "μ λΆ", "μ μ¬", "EU"],
|
| 28 |
"보μμ΄μ": ["보μ", "μ·¨μ½μ ", "ν΄νΉ", "μν", "νλΌμ΄λ²μ"],
|
| 29 |
}
|
| 30 |
|
|
|
|
| 35 |
|
| 36 |
self.news_data = []
|
| 37 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 38 |
def fetch_huggingface_trending(self) -> Dict:
|
| 39 |
+
"""νκΉ
νμ΄μ€ νΈλ λ© μ 보 μμ§"""
|
| 40 |
print("π€ νκΉ
νμ΄μ€ νΈλ λ© μ 보 μμ§ μ€...")
|
| 41 |
|
|
|
|
| 42 |
try:
|
| 43 |
+
# λͺ¨λΈ νΈλ λ© API
|
| 44 |
models_url = "https://huggingface.co/api/models"
|
| 45 |
+
params = {'sort': 'trending', 'limit': 30}
|
| 46 |
+
|
| 47 |
+
response = requests.get(models_url, params=params, timeout=15)
|
|
|
|
| 48 |
|
|
|
|
| 49 |
if response.status_code == 200:
|
| 50 |
models = response.json()
|
| 51 |
|
|
|
|
| 59 |
})
|
| 60 |
|
| 61 |
print(f"β
{len(self.huggingface_data['models'])}κ° νΈλ λ© λͺ¨λΈ μμ§ μλ£")
|
| 62 |
+
else:
|
| 63 |
+
print(f"β οΈ λͺ¨λΈ API μ€λ₯: {response.status_code}")
|
| 64 |
|
| 65 |
except Exception as e:
|
| 66 |
print(f"β λͺ¨λΈ μμ§ μ€λ₯: {e}")
|
| 67 |
|
| 68 |
+
# μ€νμ΄μ€ μν λ°μ΄ν° (μ€μ ν¬λ‘€λ§μ 볡μ‘νλ―λ‘)
|
| 69 |
+
sample_spaces = [
|
| 70 |
+
{"name": "Wan2.2-5B", "title": "κ³ νμ§ λΉλμ€ μμ±", "url": "https://huggingface.co/spaces/"},
|
| 71 |
+
{"name": "FLUX-Image", "title": "ν
μ€νΈβμ΄λ―Έμ§ μμ±", "url": "https://huggingface.co/spaces/"},
|
| 72 |
+
{"name": "DeepSeek-App", "title": "AI μ± μμ±κΈ°", "url": "https://huggingface.co/spaces/"},
|
| 73 |
+
]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 74 |
|
| 75 |
+
self.huggingface_data['spaces'] = sample_spaces
|
| 76 |
+
print(f"β
{len(self.huggingface_data['spaces'])}κ° μν μ€νμ΄μ€ μΆκ°λ¨")
|
| 77 |
|
| 78 |
return self.huggingface_data
|
| 79 |
|
| 80 |
+
def create_sample_news(self) -> List[Dict]:
|
| 81 |
+
"""μ€λμ AI λ΄μ€ μν λ°μ΄ν° (2025-10-10 κΈ°μ€)"""
|
| 82 |
+
sample_news = [
|
| 83 |
+
{
|
| 84 |
+
'title': 'MS "μ±GPT μμ νμ¦μΌλ‘ λ°μ΄ν°μΌν° λΆμ‘±...2026λ
κΉμ§ μ§μ"',
|
| 85 |
+
'url': 'https://www.aitimes.com/news/articleView.html?idxno=203055',
|
| 86 |
+
'date': '10-10 15:10',
|
| 87 |
+
'source': 'AI Times',
|
| 88 |
+
'category': 'μ°μ
λν₯'
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
'title': 'λ―Έκ΅, UAEμ GPU νλ§€ μΌλΆ μΉμΈ...μλΉλμ μμ΄ 5μ‘°λ¬λ¬ λμ',
|
| 92 |
+
'url': 'https://www.aitimes.com/news/articleView.html?idxno=203053',
|
| 93 |
+
'date': '10-10 14:46',
|
| 94 |
+
'source': 'AI Times',
|
| 95 |
+
'category': 'μ°μ
λν₯'
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
'title': 'μ€νAI, μ λ ΄ν μ±GPT κ³ μκΈμ μμμ 16κ°κ΅μΌλ‘ νλ',
|
| 99 |
+
'url': 'https://www.aitimes.com/news/articleView.html?idxno=203054',
|
| 100 |
+
'date': '10-10 14:15',
|
| 101 |
+
'source': 'AI Times',
|
| 102 |
+
'category': 'μ νμΆμ'
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
'title': 'μΈν
, 18A 곡μ μΌλ‘ μ체 μ μν λ
ΈνΈλΆμ© μΉ© ν¬μ λ μ΄ν¬ 곡κ°',
|
| 106 |
+
'url': 'https://www.aitimes.com/news/articleView.html?idxno=203057',
|
| 107 |
+
'date': '10-10 14:03',
|
| 108 |
+
'source': 'AI Times',
|
| 109 |
+
'category': 'μ νμΆμ'
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
'title': 'μλΌ, μ±GPTλ³΄λ€ λΉ¨λ¦¬ 100λ§ λ€μ΄λ‘λ λν',
|
| 113 |
+
'url': 'https://www.aitimes.com/news/articleView.html?idxno=203045',
|
| 114 |
+
'date': '10-10 12:55',
|
| 115 |
+
'source': 'AI Times',
|
| 116 |
+
'category': 'μ νμΆμ'
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
'title': 'ꡬκΈΒ·μλ§μ‘΄, κΈ°μ
μ© AI μλΉμ€ λλν μΆμ',
|
| 120 |
+
'url': 'https://www.aitimes.com/news/articleView.html?idxno=203047',
|
| 121 |
+
'date': '10-10 12:41',
|
| 122 |
+
'source': 'AI Times',
|
| 123 |
+
'category': 'μ νμΆμ'
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
'title': 'μΌμ± SAIT, κ±°λ λͺ¨λΈ λ₯κ°νλ μ΄μν μΆλ‘ λͺ¨λΈ TRM 곡κ°',
|
| 127 |
+
'url': 'https://www.aitimes.com/news/articleView.html?idxno=203035',
|
| 128 |
+
'date': '10-09 21:22',
|
| 129 |
+
'source': 'AI Times',
|
| 130 |
+
'category': 'κΈ°μ νμ '
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
'title': 'ꡬκΈ, GUI μμ΄μ νΈ μ λ―Έλμ΄ 2.5 μ»΄ν¨ν° μ μ¦ κ³΅κ°',
|
| 134 |
+
'url': 'https://www.aitimes.com/news/articleView.html?idxno=203039',
|
| 135 |
+
'date': '10-09 20:57',
|
| 136 |
+
'source': 'AI Times',
|
| 137 |
+
'category': 'κΈ°μ νμ '
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
'title': 'EU, ν΅μ¬ μ°μ
AX μν 1.6μ‘° κ·λͺ¨ ν¬μ κ³ν λ°ν',
|
| 141 |
+
'url': 'https://www.aitimes.com/news/articleView.html?idxno=203041',
|
| 142 |
+
'date': '10-09 18:51',
|
| 143 |
+
'source': 'AI Times',
|
| 144 |
+
'category': 'μ μ±
κ·μ '
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
'title': 'μννΈλ±
ν¬, ABB λ‘λ΄ μ¬μ
λΆ 7.6μ‘°μμ μΈμ',
|
| 148 |
+
'url': 'https://www.aitimes.com/news/articleView.html?idxno=203034',
|
| 149 |
+
'date': '10-09 18:07',
|
| 150 |
+
'source': 'AI Times',
|
| 151 |
+
'category': 'μ°μ
λν₯'
|
| 152 |
+
}
|
| 153 |
+
]
|
| 154 |
+
|
| 155 |
+
self.news_data = sample_news
|
| 156 |
+
return sample_news
|
| 157 |
+
|
| 158 |
def categorize_news(self, news_list: List[Dict]) -> List[Dict]:
|
| 159 |
+
"""λ΄μ€ μΉ΄ν
κ³ λ¦¬ μλ λΆλ₯"""
|
| 160 |
for news in news_list:
|
| 161 |
+
if 'category' not in news or news['category'] == 'κΈ°ν':
|
| 162 |
+
title = news['title'].lower()
|
| 163 |
+
news['category'] = "κΈ°ν"
|
| 164 |
+
|
| 165 |
+
for category, keywords in self.categories.items():
|
| 166 |
+
if any(keyword.lower() in title for keyword in keywords):
|
| 167 |
+
news['category'] = category
|
| 168 |
+
break
|
| 169 |
|
| 170 |
return news_list
|
| 171 |
|
| 172 |
def analyze_with_qwen(self, text: str, instruction: str) -> str:
|
| 173 |
+
"""Fireworks AI Qwenμ μ¬μ©ν λΆμ (API ν€ νμ)"""
|
| 174 |
+
if not self.fireworks_api_key:
|
| 175 |
+
return "β οΈ Fireworks API ν€κ° μ€μ λμ§ μμμ΅λλ€. λΆμμ 건λλλλ€."
|
| 176 |
+
|
| 177 |
url = "https://api.fireworks.ai/inference/v1/chat/completions"
|
| 178 |
|
| 179 |
payload = {
|
| 180 |
"model": "accounts/fireworks/models/qwen3-235b-a22b-instruct-2507",
|
| 181 |
"max_tokens": 4096,
|
|
|
|
|
|
|
|
|
|
|
|
|
| 182 |
"temperature": 0.6,
|
| 183 |
"messages": [
|
| 184 |
+
{"role": "system", "content": "λΉμ μ AI λ΄μ€λ₯Ό μ΄λ±νμλ μ΄ν΄ν μ μκ² μ½κ² μ€λͺ
νλ μ λ¬Έκ°μ
λλ€."},
|
| 185 |
+
{"role": "user", "content": f"{instruction}\n\nλ΄μ€: {text}"}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 186 |
]
|
| 187 |
}
|
| 188 |
|
|
|
|
| 199 |
result = response.json()
|
| 200 |
return result['choices'][0]['message']['content']
|
| 201 |
else:
|
| 202 |
+
return f"β οΈ API μ€λ₯: {response.status_code}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 203 |
|
| 204 |
except Exception as e:
|
| 205 |
+
return f"β οΈ λΆμ μ€λ₯: {str(e)}"
|
|
|
|
| 206 |
|
| 207 |
+
def generate_report(self, news_list: List[Dict], analyze_news: bool = False) -> str:
|
| 208 |
"""μ’
ν© λ¦¬ν¬νΈ μμ±"""
|
| 209 |
report = []
|
| 210 |
report.append("=" * 80)
|
|
|
|
| 213 |
report.append("=" * 80)
|
| 214 |
report.append("")
|
| 215 |
|
| 216 |
+
# 1. μΉ΄ν
κ³ λ¦¬λ³ λ΄μ€
|
| 217 |
+
report.append("π° === AI λ΄μ€ λΆμ (μΉ΄ν
κ³ λ¦¬λ³) ===")
|
| 218 |
report.append("")
|
| 219 |
|
| 220 |
categorized_news = {}
|
|
|
|
| 228 |
report.append(f"π [{category}] ({len(articles)}건)")
|
| 229 |
report.append("-" * 80)
|
| 230 |
|
| 231 |
+
for i, article in enumerate(articles, 1):
|
| 232 |
report.append(f"{i}. {article['title']}")
|
| 233 |
report.append(f" π {article['url']}")
|
| 234 |
report.append(f" π
{article.get('date', 'N/A')}")
|
| 235 |
|
| 236 |
+
# LLM λΆμ (API ν€κ° μκ³ νμ±νλ κ²½μ°λ§)
|
| 237 |
+
if analyze_news and self.fireworks_api_key and i <= 2:
|
| 238 |
print(f"π€ LLM λΆμ μ€: {article['title'][:50]}...")
|
| 239 |
+
instruction = "μ΄ λ΄μ€λ₯Ό μ΄λ±νμλ μ΄ν΄ν μ μκ² 2-3λ¬Έμ₯μΌλ‘ μ€λͺ
νκ³ , μ μ€μνμ§ 1λ¬Έμ₯, νλ μ§μΉ¨ 1-2κ°λ₯Ό μλ €μ£ΌμΈμ."
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 240 |
analysis = self.analyze_with_qwen(article['title'], instruction)
|
| 241 |
+
report.append(f"\n π€ AI λΆμ: {analysis}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 242 |
|
| 243 |
report.append("")
|
| 244 |
|
|
|
|
| 249 |
report.append("")
|
| 250 |
|
| 251 |
# λͺ¨λΈ
|
| 252 |
+
if self.huggingface_data['models']:
|
| 253 |
+
report.append("π₯ νΈλ λ© λͺ¨λΈ TOP 10")
|
| 254 |
+
report.append("-" * 80)
|
| 255 |
+
for i, model in enumerate(self.huggingface_data['models'][:10], 1):
|
| 256 |
+
report.append(f"{i:2d}. {model['name']}")
|
| 257 |
+
report.append(f" π λ€μ΄λ‘λ: {model['downloads']:,} | β€οΈ {model['likes']:,}")
|
| 258 |
+
report.append(f" π·οΈ {model['task']}")
|
| 259 |
+
report.append(f" π {model['url']}")
|
| 260 |
+
report.append("")
|
| 261 |
+
else:
|
| 262 |
+
report.append("β οΈ λͺ¨λΈ λ°μ΄ν° μμ§ μ€ν¨")
|
| 263 |
|
| 264 |
report.append("")
|
| 265 |
|
| 266 |
# μ€νμ΄μ€
|
| 267 |
+
if self.huggingface_data['spaces']:
|
| 268 |
+
report.append("π νΈλ λ© μ€νμ΄μ€ μν")
|
| 269 |
+
report.append("-" * 80)
|
| 270 |
+
for i, space in enumerate(self.huggingface_data['spaces'], 1):
|
| 271 |
+
report.append(f"{i}. {space['name']}")
|
| 272 |
+
report.append(f" π {space['title']}")
|
| 273 |
+
report.append(f" π {space['url']}")
|
| 274 |
+
report.append("")
|
| 275 |
|
| 276 |
# 3. μ’
ν© μμ½
|
| 277 |
report.append("=" * 80)
|
| 278 |
report.append("π μ’
ν© μμ½")
|
| 279 |
report.append("=" * 80)
|
| 280 |
+
report.append(f"β’ μ΄ λ΄μ€: {len(news_list)}건")
|
| 281 |
+
report.append(f"β’ μΉ΄ν
κ³ λ¦¬: {len(categorized_news)}κ°")
|
| 282 |
report.append(f"β’ νΈλ λ© λͺ¨λΈ: {len(self.huggingface_data['models'])}κ°")
|
| 283 |
report.append(f"β’ νΈλ λ© μ€νμ΄μ€: {len(self.huggingface_data['spaces'])}κ°")
|
| 284 |
report.append("")
|
| 285 |
+
report.append("β
λΆμ μλ£!")
|
| 286 |
+
report.append("")
|
| 287 |
|
| 288 |
return '\n'.join(report)
|
| 289 |
|
| 290 |
+
def run(self, use_llm: bool = False) -> str:
|
| 291 |
"""μ 체 λΆμ μ€ν"""
|
| 292 |
+
print("=" * 80)
|
| 293 |
+
print("π AI λ΄μ€ & νκΉ
νμ΄μ€ νΈλ λ© λΆμ μμ")
|
| 294 |
+
print("=" * 80)
|
| 295 |
print("")
|
| 296 |
|
| 297 |
+
# 1. μν λ΄μ€ μμ±
|
| 298 |
+
print("π° μ€λμ AI λ΄μ€ λ‘λ© μ€...")
|
| 299 |
+
news_list = self.create_sample_news()
|
| 300 |
+
print(f"β
{len(news_list)}건μ λ΄μ€ λ‘λ μλ£")
|
| 301 |
print("")
|
| 302 |
|
| 303 |
+
# 2. μΉ΄ν
κ³ λ¦¬ λΆλ₯
|
| 304 |
+
print("π·οΈ μΉ΄ν
κ³ λ¦¬ λΆλ₯ μ€...")
|
| 305 |
+
categorized = self.categorize_news(news_list)
|
| 306 |
+
print("β
λΆλ₯ μλ£")
|
| 307 |
print("")
|
| 308 |
|
| 309 |
+
# 3. νκΉ
νμ΄μ€ νΈλ λ©
|
| 310 |
self.fetch_huggingface_trending()
|
| 311 |
print("")
|
| 312 |
|
| 313 |
# 4. 리ν¬νΈ μμ±
|
| 314 |
print("π 리ν¬νΈ μμ± μ€...")
|
| 315 |
+
report = self.generate_report(categorized, analyze_news=use_llm)
|
| 316 |
+
print("β
리ν¬νΈ μμ± μλ£!")
|
| 317 |
print("")
|
|
|
|
| 318 |
|
| 319 |
return report
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 320 |
|
| 321 |
|
| 322 |
+
# ==================== λ©μΈ μ€ν ====================
|
| 323 |
|
| 324 |
def main():
|
| 325 |
"""λ©μΈ μ€ν ν¨μ"""
|
| 326 |
+
print("\n")
|
| 327 |
+
print("ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ")
|
| 328 |
+
print("β AI λ΄μ€ & νκΉ
νμ΄μ€ νΈλ λ© λΆμ μμ€ν
v1.0 β")
|
| 329 |
+
print("ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ")
|
| 330 |
+
print("\n")
|
| 331 |
|
| 332 |
+
# API ν€ μ€μ (μ νμ¬ν)
|
| 333 |
+
FIREWORKS_API_KEY = None # μ¬κΈ°μ API ν€λ₯Ό μ
λ ₯νκ±°λ NoneμΌλ‘ λμΈμ
|
| 334 |
+
BRAVE_API_KEY = None
|
| 335 |
|
| 336 |
+
if not FIREWORKS_API_KEY:
|
| 337 |
+
print("βΉοΈ Fireworks API ν€κ° μ€μ λμ§ μμμ΅λλ€.")
|
| 338 |
+
print(" - LLM λΆμ κΈ°λ₯μ λΉνμ±νλ©λλ€.")
|
| 339 |
+
print(" - κΈ°λ³Έ λ΄μ€ μμ§ λ° λΆλ₯λ μ μ μλν©λλ€.")
|
| 340 |
+
print("")
|
| 341 |
|
| 342 |
# λΆμκΈ° μ΄κΈ°ν
|
| 343 |
analyzer = AINewsAnalyzer(
|
|
|
|
| 345 |
brave_api_key=BRAVE_API_KEY
|
| 346 |
)
|
| 347 |
|
| 348 |
+
# μ€ν (use_llm=Falseλ‘ μ€μ νλ©΄ API μμ΄λ λμ)
|
| 349 |
+
report = analyzer.run(use_llm=False)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 350 |
|
| 351 |
# κ²°κ³Ό μΆλ ₯
|
|
|
|
| 352 |
print(report)
|
| 353 |
|
| 354 |
# νμΌ μ μ₯
|
| 355 |
+
timestamp = datetime.now().strftime('%Y%m%d_%H%M%S')
|
| 356 |
+
filename = f"ai_news_report_{timestamp}.txt"
|
| 357 |
+
|
| 358 |
+
try:
|
| 359 |
+
with open(filename, 'w', encoding='utf-8') as f:
|
| 360 |
+
f.write(report)
|
| 361 |
+
print(f"\nπΎ 리ν¬νΈ μ μ₯ μλ£: {filename}")
|
| 362 |
+
except Exception as e:
|
| 363 |
+
print(f"\nβ οΈ νμΌ μ μ₯ μ€ν¨: {e}")
|
| 364 |
+
|
| 365 |
+
print("\n" + "=" * 80)
|
| 366 |
+
print("νλ‘κ·Έλ¨ μ’
λ£")
|
| 367 |
+
print("=" * 80 + "\n")
|
| 368 |
|
| 369 |
|
| 370 |
if __name__ == "__main__":
|
| 371 |
+
main()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|