Spaces:
Paused
Paused
File size: 10,514 Bytes
a4c91de 4276d2b 7a35c95 a4c91de f163a93 a4c91de f163a93 a4c91de 7a35c95 a4c91de 1da5347 2a42d0f a4c91de f163a93 a4c91de f163a93 a4c91de d5916a0 4276d2b d5916a0 4276d2b d5916a0 c3cf3bf d5916a0 c3cf3bf d5916a0 c3cf3bf d5916a0 a4c91de f163a93 a4c91de c3cf3bf a4c91de f163a93 a4c91de 4276d2b a4c91de f163a93 a4c91de 4276d2b d5916a0 f163a93 a4c91de |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 |
import os
import random
import time
import re
import json
import requests
from bs4 import BeautifulSoup
from requests.adapters import HTTPAdapter
from requests.packages.urllib3.util.retry import Retry
import openai
import gradio as gr
from fpdf import FPDF as FPDF2
from datetime import datetime
# API ํค ์ค์
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
# OpenAI ์ค์
openai.api_key = OPENAI_API_KEY
def setup_session():
try:
session = requests.Session()
retries = Retry(total=5, backoff_factor=1, status_forcelist=[502, 503, 504])
session.mount('https://', HTTPAdapter(max_retries=retries))
return session
except Exception as e:
return None
def generate_naver_search_url(query):
base_url = "https://search.naver.com/search.naver?"
params = {"ssc": "tab.blog.all", "sm": "tab_jum", "query": query}
url = base_url + "&".join(f"{key}={value}" for key, value in params.items())
return url
def crawl_blog_content(url, session):
try:
headers = {
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.3",
"Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8",
"Accept-Language": "ko-KR,ko;q=0.9,en-US;q=0.8,en;q=0.7",
"Accept-Encoding": "gzip, deflate, br",
"Connection": "keep-alive",
"Referer": "https://search.naver.com/search.naver",
}
delay = random.uniform(1, 2)
time.sleep(delay)
response = session.get(url, headers=headers)
if response.status_code != 200:
return ""
soup = BeautifulSoup(response.content, "html.parser")
content = soup.find("div", attrs={'class': 'se-main-container'})
if content:
return clean_text(content.get_text())
else:
return ""
except Exception as e:
return ""
def crawl_naver_search_results(url, session):
try:
headers = {
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.3",
"Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8",
"Accept-Language": "ko-KR,ko;q=0.9,en-US;q=0.8,en;q=0.7",
"Accept-Encoding": "gzip, deflate, br",
"Connection": "keep-alive",
"Referer": "https://search.naver.com/search.naver",
}
response = session.get(url, headers=headers)
if response.status_code != 200:
return []
soup = BeautifulSoup(response.content, "html.parser")
results = []
count = 0
for li in soup.find_all("li", class_=re.compile("bx.*")):
if count >= 10:
break
for div in li.find_all("div", class_="detail_box"):
for div2 in div.find_all("div", class_="title_area"):
title = div2.text.strip()
for a in div2.find_all("a", href=True):
link = a["href"]
if "blog.naver" in link:
link = link.replace("https://", "https://m.")
results.append({"์ ๋ชฉ": title, "๋งํฌ": link})
count += 1
if count >= 10:
break
if count >= 10:
break
if count >= 10:
break
return results
except Exception as e:
return []
def clean_text(text):
text = re.sub(r'\s+', ' ', text).strip()
return text
def fetch_references(topic):
search_url = generate_naver_search_url(topic)
session = setup_session()
if session is None:
return ["์ธ์
์ค์ ์คํจ"] * 3
results = crawl_naver_search_results(search_url, session)
if len(results) < 3:
return ["์ถฉ๋ถํ ๊ฒ์ ๊ฒฐ๊ณผ๋ฅผ ์ฐพ์ง ๋ชปํ์ต๋๋ค."] * 3
selected_results = random.sample(results, 3)
references = []
for result in selected_results:
content = crawl_blog_content(result['๋งํฌ'], session)
references.append(f"์ ๋ชฉ: {result['์ ๋ชฉ']}\n๋ด์ฉ: {content}")
return references
def fetch_crawl_results(query):
references = fetch_references(query)
return references[0], references[1], references[2]
def generate_blog_post(query, prompt_template):
try:
# ์ฐธ๊ณ ๊ธ ํฌ๋กค๋ง
references = fetch_references(query)
ref1, ref2, ref3 = references
combined_content = f"์ฐธ๊ณ ๊ธ1:\n{ref1}\n\n์ฐธ๊ณ ๊ธ2:\n{ref2}\n\n์ฐธ๊ณ ๊ธ3:\n{ref3}"
# ๋๋ค ์๋ ์์ฑ
random_seed = random.randint(1, 10000)
full_prompt = f"์ฃผ์ : {query}\n\n{prompt_template}\n\n์ฐธ๊ณ ๋ด์ฉ:\n{combined_content}"
response = openai.ChatCompletion.create(
model="gpt-4o-mini",
messages=[
{"role": "system", "content": prompt_template},
{"role": "user", "content": full_prompt}
],
max_tokens=10000,
temperature=0.75,
top_p=1.0,
frequency_penalty=0.3
)
return f"์ฃผ์ : {query}\n\n{response.choices[0].message['content']}", ref1, ref2, ref3
except Exception as e:
return f"๋ธ๋ก๊ทธ ๊ธ ์์ฑ ์ค ์ค๋ฅ ๋ฐ์: {str(e)}", "", "", ""
# PDF ํด๋์ค ๋ฐ ๊ด๋ จ ํจ์ ์ ์
class PDF(FPDF2):
def __init__(self):
super().__init__()
current_dir = os.path.dirname(__file__)
self.add_font("NanumGothic", "", os.path.join(current_dir, "NanumGothic.ttf"))
self.add_font("NanumGothic", "B", os.path.join(current_dir, "NanumGothicBold.ttf"))
self.add_font("NanumGothicExtraBold", "", os.path.join(current_dir, "NanumGothicExtraBold.ttf"))
self.add_font("NanumGothicLight", "", os.path.join(current_dir, "NanumGothicLight.ttf"))
def header(self):
self.set_font('NanumGothic', '', 10)
def footer(self):
self.set_y(-15)
self.set_font('NanumGothic', '', 8)
self.cell(0, 10, f'Page {self.page_no()}', 0, 0, 'C')
def save_to_pdf(blog_post, user_topic):
pdf = PDF()
pdf.add_page()
lines = blog_post.split('\n')
title = lines[0].strip()
content = '\n'.join(lines[1:]).strip()
# ํ์ฌ ๋ ์ง์ ์๊ฐ์ ๊ฐ์ ธ์ต๋๋ค (๋ํ๋ฏผ๊ตญ ์๊ฐ ๊ธฐ์ค)
now = datetime.now()
date_str = now.strftime("%y%m%d")
time_str = now.strftime("%H%M")
# ํ์ผ๋ช
์์ฑ
filename = f"{date_str}_{time_str}_{format_filename(user_topic)}.pdf"
pdf.set_font("NanumGothic", 'B', size=14)
pdf.cell(0, 10, title, ln=True, align='C')
pdf.ln(10)
pdf.set_font("NanumGothic", '', size=11)
pdf.multi_cell(0, 5, content)
print(f"Saving PDF as: {filename}")
pdf.output(filename)
return filename
def format_filename(text):
text = re.sub(r'[^\w\s-]', '', text)
return text[:50].strip()
def save_content_to_pdf(blog_post, user_topic):
return save_to_pdf(blog_post, user_topic)
# ๊ธฐ๋ณธ ํ๋กฌํํธ ํ
ํ๋ฆฟ
DEFAULT_PROMPT_TEMPLATE = """
[๋ธ๋ก๊ทธ ๊ธ ์์ฑ ๊ธฐ๋ณธ ๊ท์น]
1. ๋ฐ๋์ ํ๊ธ๋ก ์์ฑํ๋ผ
2. ์ฃผ์ด์ง ์ฐธ๊ณ ๊ธ์ ๋ฐํ์ผ๋ก 1๊ฐ์ ์ ๋ณด์ฑ(Informativity) ๋ธ๋ก๊ทธ๋ฅผ ์์ฑ
3. ์ฃผ์ ์ ์ ๋ชฉ์ ์ ์ธํ ๊ธ์ด 1100๋จ์ด ์ด์์ด ๋๋๋ก ์์ฑ
4. ๊ธ์ ์ ๋ชฉ์ ์ ๋ณด์ฑ ๋ธ๋ก๊ทธ ํํ์ ๋ง๋ ์ ์ ํ ์ ๋ชฉ์ผ๋ก ์ถ๋ ฅ
- ์ฐธ๊ณ ๊ธ์ ์ ๋ชฉ๋ ์ฐธ๊ณ ํ๋, ๋์ผํ๊ฒ ์์ฑํ์ง ๋ง ๊ฒ
5. ๋ฐ๋์ ๋งํฌ๋ค์ด ํ์์ด ์๋ ์์ํ ํ
์คํธ๋ก๋ง ์ถ๋ ฅํ๋ผ
6. ๋ค์ํ๋ฒ ์ฐธ๊ณ ๊ธ์ ๊ฒํ ํ์ฌ ๋ด์ฉ์ ์ถฉ๋ถํ ๋ฐ์ํ๋, ์ฐธ๊ณ ๊ธ์ ํํ์ ๊ทธ๋๋ก ์์ฑํ์ง๋ ๋ง ๊ฒ
[๋ธ๋ก๊ทธ ๊ธ ์์ฑ ์ธ๋ถ ๊ท์น]
1. ์ฌ์ฉ์๊ฐ ์
๋ ฅํ ์ฃผ์ ์ ์ฃผ์ด์ง ์ฐธ๊ณ ๊ธ 3๊ฐ๋ฅผ ๋ฐํ์ผ๋ก ๋ธ๋ก๊ทธ ๊ธ 1๊ฐ๋ฅผ ์์ฑํ๋ผ
2. ์ฃผ์ด์ง ๋ชจ๋ ๊ธ์ ๋ถ์ํ์ฌ ํ๋์ ๋์ฃผ์ ๋ฅผ ์ ์ ํ๋ผ
3. ๋์ฃผ์ ์ ๋ง๊ฒ ์ฐธ๊ณ ๊ธ 3๊ฐ์ง์ ๋ด์ฉ์ด ๋ชจ๋ ๋ฐ์๋๊ฒ ์์ฑ
4. ๋
์๊ฐ ์ฝ๊ณ ์ฃผ์ ์ ์ ๋ณด๋ฅผ ์ถฉ๋ถํ ์ต๋ ํ ์ ์๋๋ก ๋ด์ฉ์ ํ์ฑํ๊ฒ ์์ฑ
5. ๋
์์ ๊ณต๊ฐ๊ณผ ๊ด์ฌ์ ๋ ์ ์๋๋ก ์์ฑํ๊ณ ๊ฒฝํ์ด๋ ์ผํ, ํฅ๋ฏธ๋ก์ด ์ฌ์ค, ์ ๋ฌธ๊ฐ์ ๊ฒฌํด ๋ฑ์ ํฌํจ
6. ๋
์์๊ฒ ์ด๋์ด ๋๋ ์ ๋ณด๋ฅผ ์์ฑ
7. ์ดํฌ๋ ์ฃผ์ด์ง ์ฐธ๊ณ ๊ธ 3๊ฐ์ง์ ์ดํฌ๋ฅผ ์ ์ ํ ๋ฐ์ํ๋ผ
- ํนํ ๋ฌธ์ฅ์ ๋ ๋ถ๋ถ์ ์ ์ ํ ๋ฐ์('~์ด์'), ('~๋๋ค' ์ ์ธ)
- ๋๋ฌด ๋ฑ๋ฑํ์ง ์๊ฒ ํธ์ํ๊ฒ ์ฝ์ ์ ์๋๋ก ์์ฐ์ค๋ฌ์ด ๋ํ์ฒด๋ฅผ ๋ฐ์
[์ ์ธ ๊ท์น]
1. ๋ฐ๋์ ์ฐธ๊ณ ๊ธ์ ํฌํจ๋ ๋งํฌ(URL)๋ ์ ์ธ
2. ์ฐธ๊ณ ๊ธ์์ '๋งํฌ๋ฅผ ํ์ธํด์ฃผ์ธ์'์ ๊ฐ์ ๋งํฌ ์ด๋์ ๋ฌธ๊ตฌ๋ ์ ์ธ
3. ์ฐธ๊ณ ๊ธ์ ์๋ ์์ฑ์, ํ์, ์ ํ๋ฒ, ๊ธฐ์(Writer, speaker, YouTuber, reporter)์ ์ด๋ฆ, ์ ์นญ, ๋๋ค์(Name, Nkickname)์ ๋ฐ๋์ ์ ์ธ
"""
# Gradio ์ฑ ์์ฑ
with gr.Blocks() as iface:
gr.Markdown("# ๋ธ๋ก๊ทธ ๊ธ ์์ฑ๊ธฐ_์ ๋ณด์ฑ_์ผ๋ฐ")
gr.Markdown("์ฃผ์ ๋ฅผ ์
๋ ฅํ๊ณ ๋ธ๋ก๊ทธ ๊ธ ์์ฑ ๋ฒํผ์ ๋๋ฅด๋ฉด ์๋์ผ๋ก ๋ธ๋ก๊ทธ ๊ธ์ ์์ฑํฉ๋๋ค.")
query_input = gr.Textbox(lines=1, placeholder="๋ธ๋ก๊ทธ ๊ธ์ ์ฃผ์ ๋ฅผ ์
๋ ฅํด์ฃผ์ธ์...", label="์ฃผ์ ")
prompt_input = gr.Textbox(lines=10, value=DEFAULT_PROMPT_TEMPLATE, label="ํ๋กฌํํธ ํ
ํ๋ฆฟ", visible=False)
generate_button = gr.Button("๋ธ๋ก๊ทธ ๊ธ ์์ฑ")
output_text = gr.Textbox(label="์์ฑ๋ ๋ธ๋ก๊ทธ ๊ธ")
ref1_text = gr.Textbox(label="์ฐธ๊ณ ๊ธ 1", lines=10, visible=False)
ref2_text = gr.Textbox(label="์ฐธ๊ณ ๊ธ 2", lines=10, visible=False)
ref3_text = gr.Textbox(label="์ฐธ๊ณ ๊ธ 3", lines=10, visible=False)
save_pdf_button = gr.Button("PDF๋ก ์ ์ฅ")
pdf_output = gr.File(label="์์ฑ๋ PDF ํ์ผ")
generate_button.click(
generate_blog_post,
inputs=[query_input, prompt_input],
outputs=[output_text, ref1_text, ref2_text, ref3_text],
show_progress=True
)
save_pdf_button.click(
save_content_to_pdf,
inputs=[output_text, query_input],
outputs=[pdf_output],
show_progress=True
)
# Gradio ์ฑ ์คํ
if __name__ == "__main__":
iface.launch() |