hist.txt

# 위키 요약 관련
def extract_main_query(text):
    sentences = re.split(r'[.?!]\s*', text)
    sentences = [s.strip() for s in sentences if s.strip()]
    if not sentences:
        return text
    last = sentences[-1]
    last = re.sub(r'[^가-힣a-zA-Z0-9 ]', '', last)
    particles = ['이', '가', '은', '는', '을', '를', '의', '에서', '에게', '한테', '보다']
    for p in particles:
        last = re.sub(rf'\b(\w+){p}\b', r'\1', last)
    return last.strip()

def get_wikipedia_summary(query):
    cleaned_query = extract_main_query(query)
    url = f"https://ko.wikipedia.org/api/rest_v1/page/summary/{cleaned_query}"
    res = requests.get(url)
    if res.status_code == 200:
        return res.json().get("extract", "요약 정보를 찾을 수 없습니다.")
    else:
        return "위키백과에서 정보를 가져올 수 없습니다."

def textrank_summarize(text, top_n=3):
    sentences = re.split(r'(?<=[.!?])\s+', text.strip())
    sentences = [s.strip() for s in sentences if len(s.strip()) > 10]
    if len(sentences) <= top_n:
        return text
    vectorizer = TfidfVectorizer()
    tfidf_matrix = vectorizer.fit_transform(sentences)
    sim_matrix = cosine_similarity(tfidf_matrix)
    np.fill_diagonal(sim_matrix, 0)
    def pagerank(matrix, damping=0.85, max_iter=100, tol=1e-4):
        N = matrix.shape[0]
        ranks = np.ones(N) / N
        row_sums = np.sum(matrix, axis=1)
        row_sums[row_sums == 0] = 1
        for _ in range(max_iter):
            prev_ranks = ranks.copy()
            for i in range(N):
                incoming = matrix[:, i]
                ranks[i] = (1 - damping) / N + damping * np.sum(incoming * prev_ranks / row_sums)
            if np.linalg.norm(ranks - prev_ranks) < tol:
                break
        return ranks
    scores = pagerank(sim_matrix)
    ranked_idx = np.argsort(scores)[::-1]
    selected_idx = sorted(ranked_idx[:top_n])
    summary = ' '.join([sentences[i] for i in selected_idx])
    return summary

def summarize_from_wikipedia(query, top_n=3):  
    raw_summary = get_wikipedia_summary(query)  
    first_summary = textrank_summarize(raw_summary, top_n=top_n)  
    second_summary = textrank_summarize(first_summary, top_n=top_n)  
    return second_summary