analysis_03_hidden.py

#!/usr/bin/env python3
"""
川普密碼 分析 #3 — 隱藏訊息搜索
藏頭詩、首字母、重複模式、數字密碼
"""

import json
import re
from collections import Counter
from pathlib import Path

BASE = Path(__file__).parent
DATA = BASE / "data"


def main():
    with open(BASE / "clean_president.json", 'r', encoding='utf-8') as f:
        posts = json.load(f)

    originals = [p for p in posts if p['has_text'] and not p['is_retweet'] and p['content_length'] > 50]

    print("=" * 70)
    print("🕵️ 分析 #3: 隱藏訊息搜索")
    print(f"   分析對象: 就任後原創貼文（>50字）{len(originals)} 篇")
    print("=" * 70)

    # --- 1. 藏頭詩分析：每篇貼文每句話第一個字母 ---
    print(f"\n📜 藏頭詩分析（每句第一個字母）:")
    print("-" * 60)

    acrostics = []
    for p in originals:
        content = p['content']
        # 用句號、驚嘆號、問號分句
        sentences = re.split(r'[.!?]+', content)
        sentences = [s.strip() for s in sentences if s.strip() and len(s.strip()) > 3]

        if len(sentences) >= 3:  # 至少 3 句才有意義
            first_letters = ''.join([s[0].upper() for s in sentences if s])
            acrostics.append({
                'date': p['created_at'][:10],
                'letters': first_letters,
                'sentence_count': len(sentences),
                'content_preview': content[:100],
                'url': p['url']
            })

    print(f"  分析了 {len(acrostics)} 篇（≥3 句）")

    # 找看起來像英文單字的
    import string
    # 常見英文字典（簡易）
    interesting_words = {'MAGA', 'TRUMP', 'WIN', 'USA', 'WAR', 'DEAL', 'HELP', 'FIRE',
                         'VOTE', 'LOVE', 'HATE', 'KILL', 'STOP', 'SAVE', 'PLAN', 'CODE',
                         'FAKE', 'TRUE', 'GOLD', 'SELL', 'BUY', 'RUN', 'FLY', 'SPY',
                         'SOS', 'FBI', 'CIA', 'GOD', 'OIL', 'GAS', 'TAX', 'PAY',
                         'END', 'NEW', 'OLD', 'BIG', 'TOP', 'RED', 'BLUE', 'FREE',
                         'RICH', 'POOR', 'FAST', 'SLOW', 'GOOD', 'EVIL', 'FEAR',
                         'HOPE', 'DOOM', 'BOOM', 'BUST', 'CASH', 'DEBT', 'BOND',
                         'WALL', 'GATE', 'DARK', 'DEEP', 'HIGH', 'LOW', 'FALL',
                         'RISE', 'LOOK', 'HIDE', 'FIND', 'LOST', 'SAFE', 'RISK',
                         'NUKE', 'BOMB', 'IRAN', 'NATO', 'ASIA', 'EURO'}

    found_words = []
    for a in acrostics:
        letters = a['letters']
        # 滑動窗口找子字串
        for wlen in range(3, min(len(letters) + 1, 8)):
            for i in range(len(letters) - wlen + 1):
                sub = letters[i:i + wlen]
                if sub in interesting_words:
                    found_words.append({
                        'word': sub,
                        'full_letters': letters,
                        'date': a['date'],
                        'url': a['url'],
                        'preview': a['content_preview']
                    })

    print(f"\n🎯 藏頭詩中找到已知英文字:")
    if found_words:
        for fw in found_words[:20]:
            print(f"  {fw['date']} | 藏頭: {fw['full_letters']} | 找到: [{fw['word']}] | {fw['preview'][:60]}...")
    else:
        print("  （未找到明顯英文單字）")

    # 顯示所有藏頭（最近 30 篇）
    print(f"\n📋 最近 30 篇藏頭字母:")
    for a in acrostics[:30]:
        print(f"  {a['date']} | {a['letters']:20s} ({a['sentence_count']}句)")

    # --- 2. 數字密碼 ---
    print(f"\n🔢 數字密碼分析:")
    print("-" * 60)

    all_numbers = Counter()
    number_posts = []

    for p in originals:
        numbers = re.findall(r'\b(\d[\d,.]*)\b', p['content'])
        cleaned_nums = []
        for n in numbers:
            # P3-5: 只移除千位逗號，保留小數點，用 float() 解析
            clean = n.replace(',', '')
            try:
                val = float(clean)
                cleaned_nums.append(val)
                all_numbers[n] += 1  # 保留原始格式
            except ValueError:
                continue

        if cleaned_nums:
            number_posts.append({
                'date': p['created_at'][:10],
                'numbers': cleaned_nums,
                'raw': numbers,
                'content': p['content'][:100]
            })

    print(f"  含數字的貼文: {len(number_posts)} 篇")
    print(f"  不同數字: {len(all_numbers)} 種")

    print(f"\n🏆 Top 30 最常出現的數字:")
    for num, count in all_numbers.most_common(30):
        print(f"  {num:>15s} → {count:3d} 次")

    # 特殊數字分析
    print(f"\n🔮 特殊數字模式:")
    all_ints = []
    for np_item in number_posts:
        all_ints.extend(int(v) for v in np_item['numbers'] if v == int(v))

    # 重複出現的具體數字
    int_counts = Counter(all_ints)
    repeated = [(n, c) for n, c in int_counts.most_common(30) if c >= 3 and n > 10]
    print(f"  重複 3 次以上的數字（>10）:")
    for n, c in repeated[:15]:
        print(f"    {n:>12,} → {c} 次")

    # --- 3. 重複短語（n-gram）分析 ---
    print(f"\n🔄 重複短語分析:")
    print("-" * 60)

    # 3-gram
    trigrams = Counter()
    bigrams = Counter()

    for p in originals:
        words = re.findall(r'[A-Za-z\']+', p['content'])
        words_lower = [w.lower() for w in words]

        for i in range(len(words_lower) - 2):
            tri = ' '.join(words_lower[i:i+3])
            trigrams[tri] += 1

        for i in range(len(words_lower) - 1):
            bi = ' '.join(words_lower[i:i+2])
            bigrams[bi] += 1

    # 過濾掉太普通的（the of a 等）
    stop_starts = {'the', 'of', 'a', 'an', 'in', 'to', 'and', 'is', 'it', 'for',
                   'on', 'at', 'by', 'with', 'from', 'as', 'that', 'this', 'was',
                   'are', 'be', 'has', 'have', 'had', 'will', 'would', 'which'}

    interesting_tri = [(t, c) for t, c in trigrams.most_common(500)
                       if c >= 5 and t.split()[0] not in stop_starts]

    print(f"  Top 25 川普口頭禪（3 詞組合，出現 ≥5 次）:")
    for phrase, count in interesting_tri[:25]:
        bar = '█' * min(count, 30)
        print(f"  {phrase:35s} {count:4d} {bar}")

    # --- 4. 標點符號密碼 ---
    print(f"\n❗ 標點符號分析:")
    print("-" * 60)

    excl_counts = []  # 每篇的驚嘆號數
    for p in originals:
        excl = p['content'].count('!')
        excl_counts.append({
            'date': p['created_at'][:10],
            'exclamation': excl,
            'question': p['content'].count('?'),
            'ellipsis': p['content'].count('...'),
            'dash': p['content'].count('—') + p['content'].count('-'),
            'length': p['content_length']
        })

    total_excl = sum(e['exclamation'] for e in excl_counts)
    total_quest = sum(e['question'] for e in excl_counts)
    total_ellip = sum(e['ellipsis'] for e in excl_counts)

    print(f"  驚嘆號 ! 總計: {total_excl} 個 (平均每篇 {total_excl/len(originals):.1f})")
    print(f"  問號 ? 總計: {total_quest} 個")
    print(f"  省略號 ... 總計: {total_ellip} 個")

    # 驚嘆號最多的貼文
    top_excl = sorted(excl_counts, key=lambda x: x['exclamation'], reverse=True)
    print(f"\n  驚嘆號最多的 5 篇:")
    for e in top_excl[:5]:
        # 找到對應的貼文
        # P4-1: next() 加 default=None
        post = next((p for p in originals if p['created_at'][:10] == e['date']), None)
        if post:
            print(f"    {e['date']} | !×{e['exclamation']} | {post['content'][:80]}...")

    # --- 5. 結尾簽名模式 ---
    print(f"\n✍️ 結尾簽名模式:")
    print("-" * 60)

    endings = Counter()
    for p in originals:
        content = p['content'].strip()
        # 取最後一句
        last_sentence = re.split(r'[.!?]', content)
        last = last_sentence[-1].strip() if last_sentence else ''
        if len(last) > 10 and len(last) < 80:
            endings[last] += 1

    print("  重複出現的結尾:")
    for ending, count in endings.most_common(15):
        if count >= 3:
            print(f"    [{count:3d}次] \"{ending}\"")

    # 存結果
    results = {
        'acrostics_sample': acrostics[:50],
        'acrostic_words_found': found_words,
        'top_numbers': dict(all_numbers.most_common(50)),
        'top_trigrams': dict(interesting_tri[:50]),
        'punctuation': {
            'total_exclamation': total_excl,
            'total_question': total_quest,
            'avg_exclamation': round(total_excl/len(originals), 1),
        },
        'top_endings': dict(endings.most_common(20)),
    }
    with open(DATA / 'results_03_hidden.json', 'w', encoding='utf-8') as f:
        json.dump(results, f, ensure_ascii=False, indent=2)

    print(f"\n💾 詳細結果存入 results_03_hidden.json")


if __name__ == '__main__':
    main()