feat: v0.0.4 - 冠字号提取功能

新增脚本：scripts/crown_extract.py
- 从一尘帖子标题/内容中提取J0开头的冠字号（J0+8位/J0+9位）
- 提取号码特征（标十、无47、PMG等）
- 提取价格信息
- 按龙钞/马钞/蛇钞/其他自动分类
- 存入 collections 表
- 每天凌晨3点定时全量扫描提取

scripts/crown_extract.py

@@ -0,0 +1,147 @@
+#!/usr/bin/env python3
+"""
+冠字号提取脚本 - 最终版
+规则：只提取J0开头的冠字号（J0+8位或J0+9位）
+"""
+import sys
+sys.path.insert(0, '/root/coolbot-data')
+
+import re
+import psycopg2
+from datetime import datetime
+
+DB_CONFIG = {
+    'host': 'pgm-bp1t1008h019ez6c.pg.rds.aliyuncs.com',
+    'port': 5432,
+    'database': 'coolbot_data',
+    'user': 'coolbot',
+    'password': 'Coolbot123'
+}
+
+FEATURE_KEYWORDS = [
+    '无47', '无34', '无347', '无4', '无3', '无2',
+    '标十', '标百', '标九', '标准十', '标准百',
+    '首日', '生日', '金马', '银马',
+    '金钩', '倒置', '满号', '圆圆',
+    '豹子', '顺子', '恐龙', '天龙',
+    'PMG', '爱藏', '尾号',
+    '大象', '麒麟', '老虎', '狮子',
+    '金马王', '天马', '龙马精神',
+    '连号', '散号', '一刀', '龙凤', '熊猫',
+]
+
+def extract_codes(text):
+    """只提取J0开头的冠字号：J0+8位或J0+9位"""
+    if not text:
+        return set()
+    codes = set()
+    for p in [r'J0\d{8}', r'J0\d{9}']:
+        for c in re.findall(p, text):
+            # J0+8位: 总长10, J0+9位: 总长11
+            if len(c) in (10, 11):
+                codes.add(c)
+    return codes
+
+def extract_features(text):
+    if not text:
+        return None
+    feat = [kw for kw in FEATURE_KEYWORDS if kw in text]
+    return '|'.join(feat) if feat else None
+
+def extract_price(text):
+    if not text:
+        return None, '元'
+    prices = []
+    for p in [r'(\d+(?:\.\d+)?)\s*元', r'(\d+(?:\.\d+)?)\s*/\s*[张件组百]']:
+        for x in re.findall(p, text):
+            try:
+                v = float(x)
+                if 1 <= v < 100000:
+                    prices.append(v)
+            except:
+                pass
+    return (min(prices), '元') if prices else (None, '元')
+
+def get_category(title):
+    if not title:
+        return '其他'
+    for kw in ['龙', '龙钞', '小龙钞', '钞王']:
+        if kw in title:
+            return '龙钞'
+    for kw in ['马', '马钞']:
+        if kw in title:
+            return '马钞'
+    for kw in ['蛇', '蛇钞']:
+        if kw in title:
+            return '蛇钞'
+    return '其他'
+
+def main():
+    print(f'[{datetime.now()}] 冠字号提取开始（仅J0冠字号）...')
+
+    conn = psycopg2.connect(**DB_CONFIG)
+    cur = conn.cursor()
+
+    # 清空旧数据
+    cur.execute('DELETE FROM collections WHERE crown_code IS NOT NULL')
+    conn.commit()
+    print(f'清空旧数据完成')
+
+    BATCH = 100
+    offset = 0
+    total_new = 0
+
+    while True:
+        cur.execute(f'''
+            SELECT id, post_id, title, content, post_type,
+                   author_username, price, price_unit, url, crawled_at
+            FROM yichens_posts
+            ORDER BY id
+            LIMIT {BATCH} OFFSET {offset}
+        ''')
+        posts = cur.fetchall()
+        if not posts:
+            break
+
+        for (pid, post_id, title, content, post_type,
+             author, price, price_unit, url, crawled_at) in posts:
+
+            text = f'{title or ""} {content or ""}'
+            codes = extract_codes(text)
+            if not codes:
+                offset += 1
+                continue
+
+            features = extract_features(text)
+            price_val, _ = extract_price(text)
+            price_val = price_val if price_val else price
+            category = get_category(title)
+
+            for code in codes:
+                try:
+                    cur.execute('''
+                        INSERT INTO collections (
+                            name, crown_code, category, post_id, post_url,
+                            author, price, price_unit,
+                            number_feature, post_title, post_type,
+                            post_crawled_at, created_at, updated_at
+                        ) VALUES (%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,NOW(),NOW())
+                    ''', (
+                        code, code, category, str(post_id), url,
+                        author, price_val, price_unit or '元',
+                        features, title, post_type,
+                        crawled_at
+                    ))
+                    total_new += 1
+                except Exception as e:
+                    pass
+
+        conn.commit()
+        offset += len(posts)
+        print(f'  已处理 {offset} 条，新增 {total_new} 条')
+
+    print(f'\n完成！总计新增: {total_new} 条')
+    conn.close()
+
+if __name__ == '__main__':
+    main()