Files
yanxuan/脚本/audit_slice_caliber.py
Your Name 62ac155641 同步择校 UI 三十四轮改版与审计产出
择校 UI(docs/择校UI):
- 搜索支持外号(东大/华科)与拼音首字母(dndx),未命中做分诊
- 目标分拖动即见三档数量,松手列出换档院校,名单按收藏/对比/本地排序
- 家乡省份接入院校库(家乡优先排序)与首页推荐(池不足时说明并可按档位判断)
- 缺档改成同尺寸占位卡并可跳院校库;三档口径收敛成一份、默认收起
- 新增拼音索引.js(907 校)

设计与交接:docs/研选E决策层_20261112 36 张稿与 README v1–v34;docs/agent_handoffs 续写。

审计与脚本:审计报告新增若干核验与分流产物,脚本同步更新。

.gitignore 追加排除 /官方资料/ 与 GB 级审计中间产物(global_writer_stages、来源发现、local_ingest_*)。
2026-11-13 22:32:10 +08:00

62 lines
3.1 KiB
Python
Raw Permalink Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
"""扫「名为全量、实为切片」的表(只读)
起因:`chsi_2027_tuimian_official` 被当作"官方推免表"参与审计,实测只含 085600–085606
七个码、1,858 行、224 校,且有 truncated_page1_only 列 —— 是 0856 专项切片。
用户已明确否掉 0856 定向口径,所以要扫全库:还有哪些表带专业码列但实际只覆盖一个窄码段,
以及它们是否被当全量分母 / 覆盖证据用。
判据(全部实测,不靠表名猜):
- 码段首位分布:只看 substr(码,1,2) 的 distinct 个数
- 切片标记:distinct 码数 <= 12 且全部落在同一 4 位前缀 → 判「窄切片」
- 对照官方目录 Bc27=1,404 个码,给出覆盖占比
"""
import sqlite3
con = sqlite3.connect(r'D:\择校\数据库\zexiao_2027.db')
tables = [r[0] for r in con.execute(
"SELECT name FROM sqlite_master WHERE type='table' AND name NOT LIKE 'sqlite_%' ORDER BY name")]
print('库里表总数:', len(tables))
CODE_CANDS = ['zydm', 'code', 'major_code', '专业代码', 'zym', 'b_code', 'c_code', 'a_code',
'school_code', 'dwdm']
rows = []
for t in tables:
cols = [r[1] for r in con.execute('PRAGMA table_info("%s")' % t)]
hit = [c for c in cols if c in ('zydm', 'code', 'major_code', '专业代码', 'b_code', 'c_code')]
if not hit:
continue
col = hit[0]
try:
n_rows = con.execute('SELECT COUNT(*) FROM "%s"' % t).fetchone()[0]
nd = con.execute('SELECT COUNT(DISTINCT CAST("%s" AS TEXT)) FROM "%s"' % (col, t)).fetchone()[0]
if nd is None or nd == 0:
continue
pref = con.execute('SELECT DISTINCT substr(CAST("%s" AS TEXT),1,4) FROM "%s" '
'WHERE "%s" IS NOT NULL' % (col, t, col)).fetchall()
pref = sorted(x[0] for x in pref if x[0])
p2 = sorted({p[:2] for p in pref if p})
rows.append((t, col, n_rows, nd, p2, pref))
except Exception as e:
print(' !! %-40s %s' % (t, str(e)[:60]))
BC27 = con.execute('SELECT COUNT(DISTINCT CAST(zydm AS TEXT)) FROM chsi_2027_dir').fetchone()[0]
print('对照分母 Bc27(官方 2027 distinct 专业码)=%d\n' % BC27)
narrow = [r for r in rows if r[3] <= 12]
print('=== 有专业码列的表 %d 张,其中 distinct 码 <=12 的窄表 %d 张 ===' % (len(rows), len(narrow)))
for t, col, n_rows, nd, p2, pref in sorted(narrow, key=lambda x: -x[2]):
print(' %-34s 列=%-10s 行=%-8s distinct码=%-4s 码段=%s 4位前缀=%s' %
(t, col, format(n_rows, ','), nd, p2[:6], [p for p in pref if p][:12]))
print('\n=== 全部窄码段的"只覆盖 0856"表 ===')
for t, col, n_rows, nd, p2, pref in rows:
if pref and all(p.startswith('0856') for p in pref):
print(' %-34s 行=%-8s distinct码=%-4s 前缀=%s' % (t, format(n_rows, ','), nd, pref))
print('\n=== 只覆盖单个门类(首位 2 段唯一)的表 ===')
for t, col, n_rows, nd, p2, pref in rows:
if len(p2) == 1:
print(' %-34s 列=%-10s 行=%-9s distinct码=%-5s 唯一门类=%s 覆盖 Bc27 的 %.2f%%' %
(t, col, format(n_rows, ','), nd, p2[0], 100.0 * nd / BC27))