match-guide-track-default/scripts/db/connection.py

"""SQLite 连接与幂等初始化（申报赛道推荐助手）。

业务表 3 张：
- teacher_profiles（教师画像，可编辑）
- guide_tracks（指南赛道，可编辑）
- match_reports（推荐报告，只读，匹配自动写入）

表间禁止 SQLite 外键，仅存业务 id（见 references/SCHEMA.md）。
"""

from __future__ import annotations

import sqlite3

from db.display_metadata import (
    init_display_metadata,
    seed_column_metadata_batch,
    upsert_table_metadata,
)
from db.timestamp_columns import init_updated_at_trigger
from util.runtime_paths import get_db_path


def get_conn() -> sqlite3.Connection:
    return sqlite3.connect(get_db_path())


_TEACHER_PROFILES_DDL = """
CREATE TABLE IF NOT EXISTS teacher_profiles (
    id             INTEGER PRIMARY KEY AUTOINCREMENT,
    teacher_name   TEXT NOT NULL UNIQUE,
    department     TEXT,
    title          TEXT,
    research_tags  TEXT,
    recent_papers  TEXT,
    project_history TEXT,
    sci_q1_count   INTEGER,
    core_paper_count INTEGER,
    created_at     INTEGER NOT NULL DEFAULT (unixepoch()),
    updated_at     INTEGER NOT NULL DEFAULT (unixepoch())
)
"""

_GUIDE_TRACKS_DDL = """
CREATE TABLE IF NOT EXISTS guide_tracks (
    id                INTEGER PRIMARY KEY AUTOINCREMENT,
    year              INTEGER NOT NULL,
    agency            TEXT,
    program_name      TEXT,
    track_name        TEXT NOT NULL,
    track_tags        TEXT,
    priority_keywords TEXT,
    min_title         TEXT,
    min_sci_q1        INTEGER NOT NULL DEFAULT 0,
    quota             INTEGER,
    source_doc        TEXT,
    created_at        INTEGER NOT NULL DEFAULT (unixepoch()),
    updated_at        INTEGER NOT NULL DEFAULT (unixepoch()),
    UNIQUE (year, agency, program_name, track_name)
)
"""

_MATCH_REPORTS_DDL = """
CREATE TABLE IF NOT EXISTS match_reports (
    id           INTEGER PRIMARY KEY AUTOINCREMENT,
    batch_no     TEXT NOT NULL,
    teacher_name TEXT,
    department   TEXT,
    season       TEXT,
    top_json     TEXT,
    tiers_json   TEXT,
    eligible_count INTEGER,
    excluded_json  TEXT,
    report_file  TEXT,
    created_at   INTEGER NOT NULL DEFAULT (unixepoch())
)
"""

_MATCH_REPORTS_INDEX = """
CREATE INDEX IF NOT EXISTS idx_match_reports_batch
ON match_reports(batch_no)
"""

_TS_COLUMN_NAMES = {"created_at", "updated_at"}


def _col(column_name: str, display_name: str, description: str, **extra: object) -> dict[str, object]:
    base: dict[str, object] = {
        "column_name": column_name,
        "display_name": display_name,
        "description": description,
        "visible": 1,
        "searchable": 0,
        "editable": 0,
    }
    base.update(extra)
    if column_name in _TS_COLUMN_NAMES:
        base["display_type"] = "datetime_unix_seconds"
    return base


_TEACHER_PROFILES_COLUMNS: tuple[dict[str, object], ...] = (
    _col("id", "编号", "主键", searchable=1),
    _col("teacher_name", "教师姓名", "教师姓名（唯一）。", searchable=1),
    _col("department", "所属学院", "二级学院/单位。", searchable=1, editable=1),
    _col("title", "职称", "讲师/副教授/教授等，用于职称门槛比对。", searchable=1, editable=1, options_json='[{"value":"讲师","label":"讲师"},{"value":"副教授","label":"副教授"},{"value":"教授","label":"教授"}]'),
    _col("research_tags", "研究标签", "研究方向关键词，逗号分隔（匹配主依据）。", searchable=1, editable=1),
    _col("recent_papers", "近3年论文", "近 3 年论文标题，分号分隔。", editable=1),
    _col("project_history", "项目经历", "在研/结题项目名称，分号分隔。", editable=1),
    _col("sci_q1_count", "SCI一区篇数", "SCI 一区论文数量（硬门槛比对）。", searchable=1, editable=1, display_type="number"),
    _col("core_paper_count", "核心论文篇数", "核心期刊论文数量。", searchable=1, editable=1, display_type="number"),
    _col("created_at", "创建时间", "记录创建时间（Unix 秒级，自动生成）。"),
    _col("updated_at", "更新时间", "记录最后更新时间（Unix 秒级，自动维护）。"),
)

_GUIDE_TRACKS_COLUMNS: tuple[dict[str, object], ...] = (
    _col("id", "编号", "主键", searchable=1),
    _col("year", "指南年度", "指南发布/申报年度。", searchable=1),
    _col("agency", "学部/归口", "如 数理学部、科研处。", searchable=1, editable=1),
    _col("program_name", "项目类型", "如 面上项目、青年项目。", searchable=1, editable=1),
    _col("track_name", "赛道名称", "指南中的方向/赛道名。", searchable=1, editable=1),
    _col("track_tags", "需求标签", "赛道方向关键词，逗号分隔（匹配主依据）。", searchable=1, editable=1),
    _col("priority_keywords", "优先方向", "优先资助方向关键词，逗号分隔（加分项）。", editable=1),
    _col("min_title", "职称门槛", "讲师及以上 / 副高及以上 / 正高及以上。", editable=1, options_json='[{"value":"讲师及以上","label":"讲师及以上"},{"value":"副高及以上","label":"副高及以上"},{"value":"正高及以上","label":"正高及以上"}]'),
    _col("min_sci_q1", "SCI一区门槛", "要求的 SCI 一区最少篇数。", editable=1, display_type="number"),
    _col("quota", "拟资助数", "该赛道拟资助项数（参考）。", editable=1, display_type="number"),
    _col("source_doc", "指南来源", "指南文件名/文号。", editable=1),
    _col("created_at", "创建时间", "记录创建时间（Unix 秒级，自动生成）。"),
    _col("updated_at", "更新时间", "记录最后更新时间（Unix 秒级，自动维护）。"),
)

_MATCH_REPORTS_COLUMNS: tuple[dict[str, object], ...] = (
    _col("id", "报告编号", "主键，推荐报告编号。", searchable=1),
    _col("batch_no", "匹配批次号", "同一次匹配请求的批次号。", searchable=1),
    _col("teacher_name", "教师姓名", "被匹配教师。", searchable=1),
    _col("department", "所属学院", "教师所属学院。", searchable=1),
    _col("season", "指南年度", "匹配所用指南年度。", searchable=1),
    _col("top_json", "Top3推荐", "Top 3 赛道（JSON，含得分与理由）。", visible=0, display_type="textarea"),
    _col("tiers_json", "分层清单", "强烈推荐/可尝试/不建议分层（JSON）。", visible=0, display_type="textarea"),
    _col("eligible_count", "可匹配赛道数", "通过硬门槛过滤后的赛道数。", searchable=1, display_type="number"),
    _col("excluded_json", "剔除清单", "被硬门槛剔除的赛道及原因（JSON）。", visible=0, display_type="textarea"),
    _col("report_file", "建议书文件", "《申报策略建议书》HTML 文件路径。", visible=0),
    _col("created_at", "创建时间", "报告创建时间（Unix 秒级，自动生成）。"),
)


def _init_teacher_profiles_metadata(cursor) -> None:
    cursor.execute(_TEACHER_PROFILES_DDL)
    init_updated_at_trigger(cursor, "teacher_profiles")
    upsert_table_metadata(
        cursor,
        table_name="teacher_profiles",
        display_name="教师画像",
        description="教师学术画像：研究标签、成果与职称，用于赛道匹配。",
        grid_help="研究标签是匹配主依据（逗号分隔）；顶栏按钮可录入，行内可编辑。",
        visible=1,
        readonly=0,
        default_sort_json='[{"column":"id","direction":"desc"}]',
        default_visible_columns_json='["teacher_name","department","title","research_tags","sci_q1_count"]',
    )
    seed_column_metadata_batch(
        cursor,
        table_name="teacher_profiles",
        columns=_TEACHER_PROFILES_COLUMNS,
        valid_column_names=[str(c["column_name"]) for c in _TEACHER_PROFILES_COLUMNS],
    )


def _init_guide_tracks_metadata(cursor) -> None:
    cursor.execute(_GUIDE_TRACKS_DDL)
    init_updated_at_trigger(cursor, "guide_tracks")
    upsert_table_metadata(
        cursor,
        table_name="guide_tracks",
        display_name="指南赛道",
        description="当年申报指南的结构化赛道库：需求标签、优先方向与硬门槛。",
        grid_help="按年度管理；需求标签为匹配主依据，硬门槛用于过滤不满足条件的教师。",
        visible=1,
        readonly=0,
        default_sort_json='[{"column":"year","direction":"desc"},{"column":"id","direction":"asc"}]',
        default_visible_columns_json='["year","program_name","track_name","track_tags","min_title","min_sci_q1"]',
    )
    seed_column_metadata_batch(
        cursor,
        table_name="guide_tracks",
        columns=_GUIDE_TRACKS_COLUMNS,
        valid_column_names=[str(c["column_name"]) for c in _GUIDE_TRACKS_COLUMNS],
    )


def _init_match_reports_metadata(cursor) -> None:
    cursor.execute(_MATCH_REPORTS_DDL)
    cursor.execute(_MATCH_REPORTS_INDEX)
    upsert_table_metadata(
        cursor,
        table_name="match_reports",
        display_name="推荐报告",
        description="每次匹配的 Top 3 赛道、分层清单与剔除原因；可重建建议书。",
        grid_help="推荐仅供参考；顶栏按钮可重建《申报策略建议书》。",
        visible=1,
        readonly=1,
        default_sort_json='[{"column":"id","direction":"desc"}]',
        default_visible_columns_json='["id","season","teacher_name","department","eligible_count"]',
    )
    seed_column_metadata_batch(
        cursor,
        table_name="match_reports",
        columns=_MATCH_REPORTS_COLUMNS,
        valid_column_names=[str(c["column_name"]) for c in _MATCH_REPORTS_COLUMNS],
    )


def init_db() -> dict[str, int]:
    """幂等创建/迁移本地 SQLite 与数据管理元数据。"""
    conn = get_conn()
    try:
        cur = conn.cursor()
        init_display_metadata(cur)
        _init_teacher_profiles_metadata(cur)
        _init_guide_tracks_metadata(cur)
        _init_match_reports_metadata(cur)
        conn.commit()
        return {}
    finally:
        conn.close()