feat: DB管线补全 + 新增orchestrator_db/program_schema/to_sql + 清理临时脚本

2026-07-11 14:55:52 +08:00
parent 40e8a50ab4
commit af37e33b98
32 changed files with 3232 additions and 255 deletions
@@ -34,6 +34,12 @@ class Config:
    gcov_threshold: float = 0.5
    max_quality_retries: int = 4

+    # gixsql for DB programs
+    gixsql_path: str = "gixsql/bin/gixpp.exe"
+    gixsql_lib_path: str = "gixsql/lib"
+    gixsql_db_path: str = ".db/gixsql"
+    gixsql_compile_flags: str = "-fixed -ext cpy --coverage"
+
    @classmethod
    def from_toml(cls, path="aurak.toml"):
        import tomllib
@@ -56,6 +62,11 @@ class Config:
        c.tolerance = cp.get("default_tolerance", c.tolerance)
        r = d.get("runner", {})
        c.runner_mode = r.get("mode", "native")
+        g = d.get("gixsql", {})
+        c.gixsql_path = g.get("path", c.gixsql_path)
+        c.gixsql_lib_path = g.get("lib_path", c.gixsql_lib_path)
+        c.gixsql_db_path = g.get("db_path", c.gixsql_db_path)
+        c.gixsql_compile_flags = g.get("compile_flags", c.gixsql_compile_flags)
        s = d.get("spark", {})
        c.spark_master = s.get("master", "local[*]")
        c.num_records = s.get("num_records", c.num_records)
@@ -0,0 +1,64 @@
+"""Program schema — per-program DB table definitions + subprogram list."""
+
+from __future__ import annotations
+from dataclasses import dataclass, field
+from pathlib import Path
+from typing import Optional
+
+
+@dataclass
+class ColumnDef:
+    name: str
+    type: str           # SQL type: "CHAR(6)", "NUMERIC(4)", "VARCHAR(30)"
+    primary_key: bool = False
+    nullable: bool = False
+    default: Optional[str] = None
+    cobol_field: Optional[str] = None   # COBOL field name if different
+
+
+@dataclass
+class TableDef:
+    name: str
+    columns: list[ColumnDef] = field(default_factory=list)
+    create_if_missing: bool = True
+    sql_name: Optional[str] = None   # COBOL SQL table name if different from YAML name
+
+
+@dataclass
+class ProgramSchema:
+    program_id: str
+    db_tables: list[TableDef] = field(default_factory=list)
+    subprograms: list[str] = field(default_factory=list)
+    db_type: str = "SQLite"
+    db_name: str = "OVERTIME.DB"
+
+    @classmethod
+    def from_yaml(cls, path: str | Path) -> ProgramSchema:
+        import yaml
+        with open(path, encoding="utf-8") as f:
+            raw = yaml.safe_load(f)
+        tables = []
+        for t in raw.get("db_tables", []):
+            cols = [ColumnDef(**c) for c in t.get("columns", [])]
+            tables.append(TableDef(
+                name=t["name"], columns=cols,
+                sql_name=t.get("sql_name"),
+            ))
+        return cls(
+            program_id=raw["program_id"],
+            db_tables=tables,
+            subprograms=raw.get("subprograms", []),
+            db_type=raw.get("db_type", "SQLite"),
+            db_name=raw.get("db_name", "OVERTIME.DB"),
+        )
+
+
+def load_schema(program_id: str, search_dirs: list[str | Path] | None = None) -> ProgramSchema:
+    """Load per-program YAML schema by program ID."""
+    if search_dirs is None:
+        search_dirs = [Path(__file__).parent / "programs"]
+    for d in search_dirs:
+        p = Path(d) / f"{program_id}.yaml"
+        if p.exists():
+            return ProgramSchema.from_yaml(p)
+    raise FileNotFoundError(f"Schema not found for {program_id} in {search_dirs}")
@@ -0,0 +1,28 @@
+program_id: KIN02UPD
+db_type: SQLite
+db_name: kin.db
+
+db_tables:
+  - name: LEAVE_RECORDS
+    columns:
+      - name: APPLICATION_ID
+        type: INTEGER
+        primary_key: true
+      - name: EMP_ID
+        type: CHAR(8)
+      - name: LEAVE_TYPE
+        type: CHAR(2)
+      - name: START_DATE
+        type: CHAR(8)
+      - name: START_TIME
+        type: CHAR(4)
+      - name: END_DATE
+        type: CHAR(8)
+      - name: END_TIME
+        type: CHAR(4)
+      - name: STATUS
+        type: CHAR(1)
+
+subprograms:
+  - SUB02MSG
+  - SUB03END
@@ -0,0 +1,37 @@
+program_id: KIN03EXP
+db_type: SQLite
+db_name: kin.db
+
+db_tables:
+  - name: LEAVE_RECORDS
+    columns:
+      - name: APPLICATION_ID
+        type: INTEGER
+        primary_key: true
+      - name: EMP_ID
+        type: CHAR(8)
+      - name: LEAVE_TYPE
+        type: CHAR(2)
+      - name: START_DATE
+        type: CHAR(8)
+      - name: START_TIME
+        type: CHAR(4)
+      - name: END_DATE
+        type: CHAR(8)
+      - name: END_TIME
+        type: CHAR(4)
+      - name: STATUS
+        type: CHAR(1)
+
+  - name: HOLIDAY_CALENDAR
+    columns:
+      - name: HOLIDAY_DATE
+        type: CHAR(8)
+        primary_key: true
+      - name: DESCRIPTION
+        type: VARCHAR(50)
+
+subprograms:
+  - SUB01DAT
+  - SUB02MSG
+  - SUB03END
@@ -0,0 +1,28 @@
+program_id: KIN06CLD
+db_type: SQLite
+db_name: kin.db
+
+db_tables:
+  - name: HOLIDAY_CALENDAR
+    columns:
+      - name: HOLIDAY_DATE
+        type: CHAR(8)
+        primary_key: true
+      - name: DESCRIPTION
+        type: VARCHAR(50)
+
+  - name: EMP_MASTER
+    columns:
+      - name: EMP_ID
+        type: CHAR(8)
+        primary_key: true
+      - name: DEPT_ID
+        type: CHAR(4)
+      - name: EMP_NAME
+        type: VARCHAR(50)
+      - name: STATUS
+        type: CHAR(1)
+
+subprograms:
+  - SUB02MSG
+  - SUB03END
@@ -0,0 +1,54 @@
+program_id: KIN08DBU
+db_type: SQLite
+db_name: kin.db
+
+db_tables:
+  - name: DAILY_RECORDS
+    columns:
+      - name: EMP_ID
+        type: CHAR(8)
+        primary_key: true
+      - name: TARGET_DATE
+        type: CHAR(8)
+        primary_key: true
+      - name: TIME_IN
+        type: CHAR(4)
+      - name: TIME_OUT
+        type: CHAR(4)
+      - name: ANNUAL_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: PERSONAL_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: OFFICIAL_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: SICK_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: UNAPPROVED_ABSENT_H
+        type: DECIMAL(6,1)
+      - name: UPDATED_AT
+        type: TIMESTAMP
+
+  - name: MONTHLY_ABSENCE
+    columns:
+      - name: EMP_ID
+        type: CHAR(8)
+        primary_key: true
+      - name: YEAR_MONTH
+        type: CHAR(6)
+        primary_key: true
+      - name: ANNUAL_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: PERSONAL_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: OFFICIAL_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: SICK_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: UNAPPROVED_ABSENT_H
+        type: DECIMAL(6,1)
+      - name: UPDATED_AT
+        type: TIMESTAMP
+
+subprograms:
+  - SUB02MSG
+  - SUB03END
@@ -0,0 +1,54 @@
+program_id: KIN09CSV
+db_type: SQLite
+db_name: kin.db
+
+db_tables:
+  - name: DAILY_RECORDS
+    columns:
+      - name: EMP_ID
+        type: CHAR(8)
+        primary_key: true
+      - name: TARGET_DATE
+        type: CHAR(8)
+        primary_key: true
+      - name: TIME_IN
+        type: CHAR(4)
+      - name: TIME_OUT
+        type: CHAR(4)
+      - name: ANNUAL_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: PERSONAL_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: OFFICIAL_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: SICK_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: UNAPPROVED_ABSENT_H
+        type: DECIMAL(6,1)
+      - name: UPDATED_AT
+        type: TIMESTAMP
+
+  - name: MONTHLY_ABSENCE
+    columns:
+      - name: EMP_ID
+        type: CHAR(8)
+        primary_key: true
+      - name: YEAR_MONTH
+        type: CHAR(6)
+        primary_key: true
+      - name: ANNUAL_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: PERSONAL_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: OFFICIAL_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: SICK_LEAVE_H
+        type: DECIMAL(6,1)
+      - name: UNAPPROVED_ABSENT_H
+        type: DECIMAL(6,1)
+      - name: UPDATED_AT
+        type: TIMESTAMP
+
+subprograms:
+  - SUB02MSG
+  - SUB03END
@@ -0,0 +1,46 @@
+program_id: ZAN06UPD
+db_type: SQLite
+
+db_tables:
+  - name: ZANTBL01
+    sql_name: OVT_APPLICATIONS
+    columns:
+      - name: EMPNO
+        type: CHAR(6)
+        primary_key: true
+      - name: WORK_DATE
+        type: CHAR(8)
+        primary_key: true
+      - name: START_TIME
+        type: NUMERIC(4)
+        cobol_field: DB-START-TIME
+      - name: END_TIME
+        type: NUMERIC(4)
+        cobol_field: DB-END-TIME
+      - name: OVERTIME
+        type: NUMERIC(4)
+        cobol_field: DB-OVERTIME
+      - name: APPROVAL_FLAG
+        type: CHAR(1)
+        cobol_field: DB-APPROVAL-FLAG
+      - name: NOTES
+        type: VARCHAR(100)
+        cobol_field: DB-NOTES
+
+  - name: ZANTBL02
+    sql_name: OVT_MONTHLY
+    columns:
+      - name: EMPNO
+        type: CHAR(6)
+        primary_key: true
+      - name: DEPTNO
+        type: CHAR(4)
+        primary_key: true
+      - name: DEPT_NAME
+        type: VARCHAR(30)
+        cobol_field: DB-DEPT-NAME
+
+subprograms:
+  - SUB01DAT
+  - SUB02MSG
+  - SUB03END