PacketStation/tests/server/test_profiles_import.py

"""The profile CSV import: preview, every error, apply refused with errors, apply clean, the audit rows."""

from __future__ import annotations

from typing import Any

import pytest
from fakedir import FakeDirectory, FakeUser
from fastapi.testclient import TestClient
from pshelpers import app_conn, needs_pg, sign_in

from packetstation.domain import profiles
from packetstation.settings import Settings

pg = needs_pg

PEOPLE = ("weld.andy", "lead.lou", "sup.wes", "dup.dave", "dup.dina", "wh.wade", "qc.quinn")


def _import(client: TestClient, text: str, apply: bool = False) -> Any:
    return client.post("/api/admin/users/import", json={"csv": text, "apply": apply})


def _profiles(settings: Settings) -> dict[str, tuple[Any, ...]]:
    with app_conn(settings) as conn:
        rows = conn.execute(
            "SELECT u.sam_account, u.welder_stamp, u.profile_active,"
            " (SELECT array_agg(s.dept_code || ':' || s.name ORDER BY s.dept_code, s.name)"
            "  FROM user_station_limits l JOIN stations s ON s.id = l.station_id WHERE l.user_id = u.id)"
            " FROM users u"
        ).fetchall()
    return {r[0]: (r[1], r[2], r[3] or []) for r in rows}


def test_parse_csv_shapes() -> None:
    cols, rows = profiles.parse_csv("Employee_ID , Welder_Stamp\r\n0042,A1\r\n\r\n,\r\n7A,\r\n")
    assert cols == ["employee_id", "welder_stamp"]
    assert rows == [(2, {"employee_id": "0042", "welder_stamp": "A1"}), (5, {"employee_id": "7A", "welder_stamp": ""})]
    for text, code in (
        ("", "bad_csv"),
        ("welder_stamp\nA1\n", "bad_csv"),
        ("employee_id,colour\n1,red\n", "bad_csv"),
        ("employee_id,employee_id\n1,1\n", "bad_csv"),
        ("employee_id\n" + "1\n" * (profiles.CSV_MAX_ROWS + 1), "bad_csv"),
        ("employee_id\n" + "x" * profiles.CSV_MAX_BYTES, "too_large"),
    ):
        with pytest.raises(Exception) as e:
            profiles.parse_csv(text)
        assert getattr(e.value, "code", None) == code, text[:40]


@pg
def test_import_preview_errors_and_apply(pg_client: TestClient, pg_settings: Settings, fake_dir: FakeDirectory) -> None:
    # An employee ID with a leading zero, kept as text
    fake_dir.users.append(FakeUser("zero.zed", "Zed Zero", "0042", ["PS-04-SouthWeld-Worker"]))
    fake_dir.users.append(FakeUser("extra.ezra", "Ezra Extra", "7777", ["PS-03-WeldParts-Worker"]))
    for sam in (*PEOPLE, "zero.zed", "extra.ezra"):
        assert sign_in(pg_client, sam).status_code == 200, sam
    sign_in(pg_client, "admin.ann")
    with app_conn(pg_settings) as conn:
        conn.execute("UPDATE users SET welder_stamp = 'W9' WHERE sam_account = 'sup.wes'")
    before = _profiles(pg_settings)

    bad = (
        "employee_id,welder_stamp,active,stations\n"
        "3003,A1,yes,03:Stall 1;03:stall 2\n"  # 2 ok
        "3003,A2,,\n"  # 3 repeated in the file
        "5555,,,\n"  # 4 unknown employee ID
        "9001,X1,,\n"  # 5 on more than one user
        "3002,bad stamp!,,\n"  # 6 bad stamp
        "6001,W9,,\n"  # 7 stamp taken (sup.wes)
        "5001,,maybe,\n"  # 8 bad active
        "0042,,,99:Stall 1\n"  # 9 unknown department
        "0042x,,,\n"  # 10 unknown (not the same as 0042)
        "42,,,\n"  # 11 unknown: 42 is not 0042
        ",A3,,\n"  # 12 no employee ID
        "2003,,,03:Stall 99\n"  # 13 unknown station
        "7777,A1,,\n"  # 14 A1 already claimed by line 2 in this file
    )
    r = _import(pg_client, bad)
    assert r.status_code == 200, r.text
    body = r.json()
    assert body["applied"] is False and body["ok"] == 1 and body["errors"] == 12
    rows = {x["line"]: x for x in body["rows"]}
    assert rows[2]["errors"] == [] and rows[2]["display_name"] == "Andy Welder"
    assert rows[2]["changes"] == {
        "welder_stamp": {"from": None, "to": "A1"},
        "station_limits": {"from": [], "to": ["03:Stall 1", "03:Stall 2"]},
    }
    expect = {
        3: "repeated in the file (first on line 2)",
        4: "Unknown employee ID 5555",
        5: "on more than one user",
        6: "Bad stamp",
        7: "Stamp W9 is taken by Wes Super",
        8: "Bad active value maybe",
        9: "Unknown department 99",
        10: "Unknown employee ID 0042x",
        11: "Unknown employee ID 42",
        12: "Employee ID is missing",
        13: "Unknown station Stall 99 in department 03",
        14: "Stamp A1 is taken by Andy Welder",
    }
    for line, text in expect.items():
        assert len(rows[line]["errors"]) == 1 and text in rows[line]["errors"][0], (line, rows[line]["errors"])

    # Apply with errors: 422 with the same body, nothing changed
    r = _import(pg_client, bad, apply=True)
    assert r.status_code == 422 and r.json()["code"] == "import_errors"
    assert r.json()["rows"] == body["rows"] and r.json()["applied"] is False
    assert _profiles(pg_settings) == before

    # A clean file: a stamp moves from Wes to Lou, Wes gets a new one, limits set and cleared, a profile goes off
    with app_conn(pg_settings) as conn:
        conn.execute(
            "INSERT INTO user_station_limits (user_id, station_id) SELECT u.id, s.id FROM users u, stations s"
            " WHERE u.sam_account = 'wh.wade' AND s.dept_code = '01' AND s.name = 'Station 1'"
        )
    clean = (
        "Employee_ID,Welder_Stamp,Active,Stations\n"
        "3003,A1,yes,03:Stall 1;03:stall 2\n"
        "3002,W9,,\n"
        "2003,W10,,\n"
        "6001,,no,-\n"
        "0042,-,,04:Stall 3\n"
        "5001,,,\n"
    )
    preview = _import(pg_client, clean).json()
    assert preview["errors"] == 0 and preview["ok"] == 6, preview
    assert {x["line"]: x["changes"] for x in preview["rows"]}[7] == {}
    r = _import(pg_client, clean, apply=True)
    assert r.status_code == 200, r.text
    assert r.json()["applied"] is True
    after = _profiles(pg_settings)
    assert after["weld.andy"] == ("A1", True, ["03:Stall 1", "03:Stall 2"])
    assert after["lead.lou"][0] == "W9" and after["sup.wes"][0] == "W10"
    assert after["wh.wade"] == (None, False, [])
    assert after["zero.zed"] == (None, True, ["04:Stall 3"])
    assert after["qc.quinn"] == before["qc.quinn"]

    with app_conn(pg_settings) as conn:
        acts = conn.execute(
            "SELECT action, subject_id, detail FROM audit_events WHERE action LIKE 'admin.profile%' ORDER BY id"
        ).fetchall()
    assert [a[0] for a in acts] == ["admin.profile_updated"] * 5 + ["admin.profile_import"]
    assert acts[-1][2] == {"rows": 6, "changed": 5, "unchanged": 1}
    assert all(a[2]["source"] == "import" for a in acts[:5])
    wade = next(a for a in acts if a[0] == "admin.profile_updated" and a[2]["after"]["profile_active"] is False)
    assert wade[2]["before"]["profile_active"] is True and wade[2]["before"]["station_limits"] != {}

    # Over 1 MB is refused before it is read
    r = _import(pg_client, "employee_id\n" + "1" * (profiles.CSV_MAX_BYTES + 10))
    assert r.status_code == 413