Проверка на данном этапе

tests/test_extract_transform.py

@@ -0,0 +1,118 @@
+import csv
+
+import pytest
+
+from src.extract import CSVFileError, FIELDS, read_records
+from src.transform import MAX_INTEGER, clean_orders
+
+
+def order(**changes):
+    return dict(order_id="1", order_date="2024-02-29", category="Книги",
+                quantity="3", unit_price="0,07") | changes
+
+
+def process(tmp_path, *rows, header=FIELDS):
+    path = tmp_path / "orders.csv"
+    with path.open("w", encoding="utf-8-sig", newline="") as output:
+        writer = csv.writer(output)
+        writer.writerow(header)
+        for row in rows:
+            writer.writerow([row[name] for name in header] if isinstance(row, dict) else row)
+    before = path.read_bytes()
+    accepted, rejected = clean_orders(read_records(path))
+    assert path.read_bytes() == before
+    numbers = [item.record_number for item in accepted]
+    numbers.extend(item.source.record_number for item in rejected)
+    assert sorted(numbers) == list(range(1, len(rows) + 1))
+    return accepted, rejected
+
+
+@pytest.mark.parametrize("value", ["2024-02-29", " 29.02.2024 "])
+def test_normalization_and_exact_money(tmp_path, value):
+    accepted, rejected = process(tmp_path, order(order_date=value, category="  "), header=FIELDS[::-1])
+    assert not rejected
+    item, = accepted
+    assert (item.order_date, item.category) == ("2024-02-29", "Не указана")
+    assert (item.unit_price_kopecks, item.revenue_kopecks) == (7, 21)
+
+
+@pytest.mark.parametrize(("name", "value"), [
+    ("order_date", "2023-02-29"), ("order_date", "1.2.2024"), ("order_date", "20240229"),
+    ("quantity", "1.2"), ("quantity", "0"), ("quantity", "-2"), ("order_id", "0"),
+    ("unit_price", "NaN"), ("unit_price", "Infinity"), ("unit_price", "-Infinity"),
+    ("unit_price", "0"), ("unit_price", "-1"), ("unit_price", "1,230"),
+    ("unit_price", "1 000.20"), ("unit_price", "1e2"),
+    ("unit_price", "92233720368547758.08"), ("quantity", str(MAX_INTEGER + 1)),
+    ("order_id", "9" * 5000),
+])
+def test_invalid_field(tmp_path, name, value):
+    accepted, rejected = process(tmp_path, order(**{name: value}))
+    assert not accepted
+    reason, = rejected[0].reasons
+    assert reason.startswith(name + ":")
+
+
+def test_sqlite_boundary_and_product_overflow(tmp_path):
+    price = "92233720368547758.07"
+    accepted, rejected = process(tmp_path, order(quantity="1", unit_price=price),
+                                 order(quantity="2", unit_price=price))
+    assert accepted[0].revenue_kopecks == MAX_INTEGER
+    assert rejected[0].reasons == ["revenue_kopecks: превышен предел INTEGER SQLite"]
+
+
+def test_multiple_missing_fields_and_exact_repeat(tmp_path):
+    bad = order(order_id=" ", order_date="", quantity="", unit_price="")
+    _, rejected = process(tmp_path, bad, bad)
+    assert rejected[0].source.values == tuple(bad.values())
+    assert [s.split(":")[0] for s in rejected[0].reasons] == [
+        "order_id", "order_date", "quantity", "unit_price"
+    ]
+    assert rejected[1].reasons == ["Точный дубль"]
+
+
+def test_duplicate_stages_and_conflict(tmp_path):
+    first = order(category="  ТоВаРы\n ДЛЯ  ДоМа  ")
+    equivalent = first | {"order_date": "29.02.2024", "order_id": "001",
+                          "category": "товары для дома", "unit_price": "0.07"}
+    conflict = order(order_id="2")
+    accepted, rejected = process(
+        tmp_path, first, first, equivalent, conflict, conflict | {"quantity": "03"},
+        conflict | {"quantity": "4"}, order(quantity="много")
+    )
+    assert [(item.record_number, item.category) for item in accepted] == [(1, "Товары для дома")]
+    assert [(r.source.record_number, r.reasons) for r in rejected] == [
+        (2, ["Точный дубль"]), (3, ["Нормализованный дубль"]),
+        (4, ["Конфликт order_id"]), (5, ["Конфликт order_id"]), (6, ["Конфликт order_id"]),
+        (7, ["quantity: ожидается положительное целое число"]),
+    ]
+
+
+def test_empty_short_and_long_records(tmp_path):
+    accepted, rejected = process(tmp_path, [], ["1"], list(order().values()) + ["лишнее"], order())
+    assert accepted[0].record_number == 4
+    assert [r.reasons for r in rejected] == [
+        ["Пустая запись CSV"], ["Число полей: ожидалось 5, получено 1"],
+        ["Число полей: ожидалось 5, получено 6"],
+    ]
+
+
+def test_header_only(tmp_path):
+    assert process(tmp_path) == ([], [])
+
+
+@pytest.mark.parametrize("header", ["", "order_id,quantity", ",".join((*FIELDS, "extra")),
+                                    ",".join((*FIELDS[:-1], "quantity"))])
+def test_invalid_header(tmp_path, header):
+    path = tmp_path / "orders.csv"
+    path.write_text(header, encoding="utf-8")
+    with pytest.raises(CSVFileError, match="неверный заголовок"):
+        clean_orders(read_records(path))
+
+
+def test_missing_or_damaged_file(tmp_path):
+    path = tmp_path / "orders.csv"
+    with pytest.raises(CSVFileError, match="файл не найден"):
+        clean_orders(read_records(path))
+    path.write_text(",".join(FIELDS) + '\n1,2024-02-29,"незакрытое поле', encoding="utf-8")
+    with pytest.raises(CSVFileError, match="ошибка структуры CSV"):
+        clean_orders(read_records(path))

tests/test_generate_data.py

@@ -0,0 +1,20 @@
+import csv
+
+from src.generate_data import generate_data
+
+
+def test_record_count_and_header(tmp_path):
+    path = tmp_path / "orders.csv"
+    generate_data(path)
+    with path.open(encoding="utf-8", newline="") as source:
+        records = csv.reader(source)
+        assert next(records) == ["order_id", "order_date", "category", "quantity", "unit_price"]
+        assert sum(1 for _ in records) == 1500
+
+
+def test_repeated_generation_is_byte_identical(tmp_path):
+    path = tmp_path / "orders.csv"
+    generate_data(path)
+    first_run = path.read_bytes()
+    generate_data(path)
+    assert path.read_bytes() == first_run