tests/test_excipient_segmentation.py

"""辅料连写串词典分词回归测试（skills/compatibility/skill._heuristic_excipients）。

复现真实问题：多个辅料无分隔符连写时只识别其中少数。修复后应按 KB 别名最大匹配
切分为全部辅料，未知段保留为独立（未知）辅料。
"""

import os
import sys

sys.path.insert(0, os.path.join(os.path.dirname(os.path.dirname(os.path.abspath(__file__))), "core"))

from skills.compatibility.skill import _heuristic_excipients  # noqa: E402
from skills.compatibility.knowledge_base import get_default_kb  # noqa: E402


def test_concatenated_excipients_segmented_to_six():
    text = "马铃薯淀粉微晶纤维素聚维酮硬脂酸钙交联羧甲基纤维素钠欧巴代"
    out = _heuristic_excipients(text)
    assert out == [
        "马铃薯淀粉", "微晶纤维素", "聚维酮", "硬脂酸钙",
        "交联羧甲基纤维素钠", "欧巴代",
    ]
    kb = get_default_kb()
    # 全部应解析为知识库已知条目（修复前 硬脂酸钙 / 欧巴代 / 马铃薯淀粉 无法识别）。
    keys = {kb.resolve(n).key for n in out}
    assert keys == {"starch", "mcc", "povidone", "calcium_stearate",
                    "croscarmellose_sodium", "opadry"}
    assert all(kb.resolve(n).known for n in out)


def test_delimited_input_preserved():
    assert _heuristic_excipients("乳糖、硬脂酸镁，微晶纤维素") == ["乳糖", "硬脂酸镁", "微晶纤维素"]


def test_single_known_and_unknown_preserved():
    assert _heuristic_excipients("乳糖") == ["乳糖"]
    assert _heuristic_excipients("完全不存在的辅料QWERTY") == ["完全不存在的辅料QWERTY"]


def test_dedup_repeated_excipients():
    out = _heuristic_excipients("乳糖、乳糖、微晶纤维素")
    assert out == ["乳糖", "微晶纤维素"]


def test_unknown_run_between_known_kept():
    """已知辅料之间夹未知串时，未知串作为独立辅料保留（不丢失）。"""
    out = _heuristic_excipients("微晶纤维素某新型辅料XYZ聚维酮")
    assert "微晶纤维素" in out and "聚维酮" in out
    assert any("某新型辅料" in x for x in out)