127 lines
4.7 KiB
Python
127 lines
4.7 KiB
Python
import pytest
|
||
|
||
from app.workers.citation_engine import BrandMatcher, CompetitorDetector
|
||
|
||
|
||
def test_brand_matcher_exact():
|
||
matcher = BrandMatcher(target_brand="华为", brand_aliases=["Huawei"])
|
||
result = matcher.match("华为是一家伟大的科技公司")
|
||
assert result["cited"] is True
|
||
assert result["match_type"] == "exact"
|
||
assert result["confidence"] == 1.0
|
||
|
||
|
||
def test_brand_matcher_alias():
|
||
matcher = BrandMatcher(target_brand="华为", brand_aliases=["Huawei"])
|
||
result = matcher.match("Huawei makes great phones")
|
||
assert result["cited"] is True
|
||
assert result["match_type"] == "alias"
|
||
assert result["confidence"] == 0.9
|
||
|
||
|
||
def test_brand_matcher_fuzzy():
|
||
matcher = BrandMatcher(target_brand="华为")
|
||
# "华伟" is a fuzzy match to "华为"
|
||
result = matcher.match("华伟 是一家科技公司")
|
||
assert result["cited"] is True
|
||
assert result["match_type"] == "fuzzy"
|
||
assert result["confidence"] > 0.4
|
||
|
||
|
||
def test_brand_matcher_no_match():
|
||
matcher = BrandMatcher(target_brand="华为")
|
||
result = matcher.match("这是一段完全不相关的文本，没有任何品牌信息")
|
||
assert result["cited"] is False
|
||
assert result["match_type"] is None
|
||
assert result["confidence"] == 0.0
|
||
|
||
|
||
def test_competitor_detector():
|
||
detector = CompetitorDetector()
|
||
text = "中国平安和中国人寿都是大型保险公司"
|
||
competitors = detector.detect(text, target_brand="中国平安")
|
||
assert "中国人寿" in competitors
|
||
assert "中国平安" not in competitors
|
||
|
||
|
||
def test_citation_position():
|
||
matcher = BrandMatcher(target_brand="华为")
|
||
text = "第一段介绍市场情况\n第二段提到华为的产品\n第三段是总结"
|
||
result = matcher.match(text)
|
||
assert result["cited"] is True
|
||
assert result["position"] == 2
|
||
assert result["citation_text"] is not None
|
||
|
||
|
||
# ---------------------------------------------------------------------------
|
||
# 补充测试
|
||
# ---------------------------------------------------------------------------
|
||
def test_brand_matcher_multiple_aliases():
|
||
"""多个别名时，应能匹配任意一个别名。"""
|
||
matcher = BrandMatcher(target_brand="华为", brand_aliases=["Huawei", "HW", "Honor"])
|
||
# 匹配第二个别名
|
||
result = matcher.match("HW released a new chip")
|
||
assert result["cited"] is True
|
||
assert result["match_type"] == "alias"
|
||
assert result["confidence"] == 0.9
|
||
# 匹配第三个别名
|
||
result2 = matcher.match("Honor phones are popular")
|
||
assert result2["cited"] is True
|
||
assert result2["match_type"] == "alias"
|
||
|
||
|
||
def test_brand_matcher_fuzzy_threshold_boundary():
|
||
"""模糊匹配阈值边界：ratio 恰好在 0.4 附近的情况。"""
|
||
matcher = BrandMatcher(target_brand="华为")
|
||
# "华伟" 与 "华为" 的 ratio 约为 0.5，应大于 0.4
|
||
result = matcher.match("华伟 是一家科技公司")
|
||
assert result["cited"] is True
|
||
assert result["match_type"] == "fuzzy"
|
||
assert result["confidence"] > 0.4
|
||
|
||
# "苹果" vs "华为" ratio 很低，不应超过 0.4
|
||
result2 = matcher.match("苹果科技公司")
|
||
assert result2["cited"] is False
|
||
assert result2["match_type"] is None
|
||
|
||
|
||
def test_brand_matcher_empty_text():
|
||
"""空字符串输入不应崩溃，应返回 cited=False。"""
|
||
matcher = BrandMatcher(target_brand="华为", brand_aliases=["Huawei"])
|
||
result = matcher.match("")
|
||
assert result["cited"] is False
|
||
assert result["confidence"] == 0.0
|
||
assert result["match_type"] is None
|
||
assert result["position"] is None
|
||
|
||
|
||
def test_competitor_detector_multi_industry():
|
||
"""竞品检测应能跨行业识别品牌。"""
|
||
detector = CompetitorDetector()
|
||
text = "华为和小米是科技公司，工商银行和招商银行是银行"
|
||
competitors_tech = detector.detect(text, target_brand="华为")
|
||
assert "小米" in competitors_tech
|
||
assert "腾讯" not in competitors_tech # 未在文本中出现
|
||
assert "华为" not in competitors_tech
|
||
|
||
competitors_finance = detector.detect(text, target_brand="工商银行")
|
||
assert "招商银行" in competitors_finance
|
||
assert "建设银行" not in competitors_finance # 未在文本中出现
|
||
|
||
|
||
def test_citation_position_multiple_paragraphs():
|
||
"""品牌在不同段落位置出现时的 position 检测。"""
|
||
matcher = BrandMatcher(target_brand="华为")
|
||
|
||
text_first = "华为位于第一段\n第二段没有\n第三段也没有"
|
||
result = matcher.match(text_first)
|
||
assert result["position"] == 1
|
||
|
||
text_third = "第一段没有\n第二段没有\n华为在第三段"
|
||
result = matcher.match(text_third)
|
||
assert result["position"] == 3
|
||
|
||
text_last = "第一段\n第二段\n第三段\n最后一段提到华为"
|
||
result = matcher.match(text_last)
|
||
assert result["position"] == 4
|