fix: TEDS algorithm optimization

SHUzhangshuo · SHUzhangshuo · commit 9a4c7fc5b89c · 2025-09-01T13:46:39.000+08:00
diff --git a/requirements.txt b/requirements.txt
@@ -1,13 +1,5 @@
-rapidFuzz
-setuptools
-jsonlines
-beautifulsoup4
-requests
-torch
-html2text
-resiliparse
-trafilatura
-# llm-web-kit==3.2.0
-https://github.com/opendatalab/magic-html/releases/download/magic_html-0.1.5-released/magic_html-0.1.5-py3-none-any.whl
-streamlit
-markdown
+# WebMainBench dependencies
+rapidfuzz>=3.0.0
+lxml>=4.6.0
+beautifulsoup4>=4.9.0
+apted>=1.2.0
diff --git a/tests/test_teds.py b/tests/test_teds.py
@@ -220,7 +220,7 @@ def test_teds_structure_same_content_different(self):
             groundtruth=gt,
             table_edit_result=self.valid_table_edit_result
         )
-        assert result.score == 0.7999999999999999
+        assert result.score == 0.96
 
 
 
@@ -326,7 +326,7 @@ def test_teds_content_similarity(self):
             table2,
             table_edit_result=self.valid_table_edit_result
         )
-        assert result.score == 0.3999999999999999
+        assert result.score == 0.9318181818181819
 
 
 class TestStructureTEDS(unittest.TestCase):
diff --git a/webmainbench/metrics/teds_metrics.py b/webmainbench/metrics/teds_metrics.py

Original file line number	Diff line number	Diff line change
`@@ -220,7 +220,7 @@ def test_teds_structure_same_content_different(self):`
`220`	`220`	`groundtruth=gt,`
`221`	`221`	`table_edit_result=self.valid_table_edit_result`
`222`	`222`	`)`
`223`		`- assert result.score == 0.7999999999999999`
	`223`	`+ assert result.score == 0.96`
`224`	`224`
`225`	`225`
`226`	`226`
`@@ -326,7 +326,7 @@ def test_teds_content_similarity(self):`
`326`	`326`	`table2,`
`327`	`327`	`table_edit_result=self.valid_table_edit_result`
`328`	`328`	`)`
`329`		`- assert result.score == 0.3999999999999999`
	`329`	`+ assert result.score == 0.9318181818181819`
`330`	`330`
`331`	`331`
`332`	`332`	`class TestStructureTEDS(unittest.TestCase):`