Refresh papers and define Agent evaluation

2026-07-27 16:28:23 +08:00
parent 8e4ff3779b
commit df475e8d90
180 changed files with 31313 additions and 160 deletions
@@ -0,0 +1,62 @@
+from __future__ import annotations
+
+import unittest
+
+from tools.collection.collect_arxiv import ArxivSearchParser, classify
+
+
+SEARCH_RESULT = """
+<ol>
+  <li class="arxiv-result">
+    <p class="list-title is-inline-block">
+      <a href="https://arxiv.org/abs/2607.12345v2">arXiv:2607.12345</a>
+    </p>
+    <p class="title is-5 mathjax">
+      A Trace-Grounded Benchmark for Tool-Using Agents
+    </p>
+    <p class="authors">
+      <span>Authors:</span>
+      <a>Alice Example</a>, <a>Bob Example</a>
+    </p>
+    <div class="tags">
+      <span class="tag is-small is-link tooltip">cs.AI</span>
+      <span class="tag is-small is-link tooltip">cs.LG</span>
+    </div>
+    <span class="abstract-full has-text-grey-dark mathjax">
+      We evaluate an LLM agent with tools and deterministic verification.
+      <a>Less</a>
+    </span>
+    <p class="is-size-7">Submitted 24 July, 2026; originally announced July 2026.</p>
+  </li>
+</ol>
+"""
+
+
+class ArxivSearchParserTest(unittest.TestCase):
+    def test_parses_search_result_into_api_compatible_record(self) -> None:
+        parser = ArxivSearchParser()
+        parser.feed(SEARCH_RESULT)
+
+        self.assertEqual(len(parser.records), 1)
+        record = parser.records[0]
+        self.assertEqual(record["id"], "2607.12345")
+        self.assertEqual(record["url"], "https://arxiv.org/abs/2607.12345")
+        self.assertEqual(record["published"], "2026-07-24")
+        self.assertEqual(record["updated"], "2026-07-24")
+        self.assertEqual(record["authors"], ["Alice Example", "Bob Example"])
+        self.assertEqual(record["categories"], ["cs.AI", "cs.LG"])
+        self.assertNotIn("Less", record["summary"])
+
+    def test_parsed_record_uses_existing_classifier(self) -> None:
+        parser = ArxivSearchParser()
+        parser.feed(SEARCH_RESULT)
+        topics, score, relevance = classify(parser.records[0], {"agent-evaluation", "tool-use"})
+
+        self.assertIn("agent-evaluation", topics)
+        self.assertIn("tool-use", topics)
+        self.assertGreaterEqual(score, 7)
+        self.assertIn(relevance, {"medium", "high"})
+
+
+if __name__ == "__main__":
+    unittest.main()