Spaces:
Running on Zero
Running on Zero
| import unittest | |
| from evaluation.rouge import RougeScorer | |
| from evaluation.bleu import BleuScorer | |
| from evaluation.metrics_manager import MetricsManager | |
| class TestEvaluation(unittest.TestCase): | |
| def setUp(self): | |
| self.rouge = RougeScorer() | |
| self.bleu = BleuScorer() | |
| self.manager = MetricsManager() | |
| def test_rouge_calculation(self): | |
| candidate = "natural language processing enables automatic text summarization" | |
| reference = "natural language processing creates automatic text summaries" | |
| scores = self.rouge.evaluate(candidate, reference) | |
| self.assertIn("rouge-1", scores) | |
| self.assertIn("rouge-2", scores) | |
| self.assertIn("rouge-l", scores) | |
| # ROUGE-1 should have positive precision, recall, and f1 | |
| self.assertGreater(scores["rouge-1"]["f1"], 0.4) | |
| self.assertGreater(scores["rouge-l"]["f1"], 0.4) | |
| def test_bleu_calculation(self): | |
| candidate = "the cat is sitting on the mat" | |
| reference = "the cat sits on the mat" | |
| scores = self.bleu.evaluate(candidate, reference) | |
| self.assertIn("bleu-1", scores) | |
| self.assertIn("bleu_cumulative", scores) | |
| self.assertGreater(scores["bleu-1"], 0.5) | |
| def test_metrics_manager(self): | |
| orig = "Natural language processing is an essential field. It enables machines to comprehend documents and produce concise summaries." | |
| summary = "NLP enables machines to produce summaries." | |
| ref = "NLP helps computers produce summaries." | |
| results = self.manager.evaluate_summary(orig, summary, reference_summary=ref, lang="en") | |
| self.assertIn("compression_ratio", results) | |
| self.assertIn("rouge", results) | |
| self.assertIn("bleu", results) | |
| self.assertLess(results["compression_ratio"], 1.0) | |
| if __name__ == "__main__": | |
| unittest.main() | |