sinonym 0.2.8__tar.gz → 0.2.9__tar.gz

{sinonym-0.2.8 → sinonym-0.2.9}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sinonym
-Version: 0.2.8
+Version: 0.2.9
 Summary: Chinese Name Detection and Normalization Module
 Project-URL: Homepage, https://github.com/allenai/sinonym
 Project-URL: Repository, https://github.com/allenai/sinonym

{sinonym-0.2.8 → sinonym-0.2.9}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "sinonym"
-version = "0.2.8"
+version = "0.2.9"
 description = "Chinese Name Detection and Normalization Module"
 readme = "README.md"
 requires-python = ">=3.10"

sinonym-0.2.9/sinonym/timo/config.yaml

@@ -0,0 +1,22 @@
+config_version: 0.0.1
+model_variants:
+  sinonym_v1:
+    instance: sinonym.timo.interface.Instance
+    prediction: sinonym.timo.interface.Prediction
+    predictor: sinonym.timo.interface.Predictor
+    predictor_config: sinonym.timo.interface.PredictorConfig
+    artifacts_s3_path: null
+    python_version: "3.10"
+    cuda: False
+    integration_test: sinonym.timo.integration_test.TestIntegration
+    docker_run_commands: []
+  sinonym_routing_v1:
+    instance: sinonym.timo.interface.RoutingInstance
+    prediction: sinonym.timo.interface.RoutedPaperPrediction
+    predictor: sinonym.timo.interface.RoutingPredictor
+    predictor_config: sinonym.timo.interface.PredictorConfig
+    artifacts_s3_path: null
+    python_version: "3.10"
+    cuda: False
+    integration_test: sinonym.timo.integration_test.TestRoutingIntegration
+    docker_run_commands: []

{sinonym-0.2.8 → sinonym-0.2.9}/sinonym/timo/integration_test.py

@@ -1,6 +1,15 @@
 import unittest
 
-from sinonym.timo.interface import Instance, Prediction, Predictor, PredictorConfig
+from sinonym.timo.interface import (
+    Instance,
+    Prediction,
+    Predictor,
+    PredictorConfig,
+    RoutedPaperPrediction,
+    RoutedPrediction,
+    RoutingInstance,
+    RoutingPredictor,
+)
 
 
 class TestIntegration(unittest.TestCase):
@@ -67,3 +76,49 @@ class TestIntegration(unittest.TestCase):
         self.assertEqual(len(result.name_order_evidence), 2)
         self.assertEqual(result.name_order_evidence[0].raw_name, "Li Wei")
         self.assertEqual(result.name_order_evidence[0].selected_format, "surname_first")
+
+
+class TestRoutingIntegration(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls):
+        cls.predictor = RoutingPredictor(config=PredictorConfig(), artifacts_dir=".")
+
+    def test_route_with_vys_pool(self):
+        # paper authors must be the leading slice of the pool
+        pp = ["Yue Lin", "Wei Wang"]
+        pool = pp + ["Jun Zhao", "Hui Li"]
+        results = self.predictor.predict_batch([RoutingInstance(pp_names=pp, vys_pool_names=pool)])
+        self.assertEqual(len(results), 1)  # one prediction per instance (timo 1:1 contract)
+        paper = results[0]
+        self.assertIsInstance(paper, RoutedPaperPrediction)
+        self.assertEqual(len(paper.authors), 2)  # one RoutedPrediction per pp author
+        for r in paper.authors:
+            self.assertIsInstance(r, RoutedPrediction)
+            self.assertIn(r.router_prediction, {"pp", "vys", "abstain", "not_person"})
+            self.assertIsNotNone(r.vys)  # vys candidate present when a pool is given
+
+    def test_route_pp_only_fallback(self):
+        results = self.predictor.predict_batch([RoutingInstance(pp_names=["Li Wei"])])
+        self.assertEqual(len(results), 1)
+        authors = results[0].authors
+        self.assertEqual(len(authors), 1)
+        self.assertIsNone(authors[0].vys)  # PP-only fallback: no venue pool
+        self.assertIsNone(authors[0].input_order_candidate)
+
+    def test_predict_batch_is_one_to_one(self):
+        instances = [
+            RoutingInstance(pp_names=["Li Wei", "Wang Weiming"]),
+            RoutingInstance(pp_names=["Zhang San"]),
+        ]
+        results = self.predictor.predict_batch(instances)
+        self.assertEqual(len(results), len(instances))  # 1:1, paper boundaries preserved
+        self.assertEqual([len(p.authors) for p in results], [2, 1])
+
+    def test_empty_paper_still_emits_one_prediction(self):
+        # an instance with no authors must not vanish from the output stream
+        results = self.predictor.predict_batch([RoutingInstance(pp_names=[])])
+        self.assertEqual(len(results), 1)
+        self.assertEqual(results[0].authors, [])
+
+    def test_predict_batch_empty(self):
+        self.assertEqual(self.predictor.predict_batch([]), [])