Merge pull request #10345 from adrianeboyd/chore/v3.1-backport-10324

Fix Tok2Vec for empty batches (#10324)

azure-pipelines.yml

@@ -23,7 +23,7 @@ jobs:
   # defined in .flake8 and overwrites the selected codes.
   - job: "Validate"
     pool:
-      vmImage: "ubuntu-18.04"
+      vmImage: "ubuntu-latest"
     steps:
       - task: UsePythonVersion@0
         inputs:
@@ -39,49 +39,49 @@ jobs:
       matrix:
         # We're only running one platform per Python version to speed up builds
         Python36Linux:
-          imageName: "ubuntu-18.04"
+          imageName: "ubuntu-latest"
           python.version: "3.6"
         #        Python36Windows:
-        #          imageName: "windows-2019"
+        #          imageName: "windows-latest"
         #          python.version: "3.6"
         #        Python36Mac:
-        #          imageName: "macos-10.14"
+        #          imageName: "macos-latest"
         #          python.version: "3.6"
         #        Python37Linux:
-        #          imageName: "ubuntu-18.04"
+        #          imageName: "ubuntu-latest"
         #          python.version: "3.7"
         Python37Windows:
-          imageName: "windows-2019"
+          imageName: "windows-latest"
           python.version: "3.7"
         #        Python37Mac:
-        #          imageName: "macos-10.14"
+        #          imageName: "macos-latest"
         #          python.version: "3.7"
         #        Python38Linux:
-        #          imageName: "ubuntu-18.04"
+        #          imageName: "ubuntu-latest"
         #          python.version: "3.8"
         #        Python38Windows:
-        #          imageName: "windows-2019"
+        #          imageName: "windows-latest"
         #          python.version: "3.8"
         Python38Mac:
-          imageName: "macos-10.14"
+          imageName: "macos-latest"
           python.version: "3.8"
         Python39Linux:
-          imageName: "ubuntu-18.04"
+          imageName: "ubuntu-latest"
           python.version: "3.9"
         #        Python39Windows:
-        #          imageName: "windows-2019"
+        #          imageName: "windows-latest"
         #          python.version: "3.9"
         #        Python39Mac:
-        #          imageName: "macos-10.14"
+        #          imageName: "macos-latest"
         #          python.version: "3.9"
         Python310Linux:
-          imageName: "ubuntu-20.04"
+          imageName: "ubuntu-latest"
           python.version: "3.10"
         Python310Windows:
-          imageName: "windows-2019"
+          imageName: "windows-latest"
           python.version: "3.10"
         Python310Mac:
-          imageName: "macos-10.15"
+          imageName: "macos-latest"
           python.version: "3.10"
       maxParallel: 4
     pool:

requirements.txt

@@ -29,7 +29,7 @@ pytest-timeout>=1.3.0,<2.0.0
 mock>=2.0.0,<3.0.0
 flake8>=3.8.0,<3.10.0
 hypothesis>=3.27.0,<7.0.0
-mypy>=0.910
+mypy==0.910
 types-dataclasses>=0.1.3; python_version < "3.7"
 types-mock>=0.1.1
 types-requests

spacy/lexeme.pyi

@@ -19,7 +19,7 @@ class Lexeme:
     @property
     def vector_norm(self) -> float: ...
     vector: Floats1d
-    rank: str
+    rank: int
     sentiment: float
     @property
     def orth_(self) -> str: ...

spacy/pipeline/tok2vec.py

@@ -118,6 +118,10 @@ class Tok2Vec(TrainablePipe):
 
         DOCS: https://spacy.io/api/tok2vec#predict
         """
+        if not any(len(doc) for doc in docs):
+            # Handle cases where there are no tokens in any docs.
+            width = self.model.get_dim("nO")
+            return [self.model.ops.alloc((0, width)) for doc in docs]
         tokvecs = self.model.predict(docs)
         batch_id = Tok2VecListener.get_batch_id(docs)
         for listener in self.listeners:

spacy/tests/pipeline/test_tok2vec.py

@@ -11,7 +11,7 @@ from spacy.lang.en import English
 from thinc.api import Config, get_current_ops
 from numpy.testing import assert_array_equal
 
-from ..util import get_batch, make_tempdir
+from ..util import get_batch, make_tempdir, add_vecs_to_vocab
 
 
 def test_empty_doc():
@@ -140,9 +140,25 @@ TRAIN_DATA = [
 ]
 
 
-def test_tok2vec_listener():
+@pytest.mark.parametrize("with_vectors", (False, True))
+def test_tok2vec_listener(with_vectors):
     orig_config = Config().from_str(cfg_string)
+    orig_config["components"]["tok2vec"]["model"]["embed"][
+        "include_static_vectors"
+    ] = with_vectors
     nlp = util.load_model_from_config(orig_config, auto_fill=True, validate=True)
+
+    if with_vectors:
+        ops = get_current_ops()
+        vectors = [
+            ("apple", ops.asarray([1, 2, 3])),
+            ("orange", ops.asarray([-1, -2, -3])),
+            ("and", ops.asarray([-1, -1, -1])),
+            ("juice", ops.asarray([5, 5, 10])),
+            ("pie", ops.asarray([7, 6.3, 8.9])),
+        ]
+        add_vecs_to_vocab(nlp.vocab, vectors)
+
     assert nlp.pipe_names == ["tok2vec", "tagger"]
     tagger = nlp.get_pipe("tagger")
     tok2vec = nlp.get_pipe("tok2vec")
@@ -169,6 +185,9 @@ def test_tok2vec_listener():
     ops = get_current_ops()
     assert_array_equal(ops.to_numpy(doc.tensor), ops.to_numpy(doc_tensor))
 
+    # test with empty doc
+    doc = nlp("")
+
     # TODO: should this warn or error?
     nlp.select_pipes(disable="tok2vec")
     assert nlp.pipe_names == ["tagger"]