From 88acbfc050a63e4c245f62419212e44d9062e8e1 Mon Sep 17 00:00:00 2001
From: Matthew Honnibal <honnibal+gh@gmail.com>
Date: Wed, 20 Jan 2021 02:47:44 +1100
Subject: [PATCH] Copy the Example objects (and their predicted Doc) in
 nlp.evaluate() and nlp.update() (#6765)

* Make copy of examples in nlp.update and nlp.evaluate

* Avoid circular import

* Fix evaluate
---
 spacy/language.py | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)

diff --git a/spacy/language.py b/spacy/language.py
index 3c4899fdc..d98a0e7a1 100644
--- a/spacy/language.py
+++ b/spacy/language.py
@@ -1084,6 +1084,7 @@ class Language:
         if len(examples) == 0:
             return losses
         validate_examples(examples, "Language.update")
+        examples = _copy_examples(examples)
         if sgd is None:
             if self._optimizer is None:
                 self._optimizer = self.create_optimizer()
@@ -1093,7 +1094,6 @@ class Language:
         for i, (name, proc) in enumerate(self.pipeline):
             component_cfg.setdefault(name, {})
             component_cfg[name].setdefault("drop", drop)
-            component_cfg[name].setdefault("set_annotations", False)
         for name, proc in self.pipeline:
             if name in exclude or not hasattr(proc, "update"):
                 continue
@@ -1299,6 +1299,7 @@ class Language:
         """
         examples = list(examples)
         validate_examples(examples, "Language.evaluate")
+        examples = _copy_examples(examples)
         if batch_size is None:
             batch_size = self.batch_size
         if component_cfg is None:
@@ -1311,8 +1312,6 @@ class Language:
             scorer = Scorer(**kwargs)
         # reset annotation in predicted docs and time tokenization
         start_time = timer()
-        for eg in examples:
-            eg.predicted = self.make_doc(eg.reference.text)
         # apply all pipeline components
         for name, pipe in self.pipeline:
             kwargs = component_cfg.get(name, {})
@@ -1821,6 +1820,15 @@ class DisabledPipes(list):
         self[:] = []
 
 
+def _copy_examples(examples: Iterable[Example]) -> List[Example]:
+    """Make a copy of a batch of examples, copying the predicted Doc as well.
+    This is used in contexts where we need to take ownership of the examples
+    so that they can be mutated, for instance during Language.evaluate and 
+    Language.update.
+    """
+    return [Example(eg.x.copy(), eg.y) for eg in examples]
+
+
 def _apply_pipes(
     make_doc: Callable[[str], Doc],
     pipes: Iterable[Callable[[Doc], Doc]],