refactor fixes (#5664)

* fixes in ud_train, UX for morphs

* update pyproject with new version of thinc

* fixes in debug_data script

* cleanup of old unused error messages

* remove obsolete TempErrors

* move error messages to errors.py

* add ENT_KB_ID to default DocBin serialization

* few fixes to simple_ner

* fix tags

bin/ud/ud_train.py

@@ -78,8 +78,7 @@ def read_data(
                 head = int(head) - 1 if head != "0" else id_
                 sent["words"].append(word)
                 sent["tags"].append(tag)
-                sent["morphology"].append(_parse_morph_string(morph))
+                sent["morphs"].append(_compile_morph_string(morph, pos))
-                sent["morphology"][-1].add("POS_%s" % pos)
                 sent["heads"].append(head)
                 sent["deps"].append("ROOT" if dep == "root" else dep)
                 sent["spaces"].append(space_after == "_")
@@ -88,12 +87,12 @@ def read_data(
             if oracle_segments:
                 docs.append(Doc(nlp.vocab, words=sent["words"], spaces=sent["spaces"]))
                 golds.append(sent)
-                assert golds[-1].morphology is not None
+                assert golds[-1]["morphs"] is not None
 
             sent_annots.append(sent)
             if raw_text and max_doc_length and len(sent_annots) >= max_doc_length:
                 doc, gold = _make_gold(nlp, None, sent_annots)
-                assert gold.morphology is not None
+                assert gold["morphs"] is not None
                 sent_annots = []
                 docs.append(doc)
                 golds.append(gold)
@@ -109,17 +108,10 @@ def read_data(
     return golds_to_gold_data(docs, golds)
 
 
-def _parse_morph_string(morph_string):
+def _compile_morph_string(morph_string, pos):
     if morph_string == '_':
-        return set()
+        return f"POS={pos}"
-    output = []
+    return morph_string + f"|POS={pos}"
-    replacements = {'1': 'one', '2': 'two', '3': 'three'}
-    for feature in morph_string.split('|'):
-        key, value = feature.split('=')
-        value = replacements.get(value, value)
-        value = value.split(',')[0]
-        output.append('%s_%s' % (key, value.lower()))
-    return set(output)
 
 
 def read_conllu(file_):
@@ -155,7 +147,7 @@ def _make_gold(nlp, text, sent_annots, drop_deps=0.0):
     sent_starts = []
     for sent in sent_annots:
         gold["heads"].extend(len(gold["words"])+head for head in sent["heads"])
-        for field in ["words", "tags", "deps", "morphology", "entities", "spaces"]:
+        for field in ["words", "tags", "deps", "morphs", "entities", "spaces"]:
             gold[field].extend(sent[field])
         sent_starts.append(True)
         sent_starts.extend([False] * (len(sent["words"]) - 1))
@@ -168,7 +160,7 @@ def _make_gold(nlp, text, sent_annots, drop_deps=0.0):
     doc = nlp.make_doc(text)
     gold.pop("spaces")
     gold["sent_starts"] = sent_starts
-    for i in range(len(gold.heads)):
+    for i in range(len(gold["heads"])):
         if random.random() < drop_deps:
             gold["heads"][i] = None
             gold["labels"][i] = None
@@ -185,7 +177,7 @@ def golds_to_gold_data(docs, golds):
     """Get out the training data format used by begin_training"""
     data = []
     for doc, gold in zip(docs, golds):
-        example = Example.from_dict(doc, gold)
+        example = Example.from_dict(doc, dict(gold))
         data.append(example)
     return data
 
@@ -354,8 +346,7 @@ def initialize_pipeline(nlp, examples, config, device):
     if config.multitask_sent:
         nlp.parser.add_multitask_objective("sent_start")
     for eg in examples:
-        gold = eg.gold
+        for tag in eg.get_aligned("TAG", as_string=True):
-        for tag in gold.tags:
             if tag is not None:
                 nlp.tagger.add_label(tag)
     if torch is not None and device != -1:
@@ -489,10 +480,6 @@ def main(
     Token.set_extension("begins_fused", default=False)
     Token.set_extension("inside_fused", default=False)
 
-    Token.set_extension("get_conllu_lines", method=get_token_conllu)
-    Token.set_extension("begins_fused", default=False)
-    Token.set_extension("inside_fused", default=False)
-
     spacy.util.fix_random_seed()
     lang.zh.Chinese.Defaults.use_jieba = False
     lang.ja.Japanese.Defaults.use_janome = False
@@ -535,10 +522,10 @@ def main(
         else:
             batches = minibatch(examples, size=batch_sizes)
         losses = {}
-        n_train_words = sum(len(eg.doc) for eg in examples)
+        n_train_words = sum(len(eg.predicted) for eg in examples)
         with tqdm.tqdm(total=n_train_words, leave=False) as pbar:
             for batch in batches:
-                pbar.update(sum(len(ex.doc) for ex in batch))
+                pbar.update(sum(len(ex.predicted) for ex in batch))
                 nlp.parser.cfg["beam_update_prob"] = next(beam_prob)
                 nlp.update(
                     batch,

examples/training/conllu.py

@@ -283,7 +283,7 @@ def initialize_pipeline(nlp, examples, config):
     nlp.parser.moves.add_action(2, "subtok")
     nlp.add_pipe(nlp.create_pipe("tagger"))
     for eg in examples:
-        for tag in eg.gold.tags:
+        for tag in eg.get_aligned("TAG", as_string=True):
             if tag is not None:
                 nlp.tagger.add_label(tag)
     # Replace labels that didn't make the frequency cutoff

examples/training/train_ner.py

@@ -56,7 +56,7 @@ def main(model=None, output_dir=None, n_iter=100):
             print("Add label", ent[2])
             ner.add_label(ent[2])
 
-    with nlp.select_pipes(enable="ner") and warnings.catch_warnings():
+    with nlp.select_pipes(enable="simple_ner") and warnings.catch_warnings():
         # show warnings for misaligned entity spans once
         warnings.filterwarnings("once", category=UserWarning, module="spacy")
 

spacy/cli/debug_data.py

@@ -102,9 +102,6 @@ def debug_data(
         corpus = Corpus(train_path, dev_path)
         try:
             train_dataset = list(corpus.train_dataset(nlp))
-            train_dataset_unpreprocessed = list(
-                corpus.train_dataset_without_preprocessing(nlp)
-            )
         except ValueError as e:
             loading_train_error_message = f"Training data cannot be loaded: {e}"
         try:
@@ -120,11 +117,9 @@ def debug_data(
     msg.good("Corpus is loadable")
 
     # Create all gold data here to avoid iterating over the train_dataset constantly
-    gold_train_data = _compile_gold(train_dataset, pipeline, nlp)
+    gold_train_data = _compile_gold(train_dataset, pipeline, nlp, make_proj=True)
-    gold_train_unpreprocessed_data = _compile_gold(
+    gold_train_unpreprocessed_data = _compile_gold(train_dataset, pipeline, nlp, make_proj=False)
-        train_dataset_unpreprocessed, pipeline
+    gold_dev_data = _compile_gold(dev_dataset, pipeline, nlp, make_proj=True)
-    )
-    gold_dev_data = _compile_gold(dev_dataset, pipeline, nlp)
 
     train_texts = gold_train_data["texts"]
     dev_texts = gold_dev_data["texts"]
@@ -497,7 +492,7 @@ def _load_file(file_path: Path, msg: Printer) -> None:
 
 
 def _compile_gold(
-    examples: Sequence[Example], pipeline: List[str], nlp: Language
+    examples: Sequence[Example], pipeline: List[str], nlp: Language, make_proj: bool
 ) -> Dict[str, Any]:
     data = {
         "ner": Counter(),
@@ -517,9 +512,9 @@ def _compile_gold(
         "n_cats_multilabel": 0,
         "texts": set(),
     }
-    for example in examples:
+    for eg in examples:
-        gold = example.reference
+        gold = eg.reference
-        doc = example.predicted
+        doc = eg.predicted
         valid_words = [x for x in gold if x is not None]
         data["words"].update(valid_words)
         data["n_words"] += len(valid_words)
@@ -530,7 +525,7 @@ def _compile_gold(
                 if nlp.vocab.strings[word] not in nlp.vocab.vectors:
                     data["words_missing_vectors"].update([word])
         if "ner" in pipeline:
-            for i, label in enumerate(gold.ner):
+            for i, label in enumerate(eg.get_aligned_ner()):
                 if label is None:
                     continue
                 if label.startswith(("B-", "U-", "L-")) and doc[i].is_space:
@@ -556,16 +551,18 @@ def _compile_gold(
             if list(gold.cats.values()).count(1.0) != 1:
                 data["n_cats_multilabel"] += 1
         if "tagger" in pipeline:
-            data["tags"].update([x for x in gold.tags if x is not None])
+            tags = eg.get_aligned("TAG", as_string=True)
+            data["tags"].update([x for x in tags if x is not None])
         if "parser" in pipeline:
-            data["deps"].update([x for x in gold.labels if x is not None])
+            aligned_heads, aligned_deps = eg.get_aligned_parse(projectivize=make_proj)
-            for i, (dep, head) in enumerate(zip(gold.labels, gold.heads)):
+            data["deps"].update([x for x in aligned_deps if x is not None])
+            for i, (dep, head) in enumerate(zip(aligned_deps, aligned_heads)):
                 if head == i:
                     data["roots"].update([dep])
                     data["n_sents"] += 1
-            if nonproj.is_nonproj_tree(gold.heads):
+            if nonproj.is_nonproj_tree(aligned_heads):
                 data["n_nonproj"] += 1
-            if nonproj.contains_cycle(gold.heads):
+            if nonproj.contains_cycle(aligned_heads):
                 data["n_cycles"] += 1
     return data
 
@@ -581,7 +578,7 @@ def _get_examples_without_label(data: Sequence[Example], label: str) -> int:
     for eg in data:
         labels = [
             label.split("-")[1]
-            for label in eg.gold.ner
+            for label in eg.get_aligned_ner()
             if label not in ("O", "-", None)
         ]
         if label not in labels:

spacy/errors.py

@@ -132,6 +132,7 @@ class Warnings(object):
             "are currently: da, de, el, en, id, lb, pt, ru, sr, ta, th.")
 
     # TODO: fix numbering after merging develop into master
+    W092 = ("Ignoring annotations for sentence starts, as dependency heads are set.")
     W093 = ("Could not find any data to train the {name} on. Is your "
             "input data correctly formatted ?")
     W094 = ("Model '{model}' ({model_version}) specifies an under-constrained "
@@ -154,7 +155,7 @@ class Warnings(object):
             "so a default configuration was used.")
     W099 = ("Expected 'dict' type for the 'model' argument of pipe '{pipe}', "
             "but got '{type}' instead, so ignoring it.")
-    W100 = ("Skipping unsupported morphological feature(s): {feature}. "
+    W100 = ("Skipping unsupported morphological feature(s): '{feature}'. "
             "Provide features as a dict {{\"Field1\": \"Value1,Value2\"}} or "
             "string \"Field1=Value1,Value2|Field2=Value3\".")
 
@@ -182,18 +183,13 @@ class Errors(object):
             "`nlp.select_pipes()`, you should remove them explicitly with "
             "`nlp.remove_pipe()` before the pipeline is restored. Names of "
             "the new components: {names}")
-    E009 = ("The `update` method expects same number of docs and golds, but "
-            "got: {n_docs} docs, {n_golds} golds.")
     E010 = ("Word vectors set to length 0. This may be because you don't have "
             "a model installed or loaded, or because your model doesn't "
             "include word vectors. For more info, see the docs:\n"
             "https://spacy.io/usage/models")
     E011 = ("Unknown operator: '{op}'. Options: {opts}")
     E012 = ("Cannot add pattern for zero tokens to matcher.\nKey: {key}")
-    E013 = ("Error selecting action in matcher")
     E014 = ("Unknown tag ID: {tag}")
-    E015 = ("Conflicting morphology exception for ({tag}, {orth}). Use "
-            "`force=True` to overwrite.")
     E016 = ("MultitaskObjective target should be function or one of: dep, "
             "tag, ent, dep_tag_offset, ent_tag.")
     E017 = ("Can only add unicode or bytes. Got type: {value_type}")
@@ -201,21 +197,8 @@ class Errors(object):
             "refers to an issue with the `Vocab` or `StringStore`.")
     E019 = ("Can't create transition with unknown action ID: {action}. Action "
             "IDs are enumerated in spacy/syntax/{src}.pyx.")
-    E020 = ("Could not find a gold-standard action to supervise the "
-            "dependency parser. The tree is non-projective (i.e. it has "
-            "crossing arcs - see spacy/syntax/nonproj.pyx for definitions). "
-            "The ArcEager transition system only supports projective trees. "
-            "To learn non-projective representations, transform the data "
-            "before training and after parsing. Either pass "
-            "`make_projective=True` to the GoldParse class, or use "
-            "spacy.syntax.nonproj.preprocess_training_data.")
-    E021 = ("Could not find a gold-standard action to supervise the "
-            "dependency parser. The GoldParse was projective. The transition "
-            "system has {n_actions} actions. State at failure: {state}")
     E022 = ("Could not find a transition with the name '{name}' in the NER "
             "model.")
-    E023 = ("Error cleaning up beam: The same state occurred twice at "
-            "memory address {addr} and position {i}.")
     E024 = ("Could not find an optimal move to supervise the parser. Usually, "
             "this means that the model can't be updated in a way that's valid "
             "and satisfies the correct annotations specified in the GoldParse. "
@@ -259,7 +242,6 @@ class Errors(object):
             "offset {start}.")
     E037 = ("Error calculating span: Can't find a token ending at character "
             "offset {end}.")
-    E038 = ("Error finding sentence for span. Infinite loop detected.")
     E039 = ("Array bounds exceeded while searching for root word. This likely "
             "means the parse tree is in an invalid state. Please report this "
             "issue here: http://github.com/explosion/spaCy/issues")
@@ -290,8 +272,6 @@ class Errors(object):
     E059 = ("One (and only one) keyword arg must be set. Got: {kwargs}")
     E060 = ("Cannot add new key to vectors: the table is full. Current shape: "
             "({rows}, {cols}).")
-    E061 = ("Bad file name: {filename}. Example of a valid file name: "
-            "'vectors.128.f.bin'")
     E062 = ("Cannot find empty bit for new lexical flag. All bits between 0 "
             "and 63 are occupied. You can replace one by specifying the "
             "`flag_id` explicitly, e.g. "
@@ -305,39 +285,17 @@ class Errors(object):
             "Query string: {string}\nOrth cached: {orth}\nOrth ID: {orth_id}")
     E065 = ("Only one of the vector table's width and shape can be specified. "
             "Got width {width} and shape {shape}.")
-    E066 = ("Error creating model helper for extracting columns. Can only "
-            "extract columns by positive integer. Got: {value}.")
     E067 = ("Invalid BILUO tag sequence: Got a tag starting with 'I' (inside "
             "an entity) without a preceding 'B' (beginning of an entity). "
             "Tag sequence:\n{tags}")
     E068 = ("Invalid BILUO tag: '{tag}'.")
-    E069 = ("Invalid gold-standard parse tree. Found cycle between word "
-            "IDs: {cycle} (tokens: {cycle_tokens}) in the document starting "
-            "with tokens: {doc_tokens}.")
-    E070 = ("Invalid gold-standard data. Number of documents ({n_docs}) "
-            "does not align with number of annotations ({n_annots}).")
     E071 = ("Error creating lexeme: specified orth ID ({orth}) does not "
             "match the one in the vocab ({vocab_orth}).")
-    E072 = ("Error serializing lexeme: expected data length {length}, "
-            "got {bad_length}.")
     E073 = ("Cannot assign vector of length {new_length}. Existing vectors "
             "are of length {length}. You can use `vocab.reset_vectors` to "
             "clear the existing vectors and resize the table.")
     E074 = ("Error interpreting compiled match pattern: patterns are expected "
             "to end with the attribute {attr}. Got: {bad_attr}.")
-    E075 = ("Error accepting match: length ({length}) > maximum length "
-            "({max_len}).")
-    E076 = ("Error setting tensor on Doc: tensor has {rows} rows, while Doc "
-            "has {words} words.")
-    E077 = ("Error computing {value}: number of Docs ({n_docs}) does not "
-            "equal number of GoldParse objects ({n_golds}) in batch.")
-    E078 = ("Error computing score: number of words in Doc ({words_doc}) does "
-            "not equal number of words in GoldParse ({words_gold}).")
-    E079 = ("Error computing states in beam: number of predicted beams "
-            "({pbeams}) does not equal number of gold beams ({gbeams}).")
-    E080 = ("Duplicate state found in beam: {key}.")
-    E081 = ("Error getting gradient in beam: number of histories ({n_hist}) "
-            "does not equal number of losses ({losses}).")
     E082 = ("Error deprojectivizing parse: number of heads ({n_heads}), "
             "projective heads ({n_proj_heads}) and labels ({n_labels}) do not "
             "match.")
@@ -345,8 +303,6 @@ class Errors(object):
             "`getter` (plus optional `setter`) is allowed. Got: {nr_defined}")
     E084 = ("Error assigning label ID {label} to span: not in StringStore.")
     E085 = ("Can't create lexeme for string '{string}'.")
-    E086 = ("Error deserializing lexeme '{string}': orth ID {orth_id} does "
-            "not match hash {hash_id} in StringStore.")
     E087 = ("Unknown displaCy style: {style}.")
     E088 = ("Text of length {length} exceeds maximum of {max_length}. The "
             "v2.x parser and NER models require roughly 1GB of temporary "
@@ -388,7 +344,6 @@ class Errors(object):
     E103 = ("Trying to set conflicting doc.ents: '{span1}' and '{span2}'. A "
             "token can only be part of one entity, so make sure the entities "
             "you're setting don't overlap.")
-    E104 = ("Can't find JSON schema for '{name}'.")
     E105 = ("The Doc.print_tree() method is now deprecated. Please use "
             "Doc.to_json() instead or write your own function.")
     E106 = ("Can't find doc._.{attr} attribute specified in the underscore "
@@ -411,8 +366,6 @@ class Errors(object):
             "practically no advantage over pickling the parent Doc directly. "
             "So instead of pickling the span, pickle the Doc it belongs to or "
             "use Span.as_doc to convert the span to a standalone Doc object.")
-    E113 = ("The newly split token can only have one root (head = 0).")
-    E114 = ("The newly split token needs to have a root (head = 0).")
     E115 = ("All subtokens must have associated heads.")
     E116 = ("Cannot currently add labels to pretrained text classifier. Add "
             "labels before training begins. This functionality was available "
@@ -435,12 +388,9 @@ class Errors(object):
             "equal to span length ({span_len}).")
     E122 = ("Cannot find token to be split. Did it get merged?")
     E123 = ("Cannot find head of token to be split. Did it get merged?")
-    E124 = ("Cannot read from file: {path}. Supported formats: {formats}")
     E125 = ("Unexpected value: {value}")
     E126 = ("Unexpected matcher predicate: '{bad}'. Expected one of: {good}. "
             "This is likely a bug in spaCy, so feel free to open an issue.")
-    E127 = ("Cannot create phrase pattern representation for length 0. This "
-            "is likely a bug in spaCy.")
     E128 = ("Unsupported serialization argument: '{arg}'. The use of keyword "
             "arguments to exclude fields from being serialized or deserialized "
             "is now deprecated. Please use the `exclude` argument instead. "
@@ -482,8 +432,6 @@ class Errors(object):
             "provided {found}.")
     E143 = ("Labels for component '{name}' not initialized. Did you forget to "
             "call add_label()?")
-    E144 = ("Could not find parameter `{param}` when building the entity "
-            "linker model.")
     E145 = ("Error reading `{param}` from input file.")
     E146 = ("Could not access `{path}`.")
     E147 = ("Unexpected error in the {method} functionality of the "
@@ -495,8 +443,6 @@ class Errors(object):
             "the component matches the model being loaded.")
     E150 = ("The language of the `nlp` object and the `vocab` should be the "
             "same, but found '{nlp}' and '{vocab}' respectively.")
-    E151 = ("Trying to call nlp.update without required annotation types. "
-            "Expected top-level keys: {exp}. Got: {unexp}.")
     E152 = ("The attribute {attr} is not supported for token patterns. "
             "Please use the option validate=True with Matcher, PhraseMatcher, "
             "or EntityRuler for more details.")
@@ -533,11 +479,6 @@ class Errors(object):
             "that case.")
     E166 = ("Can only merge DocBins with the same pre-defined attributes.\n"
             "Current DocBin: {current}\nOther DocBin: {other}")
-    E167 = ("Unknown morphological feature: '{feat}' ({feat_id}). This can "
-            "happen if the tagger was trained with a different set of "
-            "morphological features. If you're using a pretrained model, make "
-            "sure that your models are up to date:\npython -m spacy validate")
-    E168 = ("Unknown field: {field}")
     E169 = ("Can't find module: {module}")
     E170 = ("Cannot apply transition {name}: invalid for the current state.")
     E171 = ("Matcher.add received invalid on_match callback argument: expected "
@@ -548,8 +489,6 @@ class Errors(object):
     E173 = ("As of v2.2, the Lemmatizer is initialized with an instance of "
             "Lookups containing the lemmatization tables. See the docs for "
             "details: https://spacy.io/api/lemmatizer#init")
-    E174 = ("Architecture '{name}' not found in registry. Available "
-            "names: {names}")
     E175 = ("Can't remove rule for unknown match pattern ID: {key}")
     E176 = ("Alias '{alias}' is not defined in the Knowledge Base.")
     E177 = ("Ill-formed IOB input detected: {tag}")
@@ -597,10 +536,19 @@ class Errors(object):
     E198 = ("Unable to return {n} most similar vectors for the current vectors "
             "table, which contains {n_rows} vectors.")
     E199 = ("Unable to merge 0-length span at doc[{start}:{end}].")
-    E200 = ("Specifying a base model with a pretrained component '{component}' "
-            "can not be combined with adding a pretrained Tok2Vec layer.")
 
     # TODO: fix numbering after merging develop into master
+    E971 = ("Found incompatible lengths in Doc.from_array: {array_length} for the "
+            "array and {doc_length} for the Doc itself.")
+    E972 = ("Example.__init__ got None for '{arg}'. Requires Doc.")
+    E973 = ("Unexpected type for NER data")
+    E974 = ("Unknown {obj} attribute: {key}")
+    E975 = ("The method Example.from_dict expects a Doc as first argument, "
+            "but got {type}")
+    E976 = ("The method Example.from_dict expects a dict as second argument, "
+            "but received None.")
+    E977 = ("Can not compare a MorphAnalysis with a string object. "
+            "This is likely a bug in spaCy, so feel free to open an issue.")
     E978 = ("The {method} method of component {name} takes a list of Example objects, "
             "but found {types} instead.")
     E979 = ("Cannot convert {type} to an Example object.")
@@ -648,13 +596,8 @@ class Errors(object):
 @add_codes
 class TempErrors(object):
     T003 = ("Resizing pretrained Tagger models is not currently supported.")
-    T004 = ("Currently parser depth is hard-coded to 1. Received: {value}.")
     T007 = ("Can't yet set {attr} from Span. Vote for this feature on the "
             "issue tracker: http://github.com/explosion/spaCy/issues")
-    T008 = ("Bad configuration of Tagger. This is probably a bug within "
-            "spaCy. We changed the name of an internal attribute for loading "
-            "pretrained vectors, and the class has been passed the old name "
-            "(pretrained_dims) but not the new name (pretrained_vectors).")
 
 
 # fmt: on

spacy/gold/corpus.py

@@ -45,7 +45,7 @@ class Corpus:
 
     def make_examples(self, nlp, reference_docs, max_length=0):
         for reference in reference_docs:
-            if max_length >= 1 and len(reference) >= max_length:
+            if len(reference) >= max_length >= 1:
                 if reference.is_sentenced:
                     for ref_sent in reference.sents:
                         yield Example(

spacy/gold/example.pyx

@@ -2,7 +2,6 @@ import warnings
 
 import numpy
 
-from ..tokens import Token
 from ..tokens.doc cimport Doc
 from ..tokens.span cimport Span
 from ..tokens.span import Span
@@ -11,9 +10,8 @@ from .align cimport Alignment
 from .iob_utils import biluo_to_iob, biluo_tags_from_offsets, biluo_tags_from_doc
 from .iob_utils import spans_from_biluo_tags
 from .align import Alignment
-from ..errors import Errors, AlignmentError
+from ..errors import Errors, Warnings
 from ..syntax import nonproj
-from ..util import get_words_and_spaces
 
 
 cpdef Doc annotations2doc(vocab, tok_annot, doc_annot):
@@ -32,11 +30,10 @@ cpdef Doc annotations2doc(vocab, tok_annot, doc_annot):
 cdef class Example:
     def __init__(self, Doc predicted, Doc reference, *, Alignment alignment=None):
         """ Doc can either be text, or an actual Doc """
-        msg = "Example.__init__ got None for '{arg}'. Requires Doc."
         if predicted is None:
-            raise TypeError(msg.format(arg="predicted"))
+            raise TypeError(Errors.E972.format(arg="predicted"))
         if reference is None:
-            raise TypeError(msg.format(arg="reference"))
+            raise TypeError(Errors.E972.format(arg="reference"))
         self.x = predicted
         self.y = reference
         self._alignment = alignment
@@ -64,9 +61,9 @@ cdef class Example:
     @classmethod
     def from_dict(cls, Doc predicted, dict example_dict):
         if example_dict is None:
-            raise ValueError("Example.from_dict expected dict, received None")
+            raise ValueError(Errors.E976)
         if not isinstance(predicted, Doc):
-            raise TypeError(f"Argument 1 should be Doc. Got {type(predicted)}")
+            raise TypeError(Errors.E975.format(type=type(predicted)))
         example_dict = _fix_legacy_dict_data(example_dict)
         tok_dict, doc_dict = _parse_example_dict_data(example_dict)
         if "ORTH" not in tok_dict:
@@ -118,6 +115,7 @@ cdef class Example:
         aligned_deps = [None] * self.x.length
         heads = [token.head.i for token in self.y]
         deps = [token.dep_ for token in self.y]
+        if projectivize:
             heads, deps = nonproj.projectivize(heads, deps)
         for cand_i in range(self.x.length):
             gold_i = cand_to_gold[cand_i]
@@ -245,11 +243,11 @@ def _annot2array(vocab, tok_annot, doc_annot):
             elif key == "cats":
                 pass
             else:
-                raise ValueError(f"Unknown doc attribute: {key}")
+                raise ValueError(Errors.E974.format(obj="doc", key=key))
 
     for key, value in tok_annot.items():
         if key not in IDS:
-            raise ValueError(f"Unknown token attribute: {key}")
+            raise ValueError(Errors.E974.format(obj="token", key=key))
         elif key in ["ORTH", "SPACY"]:
             pass
         elif key == "HEAD":
@@ -289,7 +287,7 @@ def _add_entities_to_doc(doc, ner_data):
         doc.ents = ner_data
         doc.ents = [span for span in ner_data if span.label_]
     else:
-        raise ValueError("Unexpected type for NER data")
+        raise ValueError(Errors.E973)
 
 
 def _parse_example_dict_data(example_dict):
@@ -341,7 +339,7 @@ def _fix_legacy_dict_data(example_dict):
     if "HEAD" in token_dict and "SENT_START" in token_dict:
         # If heads are set, we don't also redundantly specify SENT_START.
         token_dict.pop("SENT_START")
-        warnings.warn("Ignoring annotations for sentence starts, as dependency heads are set")
+        warnings.warn(Warnings.W092)
     return {
         "token_annotation": token_dict,
         "doc_annotation": doc_dict

spacy/gold/gold_io.pyx

@@ -145,7 +145,7 @@ def json_to_annotations(doc):
         example["doc_annotation"] = dict(
             cats=cats,
             entities=ner_tags,
-            links=paragraph.get("links", [])   # TODO: fix/test
+            links=paragraph.get("links", [])
         )
         yield example
 

spacy/morphology.pyx

@@ -107,9 +107,9 @@ cdef class Morphology:
         Returns the hash of the new analysis.
         """
         cdef MorphAnalysisC* tag_ptr
+        if isinstance(features, str):
             if features == self.EMPTY_MORPH:
                 features = ""
-        if isinstance(features, str):
             tag_ptr = <MorphAnalysisC*>self.tags.get(<hash_t>self.strings[features])
             if tag_ptr != NULL:
                 return tag_ptr.key

spacy/pipeline/simple_ner.py

@@ -70,7 +70,7 @@ class SimpleNER(Pipe):
     def update(self, examples, set_annotations=False, drop=0.0, sgd=None, losses=None):
         if not any(_has_ner(eg) for eg in examples):
             return 0
-        docs = [eg.doc for eg in examples]
+        docs = [eg.predicted for eg in examples]
         set_dropout_rate(self.model, drop)
         scores, bp_scores = self.model.begin_update(docs)
         loss, d_scores = self.get_loss(examples, scores)
@@ -89,7 +89,8 @@ class SimpleNER(Pipe):
         d_scores = []
         truths = []
         for eg in examples:
-            gold_tags = [(tag if tag != "-" else None) for tag in eg.gold.ner]
+            tags = eg.get_aligned("TAG", as_string=True)
+            gold_tags = [(tag if tag != "-" else None) for tag in tags]
             if not self.is_biluo:
                 gold_tags = biluo_to_iob(gold_tags)
             truths.append(gold_tags)
@@ -128,8 +129,8 @@ class SimpleNER(Pipe):
         pass
 
 
-def _has_ner(eg):
+def _has_ner(example):
-    for ner_tag in eg.gold.ner:
+    for ner_tag in example.get_aligned_ner():
         if ner_tag != "-" and ner_tag is not None:
             return True
     else:

spacy/tokens/_serialize.py

@@ -9,7 +9,7 @@ from ..attrs import SPACY, ORTH, intify_attr
 from ..errors import Errors
 
 
-ALL_ATTRS = ("ORTH", "TAG", "HEAD", "DEP", "ENT_IOB", "ENT_TYPE", "LEMMA", "MORPH")
+ALL_ATTRS = ("ORTH", "TAG", "HEAD", "DEP", "ENT_IOB", "ENT_TYPE", "ENT_KB_ID", "LEMMA", "MORPH")
 
 
 class DocBin(object):

spacy/tokens/doc.pyx

@@ -816,7 +816,7 @@ cdef class Doc:
         cdef TokenC* tokens = self.c
         cdef int length = len(array)
         if length != len(self):
-            raise ValueError("Cannot set array values longer than the document.")
+            raise ValueError(Errors.E971.format(array_length=length, doc_length=len(self)))
 
         # Get set up for fast loading
         cdef Pool mem = Pool()

spacy/tokens/morphanalysis.pyx

@@ -1,6 +1,7 @@
 from libc.string cimport memset
 cimport numpy as np
 
+from ..errors import Errors
 from ..vocab cimport Vocab
 from ..typedefs cimport hash_t, attr_t
 from ..morphology cimport list_features, check_feature, get_by_field
@@ -49,6 +50,8 @@ cdef class MorphAnalysis:
         return self.key
 
     def __eq__(self, other):
+        if isinstance(other, str):
+            raise ValueError(Errors.E977)
         return self.key == other.key
 
     def __ne__(self, other):