how to explain name and proc

[Zanejins]

 nlp = spacy.load(trained_pipeline)

    doc_bin = DocBin(store_user_data=True).from_disk(test_data)
    docs = doc_bin.get_docs(nlp.vocab)
    examples = []
    for gold in docs:
        pred = Doc(
            nlp.vocab,
            words=[t.text for t in gold],
            spaces=[t.whitespace_ for t in gold],
        )
        pred.ents = gold.ents
        for name, proc in nlp.pipeline:
            pred = proc(pred)
        examples.append(Example(pred, gold))

How to explain and understand the meaning of name and proc? Thank you

[kadarakos]

Hey Zanejins,

If you load a pre-trained model:

import spacy
nlp = spacy.load("en_core_web_sm")

then you can iterate through the individual pipeline components. For example to print out the names of each pipe you can do

for name, proc in nlp.pipeline:
    print(name)

which gives the output:

tok2vec
tagger
parser
attribute_ruler
lemmatizer
ner

If you print the "proc" that stands for "processor" you get:

for name, proc in nlp.pipeline:
    print(proc)

<spacy.pipeline.tok2vec.Tok2Vec object at 0x7f4f3e303c20>
<spacy.pipeline.tagger.Tagger object at 0x7f4f3e303a60>
<spacy.pipeline.dep_parser.DependencyParser object at 0x7f4f3e11f0d0>
<spacy.pipeline.attributeruler.AttributeRuler object at 0x7f4f3e089be0>
<spacy.lang.en.lemmatizer.EnglishLemmatizer object at 0x7f4f3e084960>
<spacy.pipeline.ner.EntityRecognizer object at 0x7f4f3e18aed0>

The code that you've posted goes through all the "gold" documents in a DocBin, creating a copy of each Doc without the annotation and runs them through each pipeline component in order and saves the resulting Doc as the predicted document to be compared against the gold.