docs: spacy DocBin cookbook

[ignacioct]

Closes #420

Following our work in #420 and #1635, I'm creating a small example into the Cookbook. Is this enough, or do we need something closer to an actual phrase or dataset?

[frascuchon]

Hi, @ignacioct, and thanks for your work!!

Maybe it could be nice to use a dataset and show a "real" workflow. The conll2003 dataset could a candidate. You could do something like this:

import spacy
import rubrix as rb

from datasets import load_dataset

ds = load_dataset("conll2003", split="train")
rds = rb.DatasetForTokenClassification.from_datasets(ds, tags="ner_tags")
nlp = spacy.blank("en")  # A blank nlp pipeline works faster

db = rds.prepare_for_training(framework="spacy", lang=nlp)

[ignacioct]

import spacy
import rubrix as rb

from datasets import load_dataset

ds = load_dataset("conll2003", split="train")
rds = rb.DatasetForTokenClassification.from_datasets(ds, tags="ner_tags")
nlp = spacy.blank("en")  # A blank nlp pipeline works faster

db = rds.prepare_for_training(framework="spacy", lang=nlp)

Okay, using that dataset I believe there's one row unsupported, is this a problem for the example or we can go through?

2022-07-28 13:25:14.367 | WARNING  | datasets.builder:download_and_prepare:577 - Reusing dataset conll2003 (/Users/ignacio/.cache/huggingface/datasets/conll2003/conll2003/1.0.0/63f4ebd1bcb7148b1644497336fd74643d4ce70123334431a3c053b7ee4e96ee)
2022-07-28 13:25:14.372 | WARNING  | rubrix.client.datasets:_remove_unsupported_columns:252 - Following columns are not supported by the TokenClassificationRecord model and are ignored: ['pos_tags', 'chunk_tags']
2022-07-28 13:25:16.324 | WARNING  | rubrix.client.datasets:from_datasets:761 - Ignoring row with no tokens.

[frascuchon]

No, It's okay. It's just a warning. You can go ahead. Thanks

[ignacioct]

That is already implemented, so we can go forward and merge then :) @frascuchon

[frascuchon]

Great!