Charset should only include training characters

dan/datasets/extract/__init__.py

dan/datasets/extract/exceptions.py

dan/datasets/extract/extract.py

docs/usage/datasets/extract.md

tests/data/extraction/elements/test-page_1-line_1.json

tests/data/extraction/elements/test-page_1-line_3.json

tests/data/extraction/elements/test-page_2-line_1.json

tests/data/extraction/elements/train-page_2-line_2.json

tests/data/extraction/elements/train-page_2-line_3.json

tests/data/extraction/elements/val-page_1-line_1.json

tests/test_extract.py