Update all builtin modules to new datatypes

[markgw]

The new datatypes system is now implemented and incorporated into the master branch. All existing core modules need to be updated to use the new datatypes.

Those that haven't been updated import from old_datatypes to avoid import errors, but won't work if you try using them. They are also marked with a special flag AWAITING_UPDATE = True in their __init__.py, causing them to be marked in the documentation with !! and a note explaining that they're not updated.

They also all have TODOs, causing them to show up in the currently very long todo list.

The following modules have yet to be updated at the time of writing:

• [ ] candc
• [ ] corenlp
• [x] corpora.vocab_counter
• [ ] embeddings.dependencies
• [x] embeddings.store_tsv
• [x] embeddings.store_word2vec
• [x] embeddings.word2vec
• [ ] features.term_feature_compiler
• [ ] features.term_feature_matrix_builder
• [ ] features.vocab_builder
• [ ] features.vocab_mapper
• [x] malt.conll_parser_input
• [x] malt.parse
• [x] nltk.nist_tokenize
• [ ] opennlp.coreference
• [ ] opennlp.coreference_pipeline
• [ ] opennlp.ner
• [x] opennlp.parse
• [x] opennlp.pos
• [ ] r.script
• [ ] regex.annotated_text - drop??
• [ ] sklearn.matrix_factorization
• [x] text.char_tokenize
• [ ] text.text_normalize
• [x] text.untokenize - to be dropped, no longer necessary
• [x] utility.alias
• [x] utility.collect_files
• [ ] utility.copy_file
• [ ] visualization.bar_chart
• [x] visualization.embeddings_plot

[markgw]

Modules not yet updated have now been moved to pimlico.old_datatypes.modules, so they don't show up at all in the documentation.

Lots of modules have now been updated, so this is no longer standing in the way of v1.0 release. Ideally some more of these will be done by then.

Modules still to be updated at some point are listed in the docs at plans/index.rst.