Commit Graph

2 Commits

Author SHA1 Message Date
Daniël de Kok
2468742cb8 isort all the things 2023-06-26 11:41:03 +02:00
Adriane Boyd
4a615cacd2
Consolidate and freeze symbols (#11352)
* Consolidate and freeze symbols

Instead of having symbol values defined in three potentially conflicting
places (`spacy.attrs`, `spacy.parts_of_speech`, `spacy.symbols`), define
all symbols in `spacy.symbols` and reference those values in
`spacy.attrs` and `spacy.parts_of_speech`.

Remove deprecated and placeholder symbols from `spacy.attrs.IDS`.

Make `spacy.attrs.NAMES` and `spacy.symbols.NAMES` reverse dicts rather
than lists in order to support future use of hash values in `attr_id_t`.

Minor changes:

* Use `uint64_t` for attrs in `Doc.to_array` to support future use of
hash values
* Remove unneeded attrs filter for error message in `Doc.to_array`
* Remove unused attr `SENT_END`

* Handle dynamic size of attr_id_t in Doc.to_array

* Undo added warnings

* Refactor to make Doc.to_array more similar to Doc.from_array

* Improve refactoring
2022-09-02 09:08:40 +02:00