FastEHR.dataloader.tokenizers_local.base¶
Classes¶
Base class for custom tokenizers |
Module Contents¶
- class FastEHR.dataloader.tokenizers_local.base.TokenizerBase¶
Base class for custom tokenizers
- property vocab_size¶
- property fit_description¶
- static event_frequency(meta_information, include_measurements=True, include_diagnoses=True) polars.DataFrame¶
Get polars dataframe with three columns: event, count and relative frequencies
Returns ┌──────────────────────────┬─────────┬───────────┐ │ EVENT ┆ COUNT ┆ FREQUENCY │ │ — ┆ — ┆ — │ │ str ┆ u32 ┆ f64 │ ╞══════════════════════════╪═════════╪═══════════╡ │ <event name 1> ┆ n1 ┆ p1 │ │ <event name 2> ┆ n2 ┆ p2 │ │ … ┆ … ┆ … │ └──────────────────────────┴─────────┴───────────┘
- abstract fit(event_counts: polars.DataFrame, **kwargs)¶
- encode(sequence: list[str])¶
Take a <> of strings, output a list of integers
- decode(sequence: list[str])¶