realtabformer.data_utils

Module Contents

Classes

TabularArtefact
ModelFileName
ModelType
ColDataType
SpecialTokens

Functions

get_uuid()
fix_multi_decimal(v)
build_vocab([df, special_tokens, add_columns])
process_numeric_data(→ Tuple[pandas.Series, Dict])
process_datetime_data(→ Tuple[pandas.Series, Dict])
process_categorical_data(→ pandas.Series)
encode_partition_numeric_col(col, tr, col_zfill)
decode_partition_numeric_col(partition_col)
tokenize_numeric_col(series[, nparts, col_zfill])
encode_column_values(series)
decode_column_values(series)
encode_processed_column(idx, dtype, col)
decode_processed_column(col)
extract_processed_column(col)
is_numeric_col(col)
is_datetime_col(col)
is_categorical_col(col)
is_numeric_datetime_col(col)
process_data(→ Tuple[pandas.DataFrame, Dict])
get_token_id(→ int)
get_input_ids(→ Dict)
make_dataset(→ datasets.Dataset)
get_relational_input_ids(→ dict)
make_relational_dataset(→ datasets.Dataset)

Attributes

TEACHER_FORCING_PRE
SPECIAL_COL_SEP
NUMERIC_NA_TOKEN
INVALID_NUMS_RE

realtabformer.data_utils.TEACHER_FORCING_PRE = '_TEACHERFORCING'

realtabformer.data_utils.SPECIAL_COL_SEP = '___'

realtabformer.data_utils.NUMERIC_NA_TOKEN = '@'

realtabformer.data_utils.INVALID_NUMS_RE = '[^\\-.0-9]'

class realtabformer.data_utils.TabularArtefact

best_disc_model: str = 'best-disc-model'

mean_best_disc_model: str = 'mean-best-disc-model'

not_best_disc_model: str = 'not-best-disc-model'

last_epoch_model: str = 'last-epoch-model'

static artefacts()

class realtabformer.data_utils.ModelFileName

rtf_config_json: str = 'rtf_config.json'

rtf_model_pt: str = 'rtf_model.pt'

static names()

class realtabformer.data_utils.ModelType

tabular: str = 'tabular'

relational: str = 'relational'

static types()

class realtabformer.data_utils.ColDataType

NUMERIC: str = 'NUMERIC'

DATETIME: str = 'DATETIME'

CATEGORICAL: str = 'CATEGORICAL'

static types()

class realtabformer.data_utils.SpecialTokens

UNK: str = '[UNK]'

SEP: str = '[SEP]'

PAD: str = '[PAD]'

CLS: str = '[CLS]'

MASK: str = '[MASK]'

BOS: str = '[BOS]'

EOS: str = '[EOS]'

BMEM: str = '[BMEM]'

EMEM: str = '[EMEM]'

RMASK: str = '[RMASK]'

SPTYPE: str = '[SPTYPE]'

static tokens()

realtabformer.data_utils.get_uuid()

realtabformer.data_utils.fix_multi_decimal(v)

realtabformer.data_utils.build_vocab(df: pandas.DataFrame = None, special_tokens=None, add_columns: bool = True)

realtabformer.data_utils.process_numeric_data(series: pandas.Series, max_len: int = 10, numeric_precision: int = 4, transform_data: Dict = None) → Tuple[pandas.Series, Dict]

realtabformer.data_utils.process_datetime_data(series, transform_data: Dict = None) → Tuple[pandas.Series, Dict]

realtabformer.data_utils.process_categorical_data(series: pandas.Series) → pandas.Series

realtabformer.data_utils.encode_partition_numeric_col(col, tr, col_zfill)

realtabformer.data_utils.decode_partition_numeric_col(partition_col)

realtabformer.data_utils.tokenize_numeric_col(series: pandas.Series, nparts=2, col_zfill=2)

realtabformer.data_utils.encode_column_values(series)

realtabformer.data_utils.decode_column_values(series)

realtabformer.data_utils.encode_processed_column(idx, dtype, col)

realtabformer.data_utils.decode_processed_column(col)

realtabformer.data_utils.extract_processed_column(col)

realtabformer.data_utils.is_numeric_col(col)

realtabformer.data_utils.is_datetime_col(col)

realtabformer.data_utils.is_categorical_col(col)

realtabformer.data_utils.is_numeric_datetime_col(col)

realtabformer.data_utils.process_data(df: pandas.DataFrame, numeric_max_len=10, numeric_precision=4, numeric_nparts=2, first_col_type=None, col_transform_data: Dict = None, target_col: str = None) → Tuple[pandas.DataFrame, Dict]

realtabformer.data_utils.get_token_id(token: str, vocab_token2id: Dict[str, int], mask_rate: float = 0) → int

realtabformer.data_utils.get_input_ids(example, vocab: Dict, columns: List, mask_rate: float = 0, return_label_ids: bool | None = True, return_token_type_ids: bool | None = False, affix_bos: bool | None = True, affix_eos: bool | None = True) → Dict

realtabformer.data_utils.make_dataset(df: pandas.DataFrame, vocab: Dict, mask_rate: float = 0, affix_eos: bool = True, return_token_type_ids: bool = False) → datasets.Dataset

realtabformer.data_utils.get_relational_input_ids(example, input_idx, vocab, columns, output_dataset, in_out_idx, output_max_length: int | None = None, return_token_type_ids: bool = False) → dict

realtabformer.data_utils.make_relational_dataset(in_df: pandas.DataFrame, out_df: pandas.DataFrame, vocab: dict, in_out_idx: dict, mask_rate=0, output_max_length: int | None = None, return_token_type_ids: bool = False) → datasets.Dataset