Utils

pipelines.enrichment.shared.utils

Description/candidate utilities (port of domain/enrichment/entity/utils.py).

extract_candidates(
    description: str,
    candidate_regexes: Iterable[re.Pattern],
) -> List[str]

Extract valid payee candidates from a description.

Parameters:

Name	Type	Description	Default
`description`	`str`	Raw description text.	required
`candidate_regexes`	`Iterable[Pattern]`	Compiled extraction regexes.	required

Returns:

Type	Description
`List[str]`	List[str]: The normalized valid candidates.

is_valid_candidate(candidate: str) -> bool

Return True when a candidate looks like a plausible payee.

Parameters:

Name	Type	Description	Default
`candidate`	`str`	Extracted candidate value.	required

Returns:

Name	Type	Description
`bool`	`bool`	True when the candidate passes the length/content heuristics.

lookup_mapping(candidate: str, mapping: dict)

Look up mapping values for a candidate token.

Parameters:

Name	Type	Description	Default
`candidate`	`str`	Candidate token.	required
`mapping`	`dict`	Mapping dict of key → MappingValue.	required

Returns:

Name	Type	Description
`list`		Matching value names, or `(None, None)` when the candidate is empty.

normalize_description(text: str) -> str

Normalize a description: separators → space, lowercase, drop OCR noise.

Parameters:

Name	Type	Description	Default
`text`	`str`	Raw description.	required

Returns:

Name	Type	Description
`str`	`str`	The normalized lowercase description.

normalize_money_str(x) -> str

Format a numeric value as a fixed two-decimal string.

Parameters:

Name	Type	Description	Default
`x`	`float`	Value coercible to float.	required

Returns:

Name	Type	Description
`str`	`str`	The value formatted to exactly two decimals.

normalize_person_handle(handle: str) -> str

Normalize a person/UPI handle into a capitalized display name.

Parameters:

Name	Type	Description	Default
`handle`	`str`	Raw handle (may carry a UPI suffix or trailing digits).	required

Returns:

Name	Type	Description
`str`	`str`	The cleaned handle, or "" when only hex characters remain.

remove_number_sequence(
    text: str, amt: str, bal: str
) -> str

Strip amount/balance digit sequences from a description.

Parameters:

Name	Type	Description	Default
`text`	`str`	Description text.	required
`amt`	`str`	Amount value.	required
`bal`	`str`	Balance value.	required

Returns:

Name	Type	Description
`str`	`str`	The cleaned, whitespace-collapsed description.