Id Regulation Constants

Shared constants for Indonesian regulation document processing.

Centralises all regex patterns, lookup tables, and enum types used by both the IDRegulationParser and the IDRegulationChunker so neither stage needs to import from the other. IDRegulationParser annotates each parsed element with structure metadata; IDRegulationChunker groups those pre-tagged elements into per-article output chunks.

TODO: this module mixes constants with parsing helper functions (e.g. parse_indonesian_part_number, parse_roman_numeral). Split into a constants.py and a utils.py for id_regulation in a separate PR.

AmendmentGroupLevel

Bases: StrEnum

Structural level of a Bagian/BAB/Paragraf group quoted as amendment payload.

Distinguishes a quoted group's own hierarchy level (see IDRegulationParser's regulation_amendment_group_level metadata field) from the article-level AmendmentOperation -- a single INSERT item can introduce a whole new chapter/part/paragraph, and this is the vocabulary for which level it is.

AmendmentOperation

Bases: StrEnum

Amendment operation types for Indonesian regulations.

CitationType

Bases: StrEnum

Citation types for regulation cross-references.

NodeLabel

Bases: StrEnum

Neo4j node labels for the Indonesian regulation graph. See ONTOLOGY.md.

RegulationStructure

Bases: StrEnum

Structural element types within an Indonesian regulation document.

Used to tag each parsed element with its role in the document hierarchy, from front matter through chapters, parts, paragraphs, articles, body text, amendment directives, explanations, and the enactment tail.

RelationType

Bases: StrEnum

Neo4j relationship (edge) types for the Indonesian regulation graph. See ONTOLOGY.md.

VersionType

Bases: StrEnum

Article version types for Indonesian regulation graph nodes.

is_known_regulation_structure(regulation_structure)

Return whether a structure tag marks document content rather than page furniture.

IDRegulationParser keeps page boilerplate (the repeated masthead, standalone page numbers) in its output so nothing is silently dropped, but tags it UNKNOWN with an empty hierarchy. Those elements carry no legal meaning: IDRegulationChunker discards them, and structural assertions must skip them rather than read them as content.

Parameters:

Name	Type	Description	Default
regulation_structure	Any	An element's regulation_structure metadata value, as a RegulationStructure or its plain-string form.	required

Returns:

Name	Type	Description
bool	bool	True unless the tag is RegulationStructure.UNKNOWN.

is_meaningful_title(text)

Return whether a section-title string carries real content.

Headings sometimes capture punctuation- or whitespace-only noise (e.g. OCR dot-leader artifacts like ". ."), which should be discarded rather than stored as a chapter/part/paragraph title.

Parameters:

Name	Type	Description	Default
text	str	The candidate title text.	required

Returns:

Name	Type	Description
bool	bool	True if the text contains at least one alphanumeric character.

match_ordinal_header_re(pattern, group_name, text)

Match an ordinal-header regex, requiring ALL CAPS when the colon glue is unspaced.

Regex lookbehind can't check a variable-length word's casing, so this does it in Python instead of duplicating the check across DECISION_CLAUSE_HEADER_RE and INSTRUCTION_HEADER_RE.

Parameters:

Name	Type	Description	Default
pattern	Pattern[str]	DECISION_CLAUSE_HEADER_RE or INSTRUCTION_HEADER_RE.	required
group_name	str	The pattern's ordinal-word capture group name.	required
text	str	Stripped element text.	required

Returns:

Type	Description
Match[str] | None	re.Match[str] | None: The match, or None if rejected for an unspaced glue on a not-fully-uppercase word.

parse_indonesian_part_number(text)

Convert an Indonesian ordinal number-word phrase into its integer value.

Handles the irregular "Pertama", the optional ordinal prefix ke, and additive/multiplicative number words, e.g.: 1. "Pertama" -> 1 2. "Kesatu" -> 1 3. "Kesebelas" -> 11 4. "Kedua Puluh" -> 20 5. "Ketiga Puluh Lima" -> 35

Parameters:

Name	Type	Description	Default
text	str	The ordinal number-word phrase to parse, such as "Kesatu".	required

Returns:

Type	Description
int | None	int | None: The parsed integer value, or None if no number word is recognised.

parse_roman_numeral(text)

Convert a Roman numeral into its integer value.

Parameters:

Name	Type	Description	Default
text	str	The Roman numeral to parse, such as "IX". Case-insensitive.	required

Returns:

Type	Description
int | None	int | None: The parsed integer value, or None if the text is not a valid Roman numeral.

resolve_inline_title(inline)

Return the stripped title if it is meaningful, None otherwise.

Parameters:

Name	Type	Description	Default
inline	str | None	Candidate inline title captured alongside a section header.	required

Returns:

Type	Description
str | None	str | None: The stripped title when it carries real content, otherwise None.