Module timestamp 

Org-mode timestamp parsing and repeater logic.

Submodule layout:

• extract — pull timestamp / CREATED strings out of free-form text.
• parser — parse a single org-style timestamp into ParsedTimestamp.
• repeater — repeater grammar and occurrence math (+1d, ++2w, .+1m, +1wd).
• weekdays — localized weekday name normalization (RU → EN).

Structs

ParsedTimestamp

Result of parsing a single org-mode timestamp string.

Repeater

Repeater type and interval

TimestampParts

A timestamp located token by token, for a caller that rewrites its date.

Enums

DatePreference

Preference for closest date calculation

RepeaterType

Type of repeater (org-mode prefix)

RepeaterUnit

Repeater unit

Functions

add_months

Add months to a date, truncating the day to fit the destination month. Constant-time (no per-month loops), correct for negative months.

closest_date

Calculate closest occurrence date relative to current for the given repeater.

extract_created_normalized

Extract CREATED timestamp from already-weekday-normalized text. Callers in the parser pre-normalize so multiple extractors share one scan; tests pass already-English input.

extract_repeater_normalized

Extract the timestamp’s repeater as its canonical org string (++7d, .+1m, +1wd), or None when the timestamp carries no repeater. Input is expected already weekday-normalised (as assembled by extract_timestamp_normalized); a timestamp whose date does not parse yields None (an invalid timestamp is not synced downstream).

extract_timestamp_normalized

Extract non-CREATED timestamp from already-weekday-normalized text.

parse_org_timestamp

Parse a single org-mode timestamp like <2024-12-05 Thu 10:00 +1d> or <2024-12-05>--<2024-12-06>, optionally normalizing localized weekday names.

parse_repeater

Parse repeater string like +1d, ++2w, .+1m, +1wd

parse_timestamp_fields_normalized

Fast-path companion to parse_timestamp_fields for callers that have already weekday-normalised the input (e.g. parser::finalize_task, where info.timestamp was assembled from extract_timestamp_normalized’s regex captures over a normalize_weekdays output). Skipping the second normalisation removes a per-task Aho-Corasick scan on the timestamp substring.

parse_timestamp_parts

Locate the first timestamp in text, or return None when there is none.