todo.parse-repeat

def parse_repeat(text: str, due: Optional[str], today: str) -> Optional[Recurrence]

from fune.todo.parse_repeat import parse_repeat  # todo.parse-repeat@^1

import re
from dataclasses import replace
from typing import Optional

from .dates_add_days import add_days  ← from dates.add-days ^1.0.0 · built alongside by fune
from .dates_day_of_week import day_of_week  ← from dates.day-of-week ^1.0.0 · built alongside by fune
from .todo_item import Recurrence  ← from todo.item ^1.0.0 · built alongside by fune
from .todo_parse_date_phrase import parse_date_phrase  ← from todo.parse-date-phrase ^1.0.0 · built alongside by fune
from .todo_parse_quick_add import parse_quick_add  ← from todo.parse-quick-add ^1.0.0 · built alongside by fune

_SPACE = re.compile(r"[ \t\n\r\x0b\x0c]+")


def _lower_ascii(text: str) -> str:
    return "".join(c.lower() if "A" <= c <= "Z" else c for c in text)


def parse_repeat(text: str, due: Optional[str], today: str) -> Optional[Recurrence]:
    """A Repeat box: an "every ..." phrase as todo.parse-quick-add reads it, then optionally a start date."""
    # add_days(x, 0) is the date check: it raises dates.add-days's message.
    add_days(today, 0)
    if due is not None:
        add_days(due, 0)
    words = [w for w in _SPACE.split(text) if w != ""]
    if len(words) < 2 or _lower_ascii(words[0]) != "every":
        return None

    # The repeat is 2 words ("every week") or 3 ("every 2 weeks"); no 2-word
    # form starts a 3-word one, so the first that parses as a bare repeat wins.
    repeat: Optional[Recurrence] = None
    length = 0
    for n in (2, 3):
        if n > len(words):
            break
        draft = parse_quick_add(" ".join(words[:n]), today)
        if draft.recurrence is not None and draft.title == "":
            repeat = draft.recurrence
            length = n
            break
    if repeat is None:
        return None

    start = due
    if len(words) > length:
        start = parse_date_phrase(" ".join(words[length:]), today)
        if start is None:
            return None
    if start is None:
        return repeat

    if repeat.frequency == "weekdays":
        day = day_of_week(start)
        if day >= 6:
            start = add_days(start, 8 - day)
    elif repeat.frequency == "weekly" and length == 2 and parse_date_phrase(words[1], today) is not None:
        # "every monday": quick-add anchored it on the next Monday, so that
        # anchor's weekday is the one the series keeps.
        start = add_days(start, (day_of_week(repeat.anchor) - day_of_week(start) + 7) % 7)
    return replace(repeat, anchor=start)

Install

fune build

With that line in your source, in a Python project (language python in fune.project), fune build resolves it and its 5 dependencies, pins them in fune.lock, downloads only the Python package of each, and builds the code above into your project’s .fune/build, one readable file per capability with a header linking back here. Or pin a range in fune.project and build in one step:

fune add todo.parse-repeat

The manifest, vectors and README with only the Python implementation. Install it without the registry with fune add ./todo.parse-repeat-1.0.0-python.fune, or fetch it from a terminal with fune pull todo.parse-repeat@1.0.0:python.

The whole function, every language, is one file too: todo.parse-repeat-1.0.0.fune, 16,279 bytes, sha256 efdd5ec36a56dc62104de85dc0453a6b348ac1d515dc1f5a5e1f6b7a9c048043. It installs into a project of any language.

# fune: before todo.parse-repeat

# fune: after todo.parse-repeat

# fune: replace dates.add-days in todo.parse-repeat
# fune: replace dates.day-of-week in todo.parse-repeat
# fune: replace todo.item in todo.parse-repeat
# fune: replace todo.parse-date-phrase in todo.parse-repeat
# fune: replace todo.parse-quick-add in todo.parse-repeat

# fune: step todo.parse-repeat after <n|label>

More from the author

## What it reads

The whole box must be one `every ...` phrase, then optionally a start date:

- **The repeat** is todo.parse-quick-add's grammar, exactly (the capability calls it rather than keeping a copy): `every day`, `every weekday`, `every week`, `every monday` (any weekday name), `every month`, `every year`, `every other week`, `every 3 days`, `every ten years` (1 to 999). Case-insensitive, ASCII whitespace between words. - **The start date**, if any, is whatever todo.parse-date-phrase reads as a whole: `tomorrow`, `on fri`, `next week`, `2026-10-15`. - **Anything else is null**: words left over (`every week please`), a date first (`tomorrow every week`), `weekly` on its own (a title word to quick-add), an empty box, `every 1000 days`. A form shows its own "try every week" message for null.

## Where the series starts

The anchor is the series' first date, and so the todo's due date from then on (todo.item refuses a due date before the anchor).

1. The date typed in the box, if there is one. 2. Otherwise `due`, the todo's due date. 3. Otherwise quick-add's choice: today; for `every weekday` today if it is Monday to Friday, else next Monday; for `every <weekday>` the next such day strictly after today.

Then, for steps 1 and 2, the start moves forward to the first day the pattern allows: `every monday` on a todo due Wednesday 30 September starts Monday 5 October (quick-add, by contrast, lets a date phrase in the same line win and repeats weekly on Wednesdays; in a Repeat box the weekday is the point), and `every weekday` from a Saturday starts the Monday after. A start already on an allowed day stays: `every friday` due Friday 2 October starts that day.

Month ends and leap days are left to todo.next-occurrence: `every month` from 31 January is anchored on the 31st, so it falls on each month's last day.

## Errors

Only a malformed or impossible `today` or `due` throws, with dates.add-days's message (`"28/09/2026" is not an ISO date`). Text that is not a repeat is null, never an error.

Files