GET /schedule/work_day_rules

Which calendar days count as work. duration and due_date on a Task are computed against this, so a client that adds working days itself has to read it.

Params

Sample requests

A fortnight, studio-wide:

r = c.get("/schedule/work_day_rules",
          params={"start_date": "2026-03-02", "end_date": "2026-03-15"})

14 rows for a 14-day window, first two shown:

{
  "data": [
    {"date": "2026-03-02", "working": true, "description": null, "reason": "STUDIO_WORK_WEEK"},
    {"date": "2026-03-07", "working": false, "description": null, "reason": "STUDIO_WORK_WEEK"}
  ],
  "links": {"self": "/api/v1/schedule/work_day_rules?end_date=2026-03-15&start_date=2026-03-02"}
}

A project id that is not on the site, which is not an error:

r = c.get("/schedule/work_day_rules",
          params={"start_date": "2026-03-02", "end_date": "2026-03-15",
                  "project_id": 999999999})

{"data": [{"date": "2026-03-02", "working": true, "description": null,
           "reason": "STUDIO_WORK_WEEK"}]}

A date that will not parse:

r = c.get("/schedule/work_day_rules",
          params={"start_date": "03/02/2026", "end_date": "03/15/2026"})

{"status": "error", "error": "invalid date"}

A missing parameter:

r = c.get("/schedule/work_day_rules", params={"start_date": "2026-03-02"})

[{"id": "6e3ab12946d1c597355a5373e4ccca44", "status": 400, "code": 103,
  "title": "Request Parameters invalid.", "source": {"end_date": ["end_date is missing"]},
  "detail": null, "meta": null}]

Response codes

Edge cases

• Two error shapes on one endpoint. A missing or out-of-order parameter is a JSON:API errors array; an unparseable date is a bare {"status": "error", "error": "invalid date"} with no errors key at all. A client reading r.json()["errors"][0]["title"] raises KeyError on the second.
• A project_id or user_id that is not on the site answers 200 with the studio rule. Nothing in the body says which scope answered, other than reason, and reason reads STUDIO_WORK_WEEK for both the fallback and a genuine studio-wide answer. Check the id exists before trusting the schedule.
• Both ends are inclusive. start_date equal to end_date returns one row.
• No paging. A 730-day window returned 730 rows in 61631 bytes, with no page envelope and no links.next. Bound the range yourself.
• links.self echoes the parameters back, so two responses that differ only in project_id differ in byte length while their data is identical.
• The dates must be YYYY-MM-DD. 03/02/2026 is refused, whatever date_component_order in GET /preferences says the site displays.