kadence/backend/app/sources/google_tasks.py

"""Google Tasks as a todo source.

OAuth 2.0: the long-lived refresh token from the settings is exchanged for an
access token per run (cached for the life of the object). Obtain the refresh
token with jira-helper's `uv run google-tasks-auth` — the credentials are
interchangeable, so an existing setup works here unchanged.

Unlike jira-helper, undated tasks are kept by default: "someday" items are
still things you can spend a pomodoro on.
"""

from __future__ import annotations

from collections.abc import Iterator
from datetime import date
from typing import Any

import httpx

from app.settings_schema import GoogleTasksSourceSettings
from app.sources.base import SourceError, TodoItem

_TOKEN_URL = "https://oauth2.googleapis.com/token"
_API_BASE = "https://tasks.googleapis.com/tasks/v1"
_PAGE_SIZE = 100


class GoogleTasksSource:
    name = "google_tasks"

    def __init__(self, config: GoogleTasksSourceSettings, *, timeout: int = 30) -> None:
        if not (config.client_id and config.client_secret and config.refresh_token):
            raise SourceError(
                "Google Tasks needs a client id, client secret and refresh token."
            )
        self._config = config
        self._client = httpx.Client(timeout=timeout)
        self._access_token: str | None = None

    # ---- plumbing ----------------------------------------------------------

    def _token(self) -> str:
        if self._access_token:
            return self._access_token
        try:
            resp = self._client.post(
                _TOKEN_URL,
                data={
                    "client_id": self._config.client_id,
                    "client_secret": self._config.client_secret,
                    "refresh_token": self._config.refresh_token,
                    "grant_type": "refresh_token",
                },
            )
        except httpx.HTTPError as exc:
            raise SourceError(f"Google is unreachable: {exc}") from exc
        if not resp.is_success:
            raise SourceError(
                "Could not refresh the Google access token "
                f"({resp.status_code}). Re-run the Google auth flow for a fresh "
                "refresh token."
            )
        token = (resp.json() or {}).get("access_token")
        if not token:
            raise SourceError("Google's token response carried no access_token.")
        self._access_token = token
        return token

    def _get(self, path: str, params: dict[str, Any]) -> dict[str, Any]:
        try:
            resp = self._client.get(
                f"{_API_BASE}{path}",
                params=params,
                headers={"Authorization": f"Bearer {self._token()}"},
            )
        except httpx.HTTPError as exc:
            raise SourceError(f"Google is unreachable: {exc}") from exc
        if not resp.is_success:
            raise SourceError(f"Google GET {path} -> {resp.status_code}: {resp.text[:300]}")
        return resp.json() or {}

    def _paginate(self, path: str, params: dict[str, Any]) -> Iterator[dict[str, Any]]:
        params = dict(params)
        while True:
            data = self._get(path, params)
            yield from data.get("items", [])
            token = data.get("nextPageToken")
            if not token:
                return
            params["pageToken"] = token

    def _task_lists(self) -> list[dict[str, Any]]:
        lists = list(self._paginate("/users/@me/lists", {"maxResults": _PAGE_SIZE}))
        if not self._config.lists:
            return lists
        wanted = set(self._config.lists)
        return [
            task_list
            for task_list in lists
            if wanted & {task_list.get("id", ""), task_list.get("title", "")}
        ]

    # ---- source API --------------------------------------------------------

    def fetch(self) -> list[TodoItem]:
        items: list[TodoItem] = []
        for task_list in self._task_lists():
            list_id = task_list.get("id")
            if not list_id:
                continue
            list_title = task_list.get("title") or "(untitled list)"
            for raw in self._paginate(
                f"/lists/{list_id}/tasks",
                {
                    "maxResults": _PAGE_SIZE,
                    "showCompleted": "false",
                    "showHidden": "false",
                    "showDeleted": "false",
                },
            ):
                if raw.get("status") == "completed":
                    continue
                due = _parse_due(raw.get("due"))
                if due is None and not self._config.include_undated:
                    continue
                task_id = raw.get("id") or ""
                items.append(
                    TodoItem(
                        source=self.name,
                        # The list id is part of every write path, so it has to
                        # survive the round-trip through our database.
                        external_id=f"{list_id}:{task_id}",
                        title=(raw.get("title") or "").strip() or "(no title)",
                        description=(raw.get("notes") or "").strip() or None,
                        url=raw.get("webViewLink"),
                        project=list_title,
                        due_date=due,
                        raw={"list_id": list_id, "task_id": task_id},
                    )
                )
        return items

    def check(self) -> str:
        return f"{len(self._task_lists())} Google task list(s) visible"

    def close(self) -> None:
        self._client.close()


def _parse_due(raw: Any) -> date | None:
    """Google returns RFC-3339, but only the date part is ever meaningful."""
    if not raw:
        return None
    try:
        return date.fromisoformat(str(raw)[:10])
    except ValueError:
        return None