searxng/searx/engines/piped.py

# SPDX-License-Identifier: AGPL-3.0-or-later
"""An alternative privacy-friendly YouTube frontend which is efficient by
design.  `Piped’s architecture`_ consists of 3 components:

- :py:obj:`backend <backend_url>`
- :py:obj:`frontend <frontend_url>`
- proxy

.. _Piped’s architecture: https://docs.piped.video/docs/architecture/

Configuration
=============

The :py:obj:`backend_url` and :py:obj:`frontend_url` has to be set in the engine
named `piped` and are used by all piped engines

.. code:: yaml

  - name: piped
    engine: piped
    piped_filter: videos
    ...
    frontend_url: https://..
    backend_url:
      - https://..
      - https://..

  - name: piped.music
    engine: piped
    network: piped
    shortcut: ppdm
    piped_filter: music_songs
    ...

Known Quirks
============

The implementation to support :py:obj:`paging <searx.enginelib.Engine.paging>`
is based on the *nextpage* method of Piped's REST API / the :py:obj:`frontend
API <frontend_url>`.  This feature is *next page driven* and plays well with the
:ref:`infinite_scroll <settings ui>` setting in SearXNG but it does not really
fit into SearXNG's UI to select a page by number.

Implementations
===============
"""

from __future__ import annotations

import time
import random
from urllib.parse import urlencode
import datetime
from dateutil import parser

from searx.utils import humanize_number

# about
about = {
    "website": 'https://github.com/TeamPiped/Piped/',
    "wikidata_id": 'Q107565255',
    "official_api_documentation": 'https://docs.piped.video/docs/api-documentation/',
    "use_official_api": True,
    "require_api_key": False,
    "results": 'JSON',
}

# engine dependent config
categories = []
paging = True

# search-url
backend_url: list | str = "https://pipedapi.kavin.rocks"
"""Piped-Backend_: The core component behind Piped.  The value is an URL or a
list of URLs.  In the latter case instance will be selected randomly.  For a
complete list of official instances see Piped-Instances (`JSON
<https://piped-instances.kavin.rocks/>`__)

.. _Piped-Instances: https://github.com/TeamPiped/Piped/wiki/Instances
.. _Piped-Backend: https://github.com/TeamPiped/Piped-Backend

"""

frontend_url: str = "https://piped.video"
"""Piped-Frontend_: URL to use as link and for embeds.

.. _Piped-Frontend: https://github.com/TeamPiped/Piped
"""

piped_filter = 'all'
"""Content filter ``music_songs`` or ``videos``"""


def _backend_url() -> str:
    from searx.engines import engines  # pylint: disable=import-outside-toplevel

    url = engines['piped'].backend_url  # type: ignore
    if isinstance(url, list):
        url = random.choice(url)
    return url


def _frontend_url() -> str:
    from searx.engines import engines  # pylint: disable=import-outside-toplevel

    return engines['piped'].frontend_url  # type: ignore


def request(query, params):

    args = {
        'q': query,
        'filter': piped_filter,
    }

    path = "/search"
    if params['pageno'] > 1:
        # don't use nextpage when user selected to jump back to page 1
        nextpage = params['engine_data'].get('nextpage')
        if nextpage:
            path = "/nextpage/search"
            args['nextpage'] = nextpage

    params["url"] = _backend_url() + f"{path}?" + urlencode(args)
    return params


def response(resp):
    results = []

    json = resp.json()

    for result in json["items"]:
        # note: piped returns -1 for all upload times when filtering for music
        uploaded = result.get("uploaded", -1)

        item = {
            # the api url differs from the frontend, hence use piped.video as default
            "url": _frontend_url() + result.get("url", ""),
            "title": result.get("title", ""),
            "publishedDate": parser.parse(time.ctime(uploaded / 1000)) if uploaded != -1 else None,
            "iframe_src": _frontend_url() + '/embed' + result.get("url", ""),
            "views": humanize_number(result["views"]),
        }
        length = result.get("duration")
        if length:
            item["length"] = datetime.timedelta(seconds=length)

        if piped_filter == 'videos':
            item["template"] = "videos.html"
            # if the value of shortDescription set, but is None, return empty string
            item["content"] = result.get("shortDescription", "") or ""
            item["thumbnail"] = result.get("thumbnail", "")

        elif piped_filter == 'music_songs':
            item["template"] = "default.html"
            item["thumbnail"] = result.get("thumbnail", "")
            item["content"] = result.get("uploaderName", "") or ""

        results.append(item)

    results.append(
        {
            "engine_data": json["nextpage"],
            "key": "nextpage",
        }
    )
    return results