diff options
Diffstat (limited to 'searx/plugins/oa_doi_rewrite.py')
| -rw-r--r-- | searx/plugins/oa_doi_rewrite.py | 106 |
1 files changed, 71 insertions, 35 deletions
diff --git a/searx/plugins/oa_doi_rewrite.py b/searx/plugins/oa_doi_rewrite.py index be5a8d4a4..dac60d298 100644 --- a/searx/plugins/oa_doi_rewrite.py +++ b/searx/plugins/oa_doi_rewrite.py @@ -1,54 +1,90 @@ # SPDX-License-Identifier: AGPL-3.0-or-later # pylint: disable=missing-module-docstring - from __future__ import annotations +import typing + import re -from urllib.parse import urlparse, parse_qsl +from urllib.parse import parse_qsl from flask_babel import gettext +from searx import get_setting +from searx.plugins import Plugin, PluginInfo +from searx.extended_types import sxng_request -from searx import settings +from ._core import log +if typing.TYPE_CHECKING: + from searx.search import SearchWithPlugins + from searx.extended_types import SXNG_Request + from searx.result_types import Result, LegacyResult + from searx.plugins import PluginCfg -regex = re.compile(r'10\.\d{4,9}/[^\s]+') -name = gettext('Open Access DOI rewrite') -description = gettext('Avoid paywalls by redirecting to open-access versions of publications when available') -default_on = False -preference_section = 'general/doi_resolver' +ahmia_blacklist: list = [] -def extract_doi(url): - match = regex.search(url.path) - if match: - return match.group(0) - for _, v in parse_qsl(url.query): - match = regex.search(v) - if match: - return match.group(0) - return None +def filter_url_field(result: "Result|LegacyResult", field_name: str, url_src: str) -> bool | str: + """Returns bool ``True`` to use URL unchanged (``False`` to ignore URL). + If URL should be modified, the returned string is the new URL to use.""" + if field_name != "url": + return True # use it unchanged -def get_doi_resolver(preferences): - doi_resolvers = settings['doi_resolvers'] - selected_resolver = preferences.get_value('doi_resolver')[0] - if selected_resolver not in doi_resolvers: - selected_resolver = settings['default_doi_resolver'] - return doi_resolvers[selected_resolver] + doi = extract_doi(result.parsed_url) + if doi and len(doi) < 50: + for suffix in ("/", ".pdf", ".xml", "/full", "/meta", "/abstract"): + doi = doi.removesuffix(suffix) + new_url = get_doi_resolver() + doi + if "doi" not in result: + result["doi"] = doi + log.debug("oa_doi_rewrite: [URL field: %s] %s -> %s", field_name, url_src, new_url) + return new_url # use new url + + return True # use it unchanged + + +class SXNGPlugin(Plugin): + """Avoid paywalls by redirecting to open-access.""" + id = "oa_doi_rewrite" -def on_result(request, _search, result) -> bool: + def __init__(self, plg_cfg: "PluginCfg") -> None: + super().__init__(plg_cfg) + self.info = PluginInfo( + id=self.id, + name=gettext("Open Access DOI rewrite"), + description=gettext("Avoid paywalls by redirecting to open-access versions of publications when available"), + preference_section="general", + ) - if not result.parsed_url: + def on_result( + self, + request: "SXNG_Request", + search: "SearchWithPlugins", + result: "Result", + ) -> bool: # pylint: disable=unused-argument + if result.parsed_url: + result.filter_urls(filter_url_field) return True - doi = extract_doi(result['parsed_url']) - if doi and len(doi) < 50: - for suffix in ('/', '.pdf', '.xml', '/full', '/meta', '/abstract'): - if doi.endswith(suffix): - doi = doi[: -len(suffix)] - result['url'] = get_doi_resolver(request.preferences) + doi - result['parsed_url'] = urlparse(result['url']) - if 'doi' not in result: - result['doi'] = doi - return True + +regex = re.compile(r'10\.\d{4,9}/[^\s]+') + + +def extract_doi(url): + m = regex.search(url.path) + if m: + return m.group(0) + for _, v in parse_qsl(url.query): + m = regex.search(v) + if m: + return m.group(0) + return None + + +def get_doi_resolver() -> str: + doi_resolvers = get_setting("doi_resolvers") + selected_resolver = sxng_request.preferences.get_value('doi_resolver')[0] + if selected_resolver not in doi_resolvers: + selected_resolver = get_setting("default_doi_resolver") + return doi_resolvers[selected_resolver] |