1
0
mirror of https://github.com/searxng/searxng.git synced 2024-11-18 02:10:12 +01:00
searxng/searx/engines/bandcamp.py

96 lines
2.8 KiB
Python
Raw Normal View History

# SPDX-License-Identifier: AGPL-3.0-or-later
# lint: pylint
"""Bandcamp (Music)
2020-12-28 05:46:11 +01:00
@website https://bandcamp.com/
@provide-api no
@results HTML
@parse url, title, content, publishedDate, iframe_src, thumbnail
2020-12-28 05:46:11 +01:00
"""
from urllib.parse import urlencode, urlparse, parse_qs
from dateutil.parser import parse as dateparse
from lxml import html
from searx.utils import (
eval_xpath_getindex,
eval_xpath_list,
extract_text,
)
2020-12-28 05:46:11 +01:00
# about
about = {
"website": 'https://bandcamp.com/',
"wikidata_id": 'Q545966',
"official_api_documentation": 'https://bandcamp.com/developer',
"use_official_api": False,
"require_api_key": False,
"results": 'HTML',
}
2020-12-28 05:46:11 +01:00
categories = ['music']
paging = True
base_url = "https://bandcamp.com/"
search_string = 'search?{query}&page={page}'
iframe_src = "https://bandcamp.com/EmbeddedPlayer/{type}={result_id}/size=large/bgcol=000/linkcol=fff/artwork=small"
2020-12-28 05:46:11 +01:00
def request(query, params):
'''pre-request callback
2020-12-28 05:46:11 +01:00
params<dict>:
method : POST/GET
headers : {}
data : {} # if method == POST
url : ''
category: 'search category'
pageno : 1 # number of the requested page
'''
search_path = search_string.format(query=urlencode({'q': query}), page=params['pageno'])
2020-12-28 05:46:11 +01:00
params['url'] = base_url + search_path
return params
def response(resp):
'''post-response callback
2020-12-28 05:46:11 +01:00
resp: requests response object
'''
results = []
dom = html.fromstring(resp.text)
for result in eval_xpath_list(dom, '//li[contains(@class, "searchresult")]'):
link = eval_xpath_getindex(result, './/div[@class="itemurl"]/a', 0, default=None)
if link is None:
continue
title = result.xpath('.//div[@class="heading"]/a/text()')
content = result.xpath('.//div[@class="subhead"]/text()')
2020-12-28 05:46:11 +01:00
new_result = {
"url": extract_text(link),
"title": extract_text(title),
"content": extract_text(content),
2020-12-28 05:46:11 +01:00
}
date = eval_xpath_getindex(result, '//div[@class="released"]/text()', 0, default=None)
if date:
new_result["publishedDate"] = dateparse(date.replace("released ", ""))
thumbnail = result.xpath('.//div[@class="art"]/img/@src')
if thumbnail:
new_result['img_src'] = thumbnail[0]
result_id = parse_qs(urlparse(link.get('href')).query)["search_item_id"][0]
itemtype = extract_text(result.xpath('.//div[@class="itemtype"]')).lower()
if "album" == itemtype:
new_result["iframe_src"] = iframe_src.format(type='album', result_id=result_id)
elif "track" == itemtype:
new_result["iframe_src"] = iframe_src.format(type='track', result_id=result_id)
2020-12-28 05:46:11 +01:00
results.append(new_result)
return results