Update get_about_info for polymer (iv-org/invidious#1423) (#1429)

Update get_about_info for polymer (iv-org/invidious#1423)
This commit is contained in:
Triplesalt 2020-10-23 03:40:07 +02:00 committed by GitHub
parent 22d9d16a7a
commit ec4a22687f
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 85 additions and 51 deletions

View file

@ -775,10 +775,12 @@ def extract_channel_community_cursor(continuation)
cursor cursor
end end
INITDATA_PREQUERY = "window[\"ytInitialData\"] = {"
def get_about_info(ucid, locale) def get_about_info(ucid, locale)
about = YT_POOL.client &.get("/channel/#{ucid}/about?disable_polymer=1&gl=US&hl=en") about = YT_POOL.client &.get("/channel/#{ucid}/about?gl=US&hl=en")
if about.status_code != 200 if about.status_code != 200
about = YT_POOL.client &.get("/user/#{ucid}/about?disable_polymer=1&gl=US&hl=en") about = YT_POOL.client &.get("/user/#{ucid}/about?gl=US&hl=en")
end end
if md = about.headers["location"]?.try &.match(/\/channel\/(?<ucid>UC[a-zA-Z0-9_-]{22})/) if md = about.headers["location"]?.try &.match(/\/channel\/(?<ucid>UC[a-zA-Z0-9_-]{22})/)
@ -790,6 +792,17 @@ def get_about_info(ucid, locale)
raise error_message raise error_message
end end
initdata_pre = about.body.index(INITDATA_PREQUERY)
initdata_post = initdata_pre.nil? ? nil : about.body.index("};", initdata_pre)
if initdata_post.nil?
about = XML.parse_html(about.body)
error_message = about.xpath_node(%q(//div[@class="yt-alert-content"])).try &.content.strip
error_message ||= translate(locale, "Could not get channel info.")
raise error_message
end
initdata_pre = initdata_pre.not_nil! + INITDATA_PREQUERY.size - 1
initdata = JSON.parse(about.body[initdata_pre, initdata_post - initdata_pre + 1])
about = XML.parse_html(about.body) about = XML.parse_html(about.body)
if about.xpath_node(%q(//div[contains(@class, "channel-empty-message")])) if about.xpath_node(%q(//div[contains(@class, "channel-empty-message")]))
@ -797,73 +810,94 @@ def get_about_info(ucid, locale)
raise error_message raise error_message
end end
if about.xpath_node(%q(//span[contains(@class,"qualified-channel-title-text")]/a)).try &.content.empty? author = about.xpath_node(%q(//meta[@name="title"])).not_nil!["content"]
error_message = about.xpath_node(%q(//div[@class="yt-alert-content"])).try &.content.strip author_url = about.xpath_node(%q(//link[@rel="canonical"])).not_nil!["href"]
error_message ||= translate(locale, "Could not get channel info.") author_thumbnail = about.xpath_node(%q(//link[@rel="image_src"])).not_nil!["href"]
raise error_message
end
author = about.xpath_node(%q(//span[contains(@class,"qualified-channel-title-text")]/a)).not_nil!.content
author_url = about.xpath_node(%q(//span[contains(@class,"qualified-channel-title-text")]/a)).not_nil!["href"]
author_thumbnail = about.xpath_node(%q(//img[@class="channel-header-profile-image"])).not_nil!["src"]
ucid = about.xpath_node(%q(//meta[@itemprop="channelId"])).not_nil!["content"] ucid = about.xpath_node(%q(//meta[@itemprop="channelId"])).not_nil!["content"]
banner = about.xpath_node(%q(//div[@id="gh-banner"]/style)).not_nil!.content # Raises a KeyError on failure.
banner = "https:" + banner.match(/background-image: url\((?<url>[^)]+)\)/).not_nil!["url"] banners = initdata["header"]["c4TabbedHeaderRenderer"]?.try &.["banner"]?.try &.["thumbnails"]?
banner = banners.try &.[-1]?.try &.["url"].as_s?
if banner.includes? "channels/c4/default_banner" # if banner.includes? "channels/c4/default_banner"
banner = nil # banner = nil
end # end
description_html = about.xpath_node(%q(//div[contains(@class,"about-description")])).try &.to_s || description = initdata["metadata"]["channelMetadataRenderer"]?.try &.["description"]?.try &.as_s? || ""
%(<div class="about-description branded-page-box-padding"><pre></pre></div>) description_html = HTML.escape(description).gsub("\n", "<br>")
paid = about.xpath_node(%q(//meta[@itemprop="paid"])).not_nil!["content"] == "True" paid = about.xpath_node(%q(//meta[@itemprop="paid"])).not_nil!["content"] == "True"
is_family_friendly = about.xpath_node(%q(//meta[@itemprop="isFamilyFriendly"])).not_nil!["content"] == "True" is_family_friendly = about.xpath_node(%q(//meta[@itemprop="isFamilyFriendly"])).not_nil!["content"] == "True"
allowed_regions = about.xpath_node(%q(//meta[@itemprop="regionsAllowed"])).not_nil!["content"].split(",") allowed_regions = about.xpath_node(%q(//meta[@itemprop="regionsAllowed"])).not_nil!["content"].split(",")
related_channels = about.xpath_nodes(%q(//div[contains(@class, "branded-page-related-channels")]/ul/li)) related_channels = initdata["contents"]["twoColumnBrowseResultsRenderer"]
related_channels = related_channels.map do |node| .["secondaryContents"]?.try &.["browseSecondaryContentsRenderer"]["contents"][0]?
related_id = node["data-external-id"]? .try &.["verticalChannelSectionRenderer"]?.try &.["items"]?.try &.as_a.map do |node|
related_id ||= "" renderer = node["miniChannelRenderer"]?
related_id = renderer.try &.["channelId"]?.try &.as_s?
related_id ||= ""
anchor = node.xpath_node(%q(.//h3[contains(@class, "yt-lockup-title")]/a)) related_title = renderer.try &.["title"]?.try &.["simpleText"]?.try &.as_s?
related_title = anchor.try &.["title"] related_title ||= ""
related_title ||= ""
related_author_url = anchor.try &.["href"] related_author_url = renderer.try &.["navigationEndpoint"]?.try &.["commandMetadata"]?.try &.["webCommandMetadata"]?
related_author_url ||= "" .try &.["url"]?.try &.as_s?
related_author_url ||= ""
related_author_thumbnail = node.xpath_node(%q(.//img)).try &.["data-thumb"] related_author_thumbnails = renderer.try &.["thumbnail"]?.try &.["thumbnails"]?.try &.as_a?
related_author_thumbnail ||= "" related_author_thumbnails ||= [] of JSON::Any
AboutRelatedChannel.new({ related_author_thumbnail = ""
ucid: related_id, if related_author_thumbnails.size > 0
author: related_title, related_author_thumbnail = related_author_thumbnails[-1]["url"]?.try &.as_s?
author_url: related_author_url, related_author_thumbnail ||= ""
author_thumbnail: related_author_thumbnail, end
})
end
joined = about.xpath_node(%q(//span[contains(., "Joined")])) AboutRelatedChannel.new({
.try &.content.try { |text| Time.parse(text, "Joined %b %-d, %Y", Time::Location.local) } || Time.unix(0) ucid: related_id,
author: related_title,
author_url: related_author_url,
author_thumbnail: related_author_thumbnail,
})
end
related_channels ||= [] of AboutRelatedChannel
total_views = about.xpath_node(%q(//span[contains(., "views")]/b)) total_views = 0_i64
.try &.content.try &.gsub(/\D/, "").to_i64? || 0_i64 joined = Time.unix(0)
tabs = [] of String
sub_count = about.xpath_node(%q(.//span[contains(@class, "subscriber-count")]))
.try &.["title"].try { |text| short_text_to_number(text) } || 0
# Auto-generated channels
# https://support.google.com/youtube/answer/2579942
auto_generated = false auto_generated = false
if about.xpath_node(%q(//ul[@class="about-custom-links"]/li/a[@title="Auto-generated by YouTube"])) ||
about.xpath_node(%q(//span[@class="qualified-channel-title-badge"]/span[@title="Auto-generated by YouTube"])) tabs_json = initdata["contents"]["twoColumnBrowseResultsRenderer"]["tabs"]?.try &.as_a?
auto_generated = true if !tabs_json.nil?
# Retrieve information from the tabs array. The index we are looking for varies between channels.
tabs_json.each do |node|
# Try to find the about section which is located in only one of the tabs.
channel_about_meta = node["tabRenderer"]?.try &.["content"]?.try &.["sectionListRenderer"]?
.try &.["contents"]?.try &.[0]?.try &.["itemSectionRenderer"]?.try &.["contents"]?
.try &.[0]?.try &.["channelAboutFullMetadataRenderer"]?
if !channel_about_meta.nil?
total_views = channel_about_meta["viewCountText"]?.try &.["simpleText"]?.try &.as_s.gsub(/\D/, "").to_i64? || 0_i64
# The joined text is split to several sub strings. The reduce joins those strings before parsing the date.
joined = channel_about_meta["joinedDateText"]?.try &.["runs"]?.try &.as_a.reduce("") { |acc, node| acc + node["text"].as_s }
.try { |text| Time.parse(text, "Joined %b %-d, %Y", Time::Location.local) } || Time.unix(0)
# Auto-generated channels
# https://support.google.com/youtube/answer/2579942
# For auto-generated channels, channel_about_meta only has ["description"]["simpleText"] and ["primaryLinks"][0]["title"]["simpleText"]
if (channel_about_meta["primaryLinks"]?.try &.size || 0) == 1 && (channel_about_meta["primaryLinks"][0]?)
(channel_about_meta["primaryLinks"][0]["title"]?.try &.["simpleText"]?.try &.as_s? || "") == "Auto-generated by YouTube"
auto_generated = true
end
end
end
tabs = tabs_json.reject { |node| node["tabRenderer"]?.nil? }.map { |node| node["tabRenderer"]["title"].as_s.downcase }
end end
tabs = about.xpath_nodes(%q(//ul[@id="channel-navigation-menu"]/li/a/span)).map { |node| node.content.downcase } sub_count = initdata["header"]["c4TabbedHeaderRenderer"]?.try &.["subscriberCountText"]?.try &.["simpleText"]?.try &.as_s?
.try { |text| short_text_to_number(text.split(" ")[0]) } || 0
AboutChannel.new({ AboutChannel.new({
ucid: ucid, ucid: ucid,

View file

@ -28,7 +28,7 @@
</div> </div>
<div class="h-box"> <div class="h-box">
<p><span style="white-space:pre-wrap"><%= XML.parse_html(channel.description_html).xpath_node(%q(.//pre)).try &.content %></span></p> <p><span style="white-space:pre-wrap"><%= channel.description_html %></span></p>
</div> </div>
<div class="h-box"> <div class="h-box">