Merge pull request #1985 from SamantazFox/improve-youtube-api-helper

Improve youtube api calls
This commit is contained in:
syeopite 2021-06-07 13:59:26 -07:00 committed by GitHub
commit d82734641b
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
4 changed files with 118 additions and 72 deletions

View file

@ -229,22 +229,8 @@ def fetch_channel(ucid, db, pull_all_videos = true, locale = nil)
page = 1 page = 1
LOGGER.trace("fetch_channel: #{ucid} : Downloading channel videos page") LOGGER.trace("fetch_channel: #{ucid} : Downloading channel videos page")
response_body = get_channel_videos_response(ucid, page, auto_generated: auto_generated) initial_data = get_channel_videos_response(ucid, page, auto_generated: auto_generated)
videos = extract_videos(initial_data, author, ucid)
videos = [] of SearchVideo
begin
initial_data = JSON.parse(response_body)
raise InfoException.new("Could not extract channel JSON") if !initial_data
LOGGER.trace("fetch_channel: #{ucid} : Extracting videos from channel videos page initial_data")
videos = extract_videos(initial_data.as_h, author, ucid)
rescue ex
if response_body.includes?("To continue with your YouTube experience, please fill out the form below.") ||
response_body.includes?("https://www.google.com/sorry/index")
raise InfoException.new("Could not extract channel info. Instance is likely blocked.")
end
raise ex
end
LOGGER.trace("fetch_channel: #{ucid} : Extracting videos from channel RSS feed") LOGGER.trace("fetch_channel: #{ucid} : Extracting videos from channel RSS feed")
rss.xpath_nodes("//feed/entry").each do |entry| rss.xpath_nodes("//feed/entry").each do |entry|
@ -304,10 +290,8 @@ def fetch_channel(ucid, db, pull_all_videos = true, locale = nil)
ids = [] of String ids = [] of String
loop do loop do
response_body = get_channel_videos_response(ucid, page, auto_generated: auto_generated) initial_data = get_channel_videos_response(ucid, page, auto_generated: auto_generated)
initial_data = JSON.parse(response_body) videos = extract_videos(initial_data, author, ucid)
raise InfoException.new("Could not extract channel JSON") if !initial_data
videos = extract_videos(initial_data.as_h, author, ucid)
count = videos.size count = videos.size
videos = videos.map { |video| ChannelVideo.new({ videos = videos.map { |video| ChannelVideo.new({
@ -358,8 +342,7 @@ end
def fetch_channel_playlists(ucid, author, continuation, sort_by) def fetch_channel_playlists(ucid, author, continuation, sort_by)
if continuation if continuation
response_json = request_youtube_api_browse(continuation) response_json = request_youtube_api_browse(continuation)
result = JSON.parse(response_json) continuationItems = response_json["onResponseReceivedActions"]?
continuationItems = result["onResponseReceivedActions"]?
.try &.[0]["appendContinuationItemsAction"]["continuationItems"] .try &.[0]["appendContinuationItemsAction"]["continuationItems"]
return [] of SearchItem, nil if !continuationItems return [] of SearchItem, nil if !continuationItems
@ -964,21 +947,16 @@ def get_60_videos(ucid, author, page, auto_generated, sort_by = "newest")
videos = [] of SearchVideo videos = [] of SearchVideo
2.times do |i| 2.times do |i|
response_json = get_channel_videos_response(ucid, page * 2 + (i - 1), auto_generated: auto_generated, sort_by: sort_by) initial_data = get_channel_videos_response(ucid, page * 2 + (i - 1), auto_generated: auto_generated, sort_by: sort_by)
initial_data = JSON.parse(response_json) videos.concat extract_videos(initial_data, author, ucid)
break if !initial_data
videos.concat extract_videos(initial_data.as_h, author, ucid)
end end
return videos.size, videos return videos.size, videos
end end
def get_latest_videos(ucid) def get_latest_videos(ucid)
response_json = get_channel_videos_response(ucid) initial_data = get_channel_videos_response(ucid)
initial_data = JSON.parse(response_json)
return [] of SearchVideo if !initial_data
author = initial_data["metadata"]?.try &.["channelMetadataRenderer"]?.try &.["title"]?.try &.as_s author = initial_data["metadata"]?.try &.["channelMetadataRenderer"]?.try &.["title"]?.try &.as_s
items = extract_videos(initial_data.as_h, author, ucid)
return items return extract_videos(initial_data, author, ucid)
end end

View file

@ -4,28 +4,116 @@
# Hard-coded constants required by the API # Hard-coded constants required by the API
HARDCODED_API_KEY = "AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8" HARDCODED_API_KEY = "AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8"
HARDCODED_CLIENT_VERS = "2.20210318.08.00" HARDCODED_CLIENT_VERS = "2.20210330.08.00"
def request_youtube_api_browse(continuation) ####################################################################
# make_youtube_api_context(region)
#
# Return, as a Hash, the "context" data required to request the
# youtube API endpoints.
#
def make_youtube_api_context(region : String | Nil) : Hash
return {
"client" => {
"hl" => "en",
"gl" => region || "US", # Can't be empty!
"clientName" => "WEB",
"clientVersion" => HARDCODED_CLIENT_VERS,
},
}
end
####################################################################
# request_youtube_api_browse(continuation)
# request_youtube_api_browse(browse_id, params)
#
# Requests the youtubei/v1/browse endpoint with the required headers
# and POST data in order to get a JSON reply in english US that can
# be easily parsed.
#
# The requested data can either be:
#
# - A continuation token (ctoken). Depending on this token's
# contents, the returned data can be comments, playlist videos,
# search results, channel community tab, ...
#
# - A playlist ID (parameters MUST be an empty string)
#
def request_youtube_api_browse(continuation : String)
# JSON Request data, required by the API # JSON Request data, required by the API
data = { data = {
"context": { "context" => make_youtube_api_context("US"),
"client": { "continuation" => continuation,
"hl": "en",
"gl": "US",
"clientName": "WEB",
"clientVersion": HARDCODED_CLIENT_VERS,
},
},
"continuation": continuation,
} }
# Send the POST request and return result return _youtube_api_post_json("/youtubei/v1/browse", data)
end
def request_youtube_api_browse(browse_id : String, params : String)
# JSON Request data, required by the API
data = {
"browseId" => browse_id,
"context" => make_youtube_api_context("US"),
}
# Append the additionnal parameters if those were provided
# (this is required for channel info, playlist and community, e.g)
if params != ""
data["params"] = params
end
return _youtube_api_post_json("/youtubei/v1/browse", data)
end
####################################################################
# request_youtube_api_search(search_query, params, region)
#
# Requests the youtubei/v1/search endpoint with the required headers
# and POST data in order to get a JSON reply. As the search results
# vary depending on the region, a region code can be specified in
# order to get non-US results.
#
# The requested data is a search string, with some additional
# paramters, formatted as a base64 string.
#
def request_youtube_api_search(search_query : String, params : String, region = nil)
# JSON Request data, required by the API
data = {
"query" => URI.encode_www_form(search_query),
"context" => make_youtube_api_context(region),
"params" => params,
}
return _youtube_api_post_json("/youtubei/v1/search", data)
end
####################################################################
# _youtube_api_post_json(endpoint, data)
#
# Internal function that does the actual request to youtube servers
# and handles errors.
#
# The requested data is an endpoint (URL without the domain part)
# and the data as a Hash object.
#
def _youtube_api_post_json(endpoint, data)
# Send the POST request and parse result
response = YT_POOL.client &.post( response = YT_POOL.client &.post(
"/youtubei/v1/browse?key=#{HARDCODED_API_KEY}", "#{endpoint}?key=#{HARDCODED_API_KEY}",
headers: HTTP::Headers{"content-type" => "application/json"}, headers: HTTP::Headers{"content-type" => "application/json; charset=UTF-8"},
body: data.to_json body: data.to_json
) )
return response.body initial_data = JSON.parse(response.body).as_h
# Error handling
if initial_data.has_key?("error")
code = initial_data["error"]["code"]
message = initial_data["error"]["message"].to_s.sub(/(\\n)+\^$/, "")
raise InfoException.new("Could not extract JSON. Youtube API returned \
error #{code} with message:<br>\"#{message}\"")
end
return initial_data
end end

View file

@ -361,16 +361,7 @@ def fetch_playlist(plid, locale)
plid = "UU#{plid.lchop("UC")}" plid = "UU#{plid.lchop("UC")}"
end end
response = YT_POOL.client &.get("/playlist?list=#{plid}&hl=en") initial_data = request_youtube_api_browse("VL" + plid, params: "")
if response.status_code != 200
if response.headers["location"]?.try &.includes? "/sorry/index"
raise InfoException.new("Could not extract playlist info. Instance is likely blocked.")
else
raise InfoException.new("Not a playlist.")
end
end
initial_data = extract_initial_data(response.body)
playlist_sidebar_renderer = initial_data["sidebar"]?.try &.["playlistSidebarRenderer"]?.try &.["items"]? playlist_sidebar_renderer = initial_data["sidebar"]?.try &.["playlistSidebarRenderer"]?.try &.["items"]?
raise InfoException.new("Could not extract playlistSidebarRenderer.") if !playlist_sidebar_renderer raise InfoException.new("Could not extract playlistSidebarRenderer.") if !playlist_sidebar_renderer
@ -451,17 +442,12 @@ def get_playlist_videos(db, playlist, offset, locale = nil, continuation = nil)
offset = (offset / 100).to_i64 * 100_i64 offset = (offset / 100).to_i64 * 100_i64
ctoken = produce_playlist_continuation(playlist.id, offset) ctoken = produce_playlist_continuation(playlist.id, offset)
initial_data = JSON.parse(request_youtube_api_browse(ctoken)).as_h initial_data = request_youtube_api_browse(ctoken)
else else
response = YT_POOL.client &.get("/playlist?list=#{playlist.id}&gl=US&hl=en") initial_data = request_youtube_api_browse("VL" + playlist.id, params: "")
initial_data = extract_initial_data(response.body)
end end
if initial_data return extract_playlist_videos(initial_data)
return extract_playlist_videos(initial_data)
else
return [] of PlaylistVideo
end
end end
end end

View file

@ -246,8 +246,7 @@ def channel_search(query, page, channel)
continuation = produce_channel_search_continuation(ucid, query, page) continuation = produce_channel_search_continuation(ucid, query, page)
response_json = request_youtube_api_browse(continuation) response_json = request_youtube_api_browse(continuation)
result = JSON.parse(response_json) continuationItems = response_json["onResponseReceivedActions"]?
continuationItems = result["onResponseReceivedActions"]?
.try &.[0]["appendContinuationItemsAction"]["continuationItems"] .try &.[0]["appendContinuationItemsAction"]["continuationItems"]
return 0, [] of SearchItem if !continuationItems return 0, [] of SearchItem if !continuationItems
@ -264,14 +263,9 @@ end
def search(query, search_params = produce_search_params(content_type: "all"), region = nil) def search(query, search_params = produce_search_params(content_type: "all"), region = nil)
return 0, [] of SearchItem if query.empty? return 0, [] of SearchItem if query.empty?
body = YT_POOL.client(region, &.get("/results?search_query=#{URI.encode_www_form(query)}&sp=#{search_params}&hl=en").body) initial_data = request_youtube_api_search(query, search_params, region)
return 0, [] of SearchItem if body.empty?
initial_data = extract_initial_data(body)
items = extract_items(initial_data) items = extract_items(initial_data)
# initial_data["estimatedResults"]?.try &.as_s.to_i64
return items.size, items return items.size, items
end end