rill-node 0.0.0-stage → 2026.11.14

package/native-src/src/rill/extractors/rumble.cr

@@ -0,0 +1,325 @@
+require "html"
+require "json"
+require "set"
+require "uri"
+
+require "../errors"
+require "../extractor"
+require "../format"
+require "../http_url"
+require "../json_value"
+require "../media"
+
+module Rill::Extractors
+  class Rumble < Extractor
+    DEFAULT_BASE_URL  = "https://rumble.com"
+    NORMAL_PATH       = %r{\A/(v(?!ideos\b)[A-Za-z0-9]+)(?:-[^/]*)?(?:\.html)?/?\z}
+    EMBED_PATH        = %r{\A/embed/(v[A-Za-z0-9]+)/?\z}
+    EMBED_REFERENCE   = %r{(?:(?:https?:)?//(?:www\.)?rumble\.com)?/embed/(v[A-Za-z0-9]+)}i
+    TITLE_PATTERN     = %r{<title[^>]*>(.*?)</title>}mi
+    TITLE_AUTHOR      = /\A.+?\s-\s([^-]+?)\s-\sRumble\z/i
+    ARTIST_PATTERN    = /"content_artist"\s*:\s*"((?:\\.|[^"\\])*)"/
+    THUMBNAIL_PATTERN = /"i"\s*:\s*"((?:\\.|[^"\\])*)"/
+    DURATION_PATTERN  = /"duration"\s*:\s*(\d+)/
+
+    def initialize(@base_url : String = DEFAULT_BASE_URL)
+      @base_url = @base_url.rstrip('/')
+    end
+
+    def name : String
+      "rumble"
+    end
+
+    def matches?(url : URI) : Bool
+      return false unless url.scheme == "http" || url.scheme == "https"
+      return false unless host?(url.host)
+
+      !identifier_for(url).nil?
+    end
+
+    def extract(url : URI, client : HTTPClient) : Media
+      identifier = identifier_for(url)
+      unless identifier
+        raise UnsupportedUrlError.new(url.to_s)
+      end
+
+      embed_id = embed_identifier(url, client)
+      embed_url = URI.parse("#{@base_url}/embed/#{embed_id}")
+      body = fetch(client, embed_url, url)
+      config = config_for(body, embed_id)
+      raw_title = raw_title(body)
+
+      formats = formats_from(config)
+      if formats.empty?
+        raise MediaUnavailableError.new(url.to_s, "no playable variants were found")
+      end
+
+      Media.new(
+        id: embed_id,
+        title: title_for(raw_title, embed_id),
+        webpage_url: url,
+        uploader: uploader_for(body, raw_title),
+        duration_seconds: duration_for(config, body),
+        thumbnail_url: thumbnail_for(config, body),
+        extractor: name,
+        formats: formats,
+      )
+    end
+
+    private def identifier_for(url : URI) : String?
+      path = url.path
+      match = NORMAL_PATH.match(path) || EMBED_PATH.match(path)
+      match.try(&.[1])
+    end
+
+    private def embed_identifier(url : URI, client : HTTPClient) : String
+      identifier = identifier_for(url).not_nil!
+      return identifier if EMBED_PATH.match(url.path)
+
+      body = fetch(client, url, url)
+      match = EMBED_REFERENCE.match(body)
+      unless match
+        raise MediaUnavailableError.new(url.to_s, "page has no embed URL")
+      end
+
+      match[1]
+    end
+
+    private def fetch(client : HTTPClient, url : URI, original_url : URI) : String
+      client.get(url).body
+    rescue ex : HTTPError
+      if ex.status.code == 404 || ex.status.code == 410
+        raise MediaUnavailableError.new(original_url.to_s, "page is unavailable (#{ex.status.code})")
+      end
+
+      raise ex
+    end
+
+    private def host?(host : String?) : Bool
+      case host.try(&.downcase)
+      when "rumble.com", "www.rumble.com"
+        true
+      else
+        false
+      end
+    end
+
+    private def raw_title(body : String) : String
+      match = TITLE_PATTERN.match(body)
+      match ? HTML.unescape(match[1]).strip : ""
+    end
+
+    private def title_for(raw_title : String, identifier : String) : String
+      title = raw_title.sub(/\s*-\s*Rumble\z/i, "").strip
+      title.empty? ? identifier : title
+    end
+
+    private def uploader_for(body : String, raw_title : String) : String?
+      if match = ARTIST_PATTERN.match(body)
+        artist = json_unescape(match[1])
+        return artist unless artist.try(&.empty?)
+      end
+
+      title_author(raw_title)
+    end
+
+    private def title_author(raw_title : String) : String?
+      match = TITLE_AUTHOR.match(raw_title)
+      return nil unless match
+
+      author = HTML.unescape(match[1]).strip
+      author.empty? ? nil : author
+    end
+
+    private def thumbnail_for(config : Hash(String, JSON::Any)?, body : String) : URI?
+      value = config.try { |hash| JsonValue.string(hash["i"]?) }
+
+      unless value
+        if match = THUMBNAIL_PATTERN.match(body)
+          value = json_unescape(match[1])
+        end
+      end
+
+      value ? HttpUrl.absolute(value) : nil
+    end
+
+    private def duration_for(config : Hash(String, JSON::Any)?, body : String) : Float64?
+      if config
+        duration = JsonValue.seconds(config["duration"]?)
+        return duration if duration
+      end
+
+      match = DURATION_PATTERN.match(body)
+      return nil unless match
+
+      match[1].to_f64?
+    end
+
+    private def config_for(body : String, identifier : String) : Hash(String, JSON::Any)?
+      marker = "p.f[#{identifier.to_json}]"
+      marker_index = body.index(marker)
+      return nil unless marker_index
+
+      equals_index = body.index('=', marker_index + marker.size)
+      return nil unless equals_index
+
+      object_start = body.index('{', equals_index)
+      return nil unless object_start
+
+      object_end = matching_brace(body, object_start)
+      return nil unless object_end
+
+      json = body[object_start..object_end].sub(/,\s*"?loaded"?\s*:\s*v\(\)(?=\s*\})/, "")
+      JSON.parse(json).as_h?
+    rescue JSON::ParseException
+      nil
+    end
+
+    private def matching_brace(body : String, start : Int32) : Int32?
+      depth = 0
+      in_string = false
+      escaped = false
+      index = start
+
+      while index < body.size
+        char = body[index]
+
+        if in_string
+          if escaped
+            escaped = false
+          elsif char == '\\'
+            escaped = true
+          elsif char == '"'
+            in_string = false
+          end
+        else
+          case char
+          when '"'
+            in_string = true
+          when '{'
+            depth += 1
+          when '}'
+            depth -= 1
+            return index if depth.zero?
+          end
+        end
+
+        index += 1
+      end
+
+      nil
+    end
+
+    private def formats_from(config : Hash(String, JSON::Any)?) : Array(Format)
+      formats = [] of Format
+      return formats unless config
+
+      seen_urls = Set(String).new
+      seen_ids = Set(String).new
+
+      if ua = config["ua"]?.try(&.as_h?)
+        ua.each do |container, qualities_value|
+          next unless supported_container?(container)
+
+          qualities = qualities_value.as_h?
+          next unless qualities
+
+          qualities.each do |quality, entry_value|
+            entry = entry_value.as_h?
+            next unless entry
+
+            append_variant(formats, seen_urls, seen_ids, entry, container, quality, false)
+          end
+        end
+      end
+
+      if base = config["u"]?.try(&.as_h?)
+        base.each do |container, entry_value|
+          next unless supported_container?(container)
+
+          entry = entry_value.as_h?
+          next unless entry
+
+          append_variant(formats, seen_urls, seen_ids, entry, container, nil, true)
+        end
+      end
+
+      formats
+    end
+
+    private def supported_container?(container : String) : Bool
+      container == "mp4" || container == "webm"
+    end
+
+    private def append_variant(
+      formats : Array(Format),
+      seen_urls : Set(String),
+      seen_ids : Set(String),
+      entry : Hash(String, JSON::Any),
+      container : String,
+      quality : String?,
+      base : Bool,
+    ) : Nil
+      url_text = JsonValue.string(entry["url"]?)
+      return unless url_text
+
+      url = HttpUrl.absolute(url_text)
+      return unless url
+
+      url_key = url.to_s
+      return if seen_urls.includes?(url_key)
+
+      meta = entry["meta"]?.try(&.as_h?) || {} of String => JSON::Any
+      format_id = variant_id(seen_ids, container, quality, meta, base)
+
+      seen_urls << url_key
+      seen_ids << format_id
+      formats << Format.new(
+        format_id: format_id,
+        url: url,
+        kind: FormatKind::Combined,
+        protocol: FormatProtocol::Http,
+        container: container,
+        bitrate: non_negative(JsonValue.integer(meta["bitrate"]?)),
+        width: JsonValue.dimension(meta["w"]?),
+        height: JsonValue.dimension(meta["h"]?),
+        filesize: non_negative(JsonValue.integer(meta["size"]?)),
+        note: base ? "base" : nil,
+      )
+    end
+
+    private def variant_id(
+      seen_ids : Set(String),
+      container : String,
+      quality : String?,
+      meta : Hash(String, JSON::Any),
+      base : Bool,
+    ) : String
+      suffix = if base
+                 "base"
+               else
+                 quality || JsonValue.dimension(meta["h"]?).try(&.to_s) || "video"
+               end
+
+      candidate = "#{container}-#{suffix}"
+      return candidate unless seen_ids.includes?(candidate)
+
+      index = 2
+      while seen_ids.includes?("#{candidate}-#{index}")
+        index += 1
+      end
+
+      "#{candidate}-#{index}"
+    end
+
+    private def non_negative(value : Int64?) : Int64?
+      value if value && value >= 0
+    end
+
+    private def json_unescape(value : String) : String?
+      JSON.parse(%("#{value}")).as_s?
+    rescue JSON::ParseException
+      nil
+    end
+  end
+end