Rewrite most of Bliptv to work with their new scheme.

[dead/whatever-dl.git] / src / websites / bliptv.rb

diff --git a/src/websites/bliptv.rb b/src/websites/bliptv.rb
index 9322356a9e92353a31782e3a89c27500946b3939..33205877c61152a2974412a27b0e2a5356f023c5 100644 (file)
--- a/src/websites/bliptv.rb
+++ b/src/websites/bliptv.rb
@@ -17,57 +17,109 @@
 #
 
 require 'src/website'
+require 'cgi'
+
+class BliptvMediaFormat
+  # This is just a convenience class for parsing two parameters out of
+  # an RSS feed: 'url' and 'blip:role'.
+  def initialize(line)
+    @url = nil
+    @role = nil
+
+    url_regex  = /url=\"([^\"]+)\"/
+    role_regex = /blip:role=\"([^\"]+)\"/
+    url_matches  = url_regex.match(line)
+    role_matches = role_regex.match(line)
+
+    if not url_matches.nil? and (url_matches.length) > 1
+      @url = url_matches[1]
+    end
+
+    if not role_matches.nil? and (role_matches.length > 1)
+      @role = role_matches[1]
+    end
+  end
+
+  def role
+    return @role
+  end
+
+  def url
+    return @url
+  end
+end
+
 
 class Bliptv < Website
 
-  VALID_BLIPTV_URL_REGEX = /^(http:\/\/)?([[:alnum:]\-]+\.)?blip\.tv\/file\/(\d+)(.*)?$/
+  VALID_BLIPTV_URL_REGEX = /^(http:\/\/)?blip\.tv\/.*?(\d+)$/
 
   def self.owns_url?(url)
     return url =~ VALID_BLIPTV_URL_REGEX
   end
 
- 
+
   def get_video_url()
-    page_data = self.get_page_data(@url)    
-    filepath = parse_video_url(page_data)
-    
-    return filepath
+    video_id = self.parse_video_id()
+    rss_page_url = "http://blip.tv/rss/flash/#{video_id}"
+    rss_data = get_page_data(rss_page_url)
+    video_url = parse_video_url(rss_data)
+
+    return video_url
   end
 
-  
+
   protected;
 
-  
-  def parse_video_url(page_data)
-    # First, try to find the MOV video. The source videos are usually
-    # encoded with MOV.
-    video_url_regex = /"Quicktime \(\.mov\)", "attribute" : "(.*?\.mov)/i
-    matches = video_url_regex.match(page_data)
+  def parse_video_id()
+    video_id_regex = /(\d+)$/
+    matches = video_id_regex.match(@url)
 
-    if not matches.nil?
+    if matches.nil? or (matches.length < 2)
+      raise StandardError.new("Couldn't parse the video id from the URL.")
+    else
       return matches[1]
     end
+  end
 
-    # If that didn't work, try the WMV format, which is occasionally
-    # used for the source as well.
-    video_url_regex = /"Windows Media \(\.wmv\)", "attribute" : "(.*?\.wmv)/i
-    matches = video_url_regex.match(page_data)
 
-    if not matches.nil?
-      return matches[1]
+  def choose_best_format(formats)
+    # 'formats' is assumed to be an array of BliptvMediaFormat. We
+    # return the best one (in terms of video quality).
+    formats.each do |f|
+      if f.url.nil? or f.role.nil?
+        formats.delete(f)
+        next
+      end
+
+      return f if f.role == "Source"
+    end
+
+    if formats.length == 0
+      raise StandardError.new("No valid formats in the RSS feed.")
+    else
+      # Return whatever's left if we don't have a 'Source' video.
+      return formats[0]
     end
+  end
 
-    
-    # If neither of the source formats are present, just grab the
-    # video URL from the Flash variable and be done with it.
-    video_url_regex = /setPrimaryMediaUrl\("(.*?\.(flv|mov|wmv|mp4))/i
-    matches = video_url_regex.match(page_data)
+  def parse_video_url(page_data)
+    # All of the elements containing video URLs begin like this.
+    media_regex = /^\s*<media:content/
 
-    if matches.nil?
-      raise StandardError.new("Couldn't parse any of the video format URLs.")
+    formats = []
+    # Create an array of BliptvMediaFormat from lines matching
+    # media_regex.
+    page_data.lines.each do |line|
+      if (line =~ media_regex)
+        bp = BliptvMediaFormat.new(line)
+        formats << bp
+      end
     end
 
-    return matches[1]
+    # And return the URL from the best one.
+    # choose_best_format will raise an error if need be.
+    return choose_best_format(formats).url
   end
 
 end