pluto 0.7.0 → 0.8.0

data/lib/pluto/updater.rb

@@ -22,6 +22,9 @@ class Updater
   end
 
   def fetch_feed( url )
+    
+    ### fix: use worker.get( url )  # check http response code etc.
+    
     xml = worker.read( url )
 
     ###
@@ -42,60 +45,77 @@ class Updater
   end
 
 
-  def parse_feed( xml )
-    parser = RSS::Parser.new( xml )
-    parser.do_validate            = false
-    parser.ignore_unknown_element = true
-
-    puts "Parsing feed..."
-    feed = parser.parse
-      
-    puts "  feed.class=#{feed.class.name}"
-    feed
-  end
-
-
   def update_subscriptions( config, opts={} )
 
     ## for now - use single site w/ key planet  -- fix!! allow multiple sites (planets)
     
+    site_attribs = {
+      title: config[ 'title' ]
+    }
+    
     site_key = 'planet'
     site_rec = Site.find_by_key( site_key )
     if site_rec.nil?
-      site_rec      = Site.new
-      site_rec.key  = site_key
+      site_rec             = Site.new
+      site_attribs[ :key ] = site_key
+
+      ## use object_id: site.id and object_type: Site
+      ## change - model/table/schema!!!
+      Action.create!( title: 'new site', object: site_attribs[ :title ] )
     end
-    site_rec.title  = config[ 'title' ]
-    site_rec.save!
+    site_rec.update_attributes!( site_attribs )
+
+    # -- log update action
+    Action.create!( title: 'update subscriptions' )
 
 
     config[ 'feeds' ].each do |feed_key|
       
       feed_hash  = config[ feed_key ]
-      feed_url   = feed_hash[ 'feed_url' ]
-      
-      puts "Updating feed subscription >#{feed_key}< - >#{feed_url}<..."
+
+      feed_attribs = {
+        feed_url: feed_hash[ 'feed_url' ],
+        url:      feed_hash[ 'url'      ],
+        title:    feed_hash[ 'title'    ]   # todo: use title from feed?
+      }
+
+      puts "Updating feed subscription >#{feed_key}< - >#{feed_attribs[:feed_url]}<..."
 
       feed_rec = Feed.find_by_key( feed_key )
       if feed_rec.nil?
-        feed_rec      = Feed.new
-        feed_rec.key  = feed_key
+        feed_rec             = Feed.new
+        feed_attribs[ :key ] = feed_key
+
+        ## use object_id: feed.id and object_type: Feed
+        ## change - model/table/schema!!!
+        ## todo: add parent_action_id - why? why not?
+        Action.create!( title: 'new feed', object: feed_attribs[ :title ] )
       end
-      feed_rec.feed_url = feed_url
-      feed_rec.url      = feed_hash[ 'url' ]
-      feed_rec.title    = feed_hash[ 'title' ]    # todo: use title from feed?
-      feed_rec.save!
+      
+      feed_rec.update_attributes!( feed_attribs )
       
       ## todo:
       #  add subscription records  (feed,site)  - how? 
     end
 
   end # method update_subscriptions
-  
-  
+
+
   def update_feeds( opts={} )
 
-    logger.debug "using stdlib RSS::VERSION #{RSS::VERSION}"
+    ### move to feedutils
+    ### logger.debug "using stdlib RSS::VERSION #{RSS::VERSION}"
+
+    # -- log update action
+    Action.create!( title: 'update feeds' )
+
+    #####
+    # -- update fetched_at  timestamps for all sites
+    feeds_fetched_at = Time.now
+    Site.all.each do |site|
+      site.fetched_at = feeds_fetched_at
+      site.save!
+    end
 
     Feed.all.each do |feed_rec|
 
@@ -117,108 +137,64 @@ class Updater
       
       puts "Before parsing feed >#{feed_key}<..."
 
-      feed = parse_feed( feed_xml )
+      ## fix/todo: check for feed.nil?   -> error parsing!!!
+      #    or throw exception
+      feed = FeedUtils::Parser.parse( feed_xml )
 
-      if feed.class == RSS::Atom::Feed
-        puts "== #{feed.title.content} =="
-      else  ## assume RSS::Rss::Feed
-        puts "==  #{feed.channel.title} =="
-      end
+      feed_fetched_at = Time.now
 
-      feed.items.each do |item|
-        if feed.class == RSS::Atom::Feed
-          item_attribs = handle_feed_item_atom( item )
-        else  ## assume RSS::Rss::Feed
-          item_attribs = handle_feed_item_rss( item )
-        end
+      ## todo/check: move feed_rec update to the end (after item updates??)
 
-        # add feed_id fk_ref
-        item_attribs[ :feed_id ] = feed_rec.id
+      # update feed attribs e.g.
+      #    generator
+      #    published_at,built_at,touched_at,fetched_at
+      #    summary,title2
+      feed_attribs = {
+        fetched_at:   feed_fetched_at,
+        format:       feed.format,
+        published_at: feed.published? ? feed.published : nil,
+        touched_at:   feed.updated?   ? feed.updated   : nil,
+        built_at:     feed.built?     ? feed.built     : nil,
+        summary:      feed.summary?   ? feed.summary   : nil,
+        title2:       feed.title2?    ? feed.title2    : nil,
+        generator:    feed.generator
+      }
 
-        rec = Item.find_by_guid( item_attribs[ :guid ] )
-        if rec.nil?
-          rec      = Item.new
-          puts "** NEW"
-        else
-          puts "UPDATE"
-        end
-                
-        rec.update_attributes!( item_attribs )
-      end  # each item
+      feed_rec.update_attributes!( feed_attribs )
 
-    end # each feed
-
-  end # method run
 
+      feed.items.each do |item|
 
-  def handle_feed_item_atom( item )
-
-        ## todo: if content.content empty use summary for example
         item_attribs = {
-          title:        item.title.content,
-          url:          item.link.href,
-          published_at: item.updated.content.utc.strftime( "%Y-%m-%d %H:%M" ),
-          # content:   item.content.content,
+          fetched_at:   feed_fetched_at,
+          title:        item.title,
+          url:          item.url,
+          summary:      item.summary?   ? item.summary   : nil,
+          content:      item.content?   ? item.content   : nil,
+          published_at: item.published? ? item.published : nil,
+          touched_at:   item.updated?   ? item.updated   : nil,
+          feed_id:      feed_rec.id    # add feed_id fk_ref
         }
 
-        item_attribs[ :guid ] = item.id.content
-
-        if item.summary
-          item_attribs[ :content ] = item.summary.content
+        rec = Item.find_by_guid( item.guid )
+        if rec.nil?
+          rec      = Item.new
+          item_attribs[ :guid ] = item.guid
+          puts "** NEW | #{item.title}"
         else
-          if item.content
-            text  = item.content.content.dup
-            ## strip all html tags
-            text = text.gsub( /<[^>]+>/, '' )
-            text = text[ 0..400 ] # get first 400 chars
-            ## todo: check for length if > 400 add ... at the end???
-            item_attribs[ :content ] = text
-          end
+          ## todo: check if any attribs changed
+          puts "UPDATE | #{item.title}"
         end
 
-        puts "- #{item.title.content}"
-        puts "  link >#{item.link.href}<"
-        puts "  id (~guid) >#{item.id.content}<"
-        
-        ### todo: use/try published first? why? why not?
-        puts "  updated (~pubDate) >#{item.updated.content}< >#{item.updated.content.utc.strftime( "%Y-%m-%d %H:%M" )}< : #{item.updated.content.class.name}"
-        puts
-        
-        # puts "*** dump item:"
-        # pp item
-
-        item_attribs
-  end
+        rec.update_attributes!( item_attribs )
+      end  # each item
 
-  def handle_feed_item_rss( item )
+    end # each feed
 
-       item_attribs = {
-          title:        item.title,
-          url:          item.link,
-          published_at: item.pubDate.utc.strftime( "%Y-%m-%d %H:%M" ),
-          # content:  item.content_encoded,
-        }
-        
-        # if item.content_encoded.nil?
-          # puts " using description for content"
-          item_attribs[ :content ] = item.description
-        # end
-        
-        item_attribs[ :guid ] = item.guid.content
-        
-        puts "- #{item.title}"
-        puts "  link (#{item.link})"
-        puts "  guid (#{item.guid.content})"
-        puts "  pubDate >#{item.pubDate}< >#{item.pubDate.utc.strftime( "%Y-%m-%d %H:%M" )}< : #{item.pubDate.class.name}"
-        puts
-
-        # puts "*** dump item:"
-        # pp item
-
-        item_attribs
-  end
+  end # method run
 
 
+ 
 end # class Fetcher
 
 end # module Pluto

data/lib/pluto/version.rb

@@ -1,5 +1,5 @@
 
 module Pluto
-  VERSION = '0.7.0'
+  VERSION = '0.8.0'
 end
 

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: pluto
 version: !ruby/object:Gem::Version
-  version: 0.7.0
+  version: 0.8.0
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-09-12 00:00:00.000000000 Z
+date: 2013-09-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: pakman
-  requirement: &73611800 !ruby/object:Gem::Requirement
+  requirement: &82365260 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,10 +21,10 @@ dependencies:
         version: '0.5'
   type: :runtime
   prerelease: false
-  version_requirements: *73611800
+  version_requirements: *82365260
 - !ruby/object:Gem::Dependency
   name: fetcher
-  requirement: &73611130 !ruby/object:Gem::Requirement
+  requirement: &82364940 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -32,10 +32,10 @@ dependencies:
         version: '0.3'
   type: :runtime
   prerelease: false
-  version_requirements: *73611130
+  version_requirements: *82364940
 - !ruby/object:Gem::Dependency
   name: logutils
-  requirement: &73610460 !ruby/object:Gem::Requirement
+  requirement: &82364710 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -43,10 +43,54 @@ dependencies:
         version: '0.6'
   type: :runtime
   prerelease: false
-  version_requirements: *73610460
+  version_requirements: *82364710
+- !ruby/object:Gem::Dependency
+  name: feedutils
+  requirement: &82364390 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0.3'
+  type: :runtime
+  prerelease: false
+  version_requirements: *82364390
+- !ruby/object:Gem::Dependency
+  name: props
+  requirement: &82364120 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 1.0.0
+  type: :runtime
+  prerelease: false
+  version_requirements: *82364120
+- !ruby/object:Gem::Dependency
+  name: textutils
+  requirement: &82363830 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 0.6.8
+  type: :runtime
+  prerelease: false
+  version_requirements: *82363830
+- !ruby/object:Gem::Dependency
+  name: gli
+  requirement: &82363570 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: 2.5.6
+  type: :runtime
+  prerelease: false
+  version_requirements: *82363570
 - !ruby/object:Gem::Dependency
   name: rdoc
-  requirement: &73609750 !ruby/object:Gem::Requirement
+  requirement: &82363260 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -54,10 +98,10 @@ dependencies:
         version: '3.10'
   type: :development
   prerelease: false
-  version_requirements: *73609750
+  version_requirements: *82363260
 - !ruby/object:Gem::Dependency
   name: hoe
-  requirement: &73609370 !ruby/object:Gem::Requirement
+  requirement: &82362910 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -65,7 +109,7 @@ dependencies:
         version: '3.3'
   type: :development
   prerelease: false
-  version_requirements: *73609370
+  version_requirements: *82362910
 description: pluto - Another Planet Generator (Lets You Build Web Pages from Published
   Web Feeds)
 email: webslideshow@googlegroups.com
@@ -74,45 +118,33 @@ executables:
 extensions: []
 extra_rdoc_files:
 - Manifest.txt
-- templates/blank.txt
 files:
-- History.markdown
+- History.md
 - Manifest.txt
-- README.markdown
+- README.md
 - Rakefile
 - bin/pluto
+- config/pluto.index.yml
 - lib/pluto.rb
+- lib/pluto/cli/main.rb
 - lib/pluto/cli/opts.rb
-- lib/pluto/cli/runner.rb
 - lib/pluto/connecter.rb
 - lib/pluto/fetcher.rb
 - lib/pluto/formatter.rb
+- lib/pluto/installer.rb
+- lib/pluto/lister.rb
+- lib/pluto/manifest_helpers.rb
 - lib/pluto/models.rb
 - lib/pluto/schema.rb
-- lib/pluto/server.rb
-- lib/pluto/server/public/feed-icon-10x10.png
-- lib/pluto/server/public/style.css
-- lib/pluto/server/views/_debug.erb
-- lib/pluto/server/views/_version.erb
-- lib/pluto/server/views/debug.erb
-- lib/pluto/server/views/feeds.erb
-- lib/pluto/server/views/index.erb
-- lib/pluto/server/views/items.erb
-- lib/pluto/server/views/layout.erb
-- lib/pluto/server/views/sites.erb
 - lib/pluto/updater.rb
 - lib/pluto/version.rb
-- templates/blank.html.erb
-- templates/blank.top.html.erb
-- templates/blank.txt
-- templates/feed-icon-10x10.png
-homepage: https://github.com/geraldb/pluto
+homepage: https://github.com/feedreader/pluto
 licenses:
 - Public Domain
 post_install_message: 
 rdoc_options:
 - --main
-- README.markdown
+- README.md
 require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement