Changeset 103

Timestamp:

09/24/08 18:24:27 (2 months ago)

Author:

deveiant

Message:

Made the db converter look in more paths for the WordNet? dict files.

Location:

trunk

Files:

• Rakefile.local (modified) (1 diff)
• convertdb.rb (modified) (6 diffs)
• utils.rb (modified) (3 diffs)

trunk/Rakefile.local

@@ -61 +61 @@
     end
 end
+

trunk/convertdb.rb

@@ -44 +44 @@
 require 'optparse'
 require 'fileutils'
+require 'uri'
+require 'net/http'
 
 
@@ -79 +81 @@
                Pathname.new( WordNet::Lexicon::DEFAULT_DB_ENV ).basename
 
+    # Paths to search for WordNet dictionary files
+    WNDB_LOCAL_PATHS = [
+        '/usr/local/WordNet-3.0/dict',      # Default install
+        '/usr/WordNet-3.0/dict',            # Default with --prefix=/usr
+        '/usr/local/share/WordNet',         # FreeBSD
+        './dict',                           # Extracted locally
+    ]
+
+    # URL to the latest database file archive
+    WNDB_TARBALL_URL = URI.parse( 'http://wordnet.princeton.edu/3.0/WordNet-3.0.tar.gz' )
+
 
     ### Create a new converter that will dump WordNet dictionary files into a BerkeleyDB 
@@ -98 +111 @@
         exit unless /^y/i =~ prompt_with_default("Continue?", "y")
 
-        # Open the database and check to be sure it's empty. Confirm overwrite if
-        # not. Checkpoint and set up logging proc if debugging.
+        # Confirm if we're going to clobber an existing database directory
         if @builddir.exist? && ( @builddir + 'data' ).exist?
             message ">>> Warning: Existing data in the Ruby-WordNet databases\n"\
@@ -109 +121 @@
 
         # Find the source data files
-        default = nil
-        wndirs = Pathname.glob( Pathname.getwd + 'WordNet-*' )
-        localdict = Pathname.getwd + 'dict'
-        if !wndirs.empty?
-            default = wndirs.first + 'dict'
-        elsif localdict.exist?
-            default = localdict
-        else
-            default = '/usr/local/WordNet-3.0/dict'
-        end
-
-        message "Where can I find the WordNet data files?\n"
-        datadir = prompt_with_default( "Data directory", default )
-        datadir = Pathname.new( datadir )
-
-        abort( "Directory '#{datadir}' does not exist" ) unless datadir.exist?
-        abort( "'#{datadir}' is not a directory" ) unless datadir.directory?
+        datadir = find_data_files()
         testfile = datadir + "data.noun"
-        abort( "'#{datadir}' doesn't seem to contain the necessary files.") unless testfile.exist?
+        abort( "'#{datadir}' doesn't seem to contain the necessary files." ) unless testfile.exist?
 
         # Open the lexicon readwrite into the temporary datadir
@@ -201 +197 @@
     end
 
-
-    #######
-    private
-    #######
 
     # Index entry patterns
@@ -375 +367 @@
     end
 
+
+    ### Find the path to the WordNet dict directory with the files we're 
+    ### going to parse in it.
+    def find_data_files
+        datadir = WNDB_LOCAL_PATHS.collect {|pn| Pathname.new(pn) }.find {|pn| pn.exist? }
+        
+        unless datadir
+            message "Where can I find the WordNet data files?\n"
+            answer = prompt_with_default( "Data directory", WNDB_LOCAL_PATHS.first )
+            datadir = Pathname.new( answer )
+        end
+
+        abort( "Directory '#{datadir}' does not exist" ) unless datadir.exist?
+        abort( "'#{datadir}' is not a directory" ) unless datadir.directory?
+
+        return datadir
+    end
+    
 end # class WordNetConverter
 

trunk/utils.rb

@@ -294 +294 @@
     ### Display a description of a potentially-dangerous task, and prompt
     ### for confirmation. If the user answers with anything that begins
-    ### with 'y', yield to the block, else raise with an error.
-    def ask_for_confirmation( description )
+    ### with 'y', yield to the block. If +abort_on_decline+ is +true+,
+    ### any non-'y' answer will fail with an error message.
+    def ask_for_confirmation( description, abort_on_decline=true )
         puts description
 
@@ -302 +303 @@
         end
 
-        case answer
-        when /^y/i
-            yield
-        else
+        if answer =~ /^y/i
+            return yield
+        elsif abort_on_decline
             error "Aborted."
             fail
         end
+
+        return false
     end
 
@@ -810 +812 @@
         end
     end
+    
+    
+    ### Download the file at +sourceuri+ via HTTP and write it to +targetfile+.
+    def download( sourceuri, targetfile=nil )
+        oldsync = $defout.sync
+        $defout.sync = true
+        require 'net/http'
+        require 'uri'
+
+        targetpath = Pathname.new( targetfile )
+
+        log "Downloading %s to %s" % [sourceuri, targetfile]
+        targetpath.open( File::WRONLY|File::TRUNC|File::CREAT, 0644 ) do |ofh|
+
+            url = sourceuri.is_a?( URI ) ? sourceuri : URI.parse( sourceuri )
+            downloaded = false
+            limit = 5
+
+            until downloaded or limit.zero?
+                Net::HTTP.start( url.host, url.port ) do |http|
+                    req = Net::HTTP::Get.new( url.path )
+
+                    http.request( req ) do |res|
+                        if res.is_a?( Net::HTTPSuccess )
+                            log "Downloading..."
+                            res.read_body do |buf|
+                                ofh.print( buf )
+                            end
+                            downloaded = true
+                            puts "done."
+
+                        elsif res.is_a?( Net::HTTPRedirection )
+                            url = URI.parse( res['location'] )
+                            log "...following redirection to: %s" % [ url ]
+                            limit -= 1
+                            sleep 0.2
+                            next
+
+                        else
+                            res.error!
+                        end
+                    end
+                end
+            end
+
+        end
+
+        return targetpath
+    ensure
+        $defout.sync = oldsync
+    end
 
 end # module UtilityFunctions