Revise documentation and download strategy.

Signed-off-by: Anurag Priyam <[email protected]>

lib/ncbi-blast-dbs.rake

@@ -4,15 +4,16 @@ require 'net/ftp'
 # local copy is older than at the given URL, or if the local copy is corrupt.
 def download(url)
   file = File.basename(url)
-  # Download tarball if the local copy is older than at the given URL or fetch
-  # it for the first time.
+  # Resume an interrupted download or fetch the file for the first time. If
+  # the file on the server is newer, then it is downloaded from start.
+  sh "wget -Nc #{url}"
+  # If the local copy is already fully retrieved, then the previous command
+  # ignores the timestamp. So we check with the server again if the file on
+  # the server is newer and if so download the new copy.
   sh "wget -N #{url}"
-  # Resume aborted download. Do nothing if the file is already fully retrieved
-  # (at the cost is a round trip to server).
-  sh "wget -c #{url}"
 
-  # Always download md5 and verify the tarball. Re-download tarball if corrupt;
-  # extract otherwise.
+  # Immediately download md5 and verify the tarball. Re-download tarball if
+  # corrupt; extract otherwise.
   sh "wget #{url}.md5 && md5sum -c #{file}.md5" do |matched, _|
     if !matched
       sh "rm #{file} #{file}.md5"; download(url)
@@ -22,6 +23,11 @@ def download(url)
   end
 end
 
+# Connects to NCBI's FTP server, gets the URL of all database volumes and
+# returns them grouped by database name:
+#
+#     {'nr' => ['ftp://...', ...], 'nt' => [...], ...}
+#
 def databases
   host, dir = 'ftp.ncbi.nlm.nih.gov', 'blast/db'
   usr, pswd = 'anonymous', ENV['email']
@@ -35,10 +41,14 @@ def databases
   end
 end
 
+# Create user-facing task for each database to drive the download of its
+# volumes in parallel.
 databases.each do |name, files|
   multitask(name => files.map { |file| task(file) { download(file) } })
 end
 
+# List name of all databases that can be downloaded if executed without
+# any arguments.
 task :default do
   puts databases.keys.join(', ')
 end