arvados 0.1.20140909151622 → 0.1.20140910123800

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 14c8142d52bb086071d797a71f30740d55e3b113
-  data.tar.gz: b81833ea78e17af27936c1fafa0b4ede6f79e3f5
+  metadata.gz: 76f2025b289c3869d9ab13a42e3242c7eb4815ce
+  data.tar.gz: 6c12428dfab28718ee596af488dd46ff7398246e
 SHA512:
-  metadata.gz: f6bb6fde478daa87acde8018ab55cf25ba809ae9fb313bac9b7bac86d9e2cf37bb09f249cfe65d30e66c4fc5c48c7d087eab88149074d3e1588fdbdab8bd05cd
-  data.tar.gz: 56fff84e629e1a4c1eb5e6d95c486b673740c41204601f4239570719fa4cbd5d8b4207ef7346bfb07aa0fec67862d63749098b079875e2d6f80f5f4b6b9bd866
+  metadata.gz: 613ee75d6a08e13313b3ec83e34d8936a408c2c4b43a31f4d85de1895033010a3eb743b698c5eace067875456d47bf04fd54f23a9a11a363b1fba3ecefd0ff7d
+  data.tar.gz: dcad0bbab01458824d6d0792ffd1e4bc9442a69c6e46c9f6c28f7256e5532d8cf961287c5519b715d150f2d634368691fe52e388c355b94458ec2f2ef96639f7

data/lib/arvados/keep.rb

@@ -94,9 +94,10 @@ module Keep
     # Class to parse a manifest text and provide common views of that data.
     def initialize(manifest_text)
       @text = manifest_text
+      @files = nil
     end
 
-    def each_stream
+    def each_line
       return to_enum(__method__) unless block_given?
       @text.each_line do |line|
         tokens = line.split
@@ -110,17 +111,8 @@ module Keep
       end
     end
 
-    def each_file
-      return to_enum(__method__) unless block_given?
-      each_stream do |streamname, blocklist, filelist|
-        filelist.each do |filespec|
-          start_pos, filesize, filename = filespec.split(':', 3)
-          yield [streamname, filename, filesize.to_i]
-        end
-      end
-    end
-
     def unescape(s)
+      # Parse backslash escapes in a Keep manifest stream or file name.
       s.gsub(/\\(\\|[0-7]{3})/) do |_|
         case $1
         when '\\'
@@ -130,5 +122,68 @@ module Keep
         end
       end
     end
+
+    def each_file_spec(speclist)
+      return to_enum(__method__, speclist) unless block_given?
+      speclist.each do |filespec|
+        start_pos, filesize, filename = filespec.split(':', 3)
+        yield [start_pos.to_i, filesize.to_i, filename]
+      end
+    end
+
+    def files
+      if @files.nil?
+        file_sizes = Hash.new(0)
+        each_line do |streamname, blocklist, filelist|
+          each_file_spec(filelist) do |_, filesize, filename|
+            file_sizes[[streamname, filename]] += filesize
+          end
+        end
+        @files = file_sizes.each_pair.map do |(streamname, filename), size|
+          [streamname, filename, size]
+        end
+      end
+      @files
+    end
+
+    def files_count(stop_after=nil)
+      # Return the number of files represented in this manifest.
+      # If stop_after is provided, files_count will read the manifest
+      # incrementally, and return immediately when it counts that number of
+      # files.  This can help you avoid parsing the entire manifest if you
+      # just want to check if a small number of files are specified.
+      if stop_after.nil? or not @files.nil?
+        return files.size
+      end
+      seen_files = {}
+      each_line do |streamname, blocklist, filelist|
+        each_file_spec(filelist) do |_, _, filename|
+          seen_files[[streamname, filename]] = true
+          return stop_after if (seen_files.size >= stop_after)
+        end
+      end
+      seen_files.size
+    end
+
+    def exact_file_count?(want_count)
+      files_count(want_count + 1) == want_count
+    end
+
+    def minimum_file_count?(want_count)
+      files_count(want_count) >= want_count
+    end
+
+    def has_file?(want_stream, want_file=nil)
+      if want_file.nil?
+        want_stream, want_file = File.split(want_stream)
+      end
+      each_line do |stream_name, _, filelist|
+        if (stream_name == want_stream) and
+            each_file_spec(filelist).any? { |_, _, name| name == want_file }
+          return true
+        end
+      end
+      false
+    end
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: arvados
 version: !ruby/object:Gem::Version
-  version: 0.1.20140909151622
+  version: 0.1.20140910123800
 platform: ruby
 authors:
 - Arvados Authors
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-09-09 00:00:00.000000000 Z
+date: 2014-09-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: google-api-client
@@ -86,7 +86,7 @@ dependencies:
     - - "<"
       - !ruby/object:Gem::Version
         version: 1.0.0
-description: This is the Arvados SDK gem, git revision b0440160e64caecbff160a26f62301dd15d84c7f
+description: This is the Arvados SDK gem, git revision bafb417c531d6094697a13a7465313d1cc7c0bc9
 email: gem-dev@curoverse.com
 executables: []
 extensions: []