bmg 0.18.14 → 0.18.15

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 8541ab1e17726eb751772cc29a3d198f8bc021d7
-  data.tar.gz: a4e07beb7b994d3ac6dda214e3a75822a1a3eb07
+  metadata.gz: 14260453d9c3f5d24b5ef6975072e452af309485
+  data.tar.gz: f6de19178d2b4404c9fbe248d42f9c596cee410e
 SHA512:
-  metadata.gz: '063179c6a79505b6baaaf800f56a35dc7084b95ac4fa2c73acc92e809bb1d164d99bb3ce6120863d3b87886f8e21b3bc3579a1043e01dc2938ff6c278a4c08c3'
-  data.tar.gz: 3fe57b85503565d547522f4a41ad608ca3839add2305e5108b212f269a4d76adb4d681da58cca01bea564c1ce24b70922215923f3df332d5a7d51dadc09a2ee8
+  metadata.gz: e53da5df4abb7095036dc275948a70e4af9f1945e2fd446244e765d7fead1258a9e564d9076baa3efc4ee29566810a5176c76d2c3c953de906a7ff36a3a86dff
+  data.tar.gz: a6f0cb03795e5b954563f774bc65c7879a78049928d06da935017a8dfcdb2bdd563c4ab08922cb95f774d68714ab108bb480d0722249ec5e6cc67fed5ed0064a

data/lib/bmg/reader/csv.rb

@@ -10,20 +10,18 @@ module Bmg
       }
 
       def initialize(type, path_or_io, options = {})
-        @type = type
+        require 'csv'
+
         @path_or_io = path_or_io
-        @options = DEFAULT_OPTIONS.merge(options)
-        if @options[:smart] && !@path_or_io.is_a?(IO)
-          @options[:col_sep] ||= infer_col_sep
-          @options[:quote_char] ||= infer_quote_char
-        end
+        @options = handle_options(options)
+        @type = handle_type(type)
       end
 
       def each
         return to_enum unless block_given?
-        require 'csv'
-        with_io do |io|
-          ::CSV.new(io, **csv_options).each do |row|
+
+        with_csv do |csv|
+          csv.each do |row|
             yield tuple(row)
           end
         end
@@ -44,33 +42,70 @@ module Bmg
         row.to_hash.each_with_object({}){|(k,v),h| h[k.to_sym] = v }
       end
 
+      def handle_type(type)
+        return type if type.knows_attrlist?
+
+        type.with_attrlist(infer_attrlist)
+      end
+
+      def infer_attrlist
+        with_csv do |csv|
+          csv.each do |row|
+            return tuple(row).keys
+          end
+        end
+      end
+
+      def handle_options(options)
+        options = DEFAULT_OPTIONS.merge(options)
+        if options[:smart] && !@path_or_io.is_a?(IO)
+          options[:col_sep] ||= infer_col_sep
+          options[:quote_char] ||= infer_quote_char(options[:col_sep])
+        end
+        options
+      end
+
       def infer_col_sep
         sniff(text_portion, [",","\t",";"], ",")
       end
 
-      def infer_quote_char
-        sniff(text_portion, ["'","\""], "\"")
+      def infer_quote_char(col_sep)
+        sniff(text_portion, ['"',"'"], '"'){|quote|
+          /#{quote}#{col_sep}#{quote}|^#{quote}|#{quote}$/
+        }
       end
 
       def text_portion
-        @text_portion ||= with_io{|io| io.readlines(10).join("\n") }
+        @text_portion ||= with_io{|io| io.readlines(50).join("\n") }
       end
 
       def with_io(&bl)
         case @path_or_io
         when IO, StringIO
+          @path_or_io.rewind if @path_or_io.respond_to?(:rewind)
           bl.call(@path_or_io)
         else
           File.open(@path_or_io, "r", &bl)
         end
       end
 
+      def with_csv(&bl)
+        with_io do |io|
+          yield ::CSV.new(io, **csv_options)
+        end
+      end
+
       # Finds the best candidate among `candidates` for a separator
       # found in `str`. If none is found, returns `default`.
-      def sniff(str, candidates, default)
+      def sniff(str, candidates, default, &bl)
         snif = {}
         candidates.each {|delim|
-          snif[delim] = str.count(delim)
+          counter = bl ? bl.call(delim) : delim
+          snif[delim] = if counter.is_a?(Regexp)
+            str.scan(counter).length
+          else
+            str.count(counter)
+          end
         }
         snif = snif.sort {|a,b| b[1] <=> a[1] }
         snif.size > 0 ? snif[0][0] : default

data/lib/bmg/version.rb

@@ -2,7 +2,7 @@ module Bmg
   module Version
     MAJOR = 0
     MINOR = 18
-    TINY  = 14
+    TINY  = 15
   end
   VERSION = "#{Version::MAJOR}.#{Version::MINOR}.#{Version::TINY}"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: bmg
 version: !ruby/object:Gem::Version
-  version: 0.18.14
+  version: 0.18.15
 platform: ruby
 authors:
 - Bernard Lambeau