revs-utils 1.0.14 → 1.0.15

checksums.yaml

@@ -1,15 +1,15 @@
 ---
 !binary "U0hBMQ==":
   metadata.gz: !binary |-
-    NWQyNzQ0ZDdhYmMzZmY3OTRiOGQ4MDAwYjJmODhmOTlkNDYwOGI0Zg==
+    MDA0MDNjYzdmYThhMTQzNDNkYzdkYjY4YWE4ZGI5MDIyY2JkZGI1MQ==
   data.tar.gz: !binary |-
-    ZDNjMjY4NjExMzBlODMzZDc5MzgzOWE4YWRlNzBmODhiMWRlZmQ1Yw==
+    OTYwODZjNTIzN2IzZDdmMTA3Mjg3MTA0N2FjNGE0MjNiZjMwN2MwNQ==
 SHA512:
   metadata.gz: !binary |-
-    NDRhYjMwYjFlYWMzNWZjMGZkYzBjMzg5ZWU2NmY3NjI3YzNmNDI2ZjhlNTdl
-    MTE5MmEzYzVlZTliMzQ1NTA2MDU4MTkyNTAyZjU5NmNmNTZmZDkwZjk5MTYx
-    NWRkYTRmYmZiNmQyZTZkZDU1Y2UwMjczMDBiNDYxOTc2MjMxYzY=
+    ZTA5MDU0OTgzNWIwNjdjZWI0M2RmOTg1OGM5MDk4Y2I3ZmU5YWE5YjVhYWQ2
+    NTQ1NWExMmJiNmU5MGNhMzUzYTdhMzk4ZWYwNzkyNTNlYzljNTVmOTBmN2Vl
+    OTdiMTg2MjQ3YWUxOGQ3YWMxYzk5MmVlNTA1YjkwM2U2ZjRjNzA=
   data.tar.gz: !binary |-
-    NWFhYjRkNTYwZDI3Y2JjMDQ2YTVkMDI2NzEwZTFjYTM0ZTc3MTQzOGYxNWMw
-    ODZjZDIxNzU4ODFlYzBmOTkyZjIwNDA0Y2UwNzg0YThjMGEyNjc3MmMzZWVm
-    YjBiYmUzOWI4OTc5YmE5N2ZhMzAzMzAwMDc5YzhlNWM5Y2U4YWE=
+    NTBmNTczNzZhZDI2YzI5OTg0YjMyOWQyOGMyMDJlYjM1ZDBmZGUxYzEzNGI1
+    YjllMzEyMzkxZWJhNzk3MGNkZjRhNjcyZDljOGFlMDhkNDJkZDY3NjBhYTc0
+    Nzg0YTJkYWU1MTlmNjViMGU2ODcyNTk1MGM5ZDVkZGQ2YmU2MWQ=

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    revs-utils (1.0.13)
+    revs-utils (1.0.15)
       actionpack (~> 3)
       chronic
       countries (= 0.9.2)
@@ -42,14 +42,14 @@ GEM
       rake (>= 0.8.7)
       rest-client
     mime-types (2.0)
-    multi_json (1.10.1)
+    multi_json (1.11.0)
     rack (1.4.5)
     rack-cache (1.2)
       rack (>= 0.4)
     rack-test (0.6.3)
       rack (>= 1.0)
     rake (10.1.0)
-    rdf (1.1.7)
+    rdf (1.1.11)
       link_header (~> 0.0, >= 0.0.8)
     rest-client (1.6.7)
       mime-types (>= 1.16)

data/config/manifest_headers.yml

@@ -30,5 +30,15 @@ metadata:
   hide: hide
   format: format
   collection_name: collection_name
-  
-  
+known_formats: 
+  - black-and-white film
+  - color film
+  - slides
+  - foldouts
+  - pamphlets
+  - oversize items
+  - photographic prints
+  - black-and-white negatives
+  - color negatives
+  - black-and-white transparencies
+  - color transparencies

data/lib/revs-utils/version.rb

@@ -1,5 +1,5 @@
 module Revs
   module Utils
-    VERSION = "1.0.14"
+    VERSION = "1.0.15"
   end
 end

data/lib/revs-utils.rb

@@ -14,7 +14,7 @@ REVS_LC_TERMS_FILENAME=File.join(PROJECT_ROOT,'files','revs-lc-marque-terms.obj'
 REVS_MANIFEST_HEADERS_FILEPATH = File.join(PROJECT_ROOT,'config',"manifest_headers.yml")
 REGISTER = "register"
 METADATA = "metadata"
-
+FORMATS = "known_formats"
 
 module Revs
   module Utils
@@ -28,6 +28,9 @@ module Revs
       REVS_MANIFEST_HEADERS_FILE = File.open(REVS_MANIFEST_HEADERS_FILEPATH)
       REVS_MANIFEST_HEADERS = YAML.load( REVS_MANIFEST_HEADERS_FILE)
       
+      def revs_known_formats
+        get_manifest_section(FORMATS)
+      end
       
       def get_manifest_section(section)
         return REVS_MANIFEST_HEADERS[section]
@@ -67,40 +70,59 @@ module Revs
         sources = Array.new
         files.each do |file|
           file.each do |row|
-            #Make sure the sourcid and filename are the same
+            #Make sure the sourceid and filename are the same
             fname = row[get_manifest_section(REGISTER)['filename']].chomp(File.extname(row[get_manifest_section(REGISTER)['filename']]))
             return false if row[get_manifest_section(REGISTER)['sourceid']] != fname
             sources << row[get_manifest_section(REGISTER)['sourceid']]
-          end
-          
-          
-         
+          end         
         end
         return sources.uniq.size == sources.size
       
       end
-      
-      
+            
       #Pass this function a CSV file and it will return true if the proper headers are there and each entry has the required fields filled in
       def valid_to_register(file_path)
-        
         file = read_csv_with_headers(file_path)
+        return check_valid_to_register(file)
+      end
+      
+      #Pass this function a CSV file and it will return true if the proper headers are there and each entry has the required fields filled in.  
+      def valid_for_metadata(file_path)
+        file = read_csv_with_headers(file_path)
+        return check_headers(file)
+      end
+      
+      # pass in csv data and it will tell if you everything is safe to register based on having labels, unique sourceIDs and filenames matching sourceIDs
+      def check_valid_to_register(csv_data)
         #Make sure all the required headers are there
-        return false if not get_manifest_section(REGISTER).values-file[0].keys == []
-        
+        return false if not get_manifest_section(REGISTER).values-csv_data[0].keys == []
+        sources=Array.new
         #Make sure all files have entries for those required headers
-        file.each do |row|
+        csv_data.each do |row|
           get_manifest_section(REGISTER).keys.each do |header| # label should be there as a column but does not always need a value
             return false if header.downcase !='label' && row[header].blank? #Alternatively consider row[header].class != String or row[header].size <= 0
           end
+          fname = row[get_manifest_section(REGISTER)['filename']].chomp(File.extname(row[get_manifest_section(REGISTER)['filename']]))
+          return false if row[get_manifest_section(REGISTER)['sourceid']] != fname
+          sources << row[get_manifest_section(REGISTER)['sourceid']]
         end
-       return true
+        return sources.uniq.size == sources.size
       end
       
-      #Pass this function a CSV file and it will return true if the proper headers are there and each entry has the required fields filled in.  
-      def valid_for_metadata(file_path)
-        file = read_csv_with_headers(file_path)
-        file_headers=file[0].keys.reject(&:blank?).collect(&:downcase)
+      # looks at certain metadata fields in manifest to confirm validity (such as dates and formats)
+      def check_metadata(csv_data)
+        bad_rows=0
+        csv_data.each do |row|
+          valid_date=revs_is_valid_datestring?(row[get_manifest_section(METADATA)['year']] || row[get_manifest_section(METADATA)['date']])
+          valid_format=revs_is_valid_format?(row[get_manifest_section(METADATA)['format']])
+          bad_rows+=1 unless (valid_date && valid_format)
+        end
+        return bad_rows
+      end
+      
+      # pass in csv data from a file read in and it will tell you if the headers are valid
+      def check_headers(csv_data)
+        file_headers=csv_data[0].keys.reject(&:blank?).collect(&:downcase)
         #The file doesn't need to have all the metadata values, it just can't have headers that aren't used for metadata or registration
         if file_headers.include?('date') && file_headers.include?('year') # can't have both date and year 
           return false
@@ -110,7 +132,7 @@ module Revs
           return file_headers-get_manifest_section(METADATA).values-get_manifest_section(REGISTER).values == []
         end
       end
-
+      
       def clean_collection_name(name)
         return "" if name.blank? || name.nil?
         name=name.to_s
@@ -147,7 +169,13 @@ module Revs
         return row
       end
 
-      # check a single format single and fix some common issues
+      # checks to see if we have a valid format
+      def revs_is_valid_format?(format)
+        formats=format.split("|").collect{|f| f.strip}
+        !formats.collect {|f| revs_known_formats.include?(f)}.uniq.include?(false)
+      end
+      
+      # check a single format and fix some common issues
       def revs_check_format(format)
         return revs_check_formats([format]).first
       end
@@ -237,6 +265,14 @@ module Revs
         date_string.to_s.strip.scan(/\D/).empty? and (starting_year..Date.today.year).include?(date_string.to_i)
       end
 
+      # tell us if the incoming datestring supplied in the manifest column is a valid date, year or list of years
+      def revs_is_valid_datestring?(date_string)
+        return true if date_string.nil? || date_string.empty?
+        is_full_date=(get_full_date(date_string) != false)
+        is_year=!parse_years(date_string).empty?
+        return is_year || is_full_date
+      end
+      
       # tell us if the string passed is in is a full date of the format M/D/YYYY or m-d-yyyy or m-d-yy or M/D/YY, and returns the date object if it is valid
       def get_full_date(date_string)
         begin

data/spec/revs-utils_spec.rb

@@ -108,7 +108,27 @@ describe "Revs-Utils" do
      @revs.is_valid_year?('1700').should be_false # too old! no cars even existed yet
      @revs.is_valid_year?('1700',1600).should be_true # unless we allow it to be ok
   end
-   
+
+  it "should indicate if we have unknown formats" do
+    @revs.revs_is_valid_format?('slides').should be_true
+    @revs.revs_is_valid_format?('slide').should be_false
+    @revs.revs_is_valid_format?('slides | slide').should be_false
+    @revs.revs_is_valid_format?('slides | black-and-white negatives').should be_true
+    @revs.revs_is_valid_format?('black-and-white-negatives').should be_false
+    @revs.revs_is_valid_format?('black-and-white negatives').should be_true   
+  end
+  
+  it "should indicate if we have a valid datestring" do
+    @revs.revs_is_valid_datestring?('1959').should be_true
+    @revs.revs_is_valid_datestring?('bogus').should be_false
+    @revs.revs_is_valid_datestring?('').should be_true
+    @revs.revs_is_valid_datestring?(nil).should be_true
+    @revs.revs_is_valid_datestring?([]).should be_true
+    @revs.revs_is_valid_datestring?('2/2/1950').should be_true
+    @revs.revs_is_valid_datestring?('2/31/1950').should be_false
+    @revs.revs_is_valid_datestring?('2/2/50').should be_true
+    @revs.revs_is_valid_datestring?('195x').should be_true
+ end   
    
    it "should lookup the country correctly" do
      @revs.revs_get_country('USA').should == "United States"
@@ -137,6 +157,7 @@ describe "Revs-Utils" do
   it "should parse 1950s correctly" do
     
     @revs.parse_years('1950s').should == ['1950','1951','1952','1953','1954','1955','1956','1957','1958','1959']    
+    @revs.parse_years("1950's").should == ['1950','1951','1952','1953','1954','1955','1956','1957','1958','1959']    
     
   end
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: revs-utils
 version: !ruby/object:Gem::Version
-  version: 1.0.14
+  version: 1.0.15
 platform: ruby
 authors:
 - Peter Mangiafico
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-01-26 00:00:00.000000000 Z
+date: 2015-04-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: countries