fluent-plugin-email-obfuscate 0.0.4 → 0.0.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: b5589da8d1f315ff538febbe33596f14c9f716d1
-  data.tar.gz: e78e4f9b610888244cfd4a2e5716f42f3a433743
+  metadata.gz: 7ecc5c56f7719e9bae0229b222a93f9802999dc8
+  data.tar.gz: 5fa7276948f2f1ea0ec8fbecda015b30870f419f
 SHA512:
-  metadata.gz: df54f8e6f6ac2edd5a5717f5be94a67a8ca9c240078b21ebe103fba1397afc0e2bac8de6297418f496001b3c9898df6e223a7f93a2e9510fa24590426905f029
-  data.tar.gz: f3f6bbe3a21a66a5360821b76e6f6420f3f6c20eea9586763060e6f4c7f2122d2b19acb0d8ad8331720fc3f4297d2b3a317c43ad1c0690b46c0538f48d71b756
+  metadata.gz: 6b28e85baf32c876efb6c184b3da1ea7c29317e0a52b8651b8a31fac4a58dd071f1944df3f10a59c500884bf32b63bff7bf57a083c8c0760b85e0dc67c7da26b
+  data.tar.gz: 8c3a7376acfa39ad45450493ef81d9fa74936de85e06f45e069470aaad8e7a6472929507ec8cc8a2da5d6ab50dfbf76d09b39f354613c53ed36be03e6e0076d7

data/.travis.yml

@@ -1,4 +1,6 @@
 language: ruby
 rvm:
- - 2.3
+ - 2.6
+ - 2.5
  - 2.4
+ - 2.3

data/README.md

@@ -82,11 +82,12 @@ Use `email_obfuscate` filter.
 
 ### mode (default: partial_name)
 
-mode            | Action | Example
-:--             | :--    | :-- 
-`domain_only`   | Only replace all characters in the domain part | `testuser@*******.***`
-`partial_name`  | Replace all characters in domain and partially in the name | `testu***@*******.***`
-`full`          | Replace all characters in name and domain part of address | `********@*******.***`
+mode              | Action | Example
+:--               | :--    | :-- 
+`domain_only`     | Only replace all characters in the domain part | `testuser@*******.***`
+`partial_name`    | Replace all characters in domain and partially in the name | `testu***@*******.***`
+`name_substring`  | Replace all characters in name and maintain surrounding context | `My old email address ********@example.com does not work any more`
+`full`            | Replace all characters in name and domain part of address | `********@*******.***`
 
 _Note: `.` and `@` are never replaced_
 
@@ -102,12 +103,12 @@ Input                        | Action
 
 ## Todo
 
-* Add tests!
 * Support configuration of which fields to act upon
+* Fine grained configuration of email delimiters
 * . . .
 
 ## Copyright
 
-* Copyright(c) 2018 JamesJJ 
+* Copyright(c) 2018-2019 JamesJJ 
 * License
   * Apache License, Version 2.0

data/fluent-plugin-email-obfuscate.gemspec

@@ -3,7 +3,7 @@ $LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
 
 Gem::Specification.new do |spec|
   spec.name    = "fluent-plugin-email-obfuscate"
-  spec.version = "0.0.4"
+  spec.version = "0.0.5"
   spec.authors = ["JamesJJ"]
   spec.email   = ["jj@fcg.fyi"]
 

data/lib/fluent/plugin/filter_email_obfuscate.rb

@@ -20,9 +20,10 @@ module Fluent
     class EmailObfuscateFilter < Fluent::Plugin::Filter
       Fluent::Plugin.register_filter("email_obfuscate", self)
 
-      config_param :mode, :enum, list: [:partial_name, :full, :domain_only],  default: :partial_name,
+      config_param :mode, :enum, list: [:name_substring, :partial_name, :full, :domain_only],  default: :partial_name,
         desc: <<-DESC
 'full' will replace all characters.
+'name_substring' will replace the 'name' half of the address, and attempt to retain surrounding context.
 'partial_name' will replace all characters in the 'domain' half of the address, and a subset of the 'name'.
 'domain_only' will only replace characters in the 'domain' half of the address.
 '.' and '@' are never replaced.
@@ -59,6 +60,8 @@ DESC
              m[1] + m[2].tr("@.a-zA-Z0-9", "@.*")
            when :full
              m[1].gsub(/./, '*') + m[2].tr("@.a-zA-Z0-9", "@.*")
+           when :name_substring
+             m[1].gsub(/./, '*') + m[2]
            else
              hide_partial(m[1]) + m[2].tr("@.a-zA-Z0-9", "@.*")
            end
@@ -75,8 +78,15 @@ DESC
       def deep_process(o)
         case o
         when String
-          os = o.scan(/<([^<>]+@[^<>]+)>/)
-          o = os.length.zero? ? obfuscate(o) : deep_process(os).join(", ")
+          case @mode
+          when :name_substring
+            o = o.gsub(/([a-zA-Z0-9_+\.\-]+@[a-zA-Z0-9\.\-]+)/) { |om|
+              obfuscate(om)
+            }
+          else
+            os = o.scan(/<([^<>]+@[^<>]+)>/)
+            o = os.length.zero? ? obfuscate(o) : deep_process(os).join(", ")
+          end
         when Array
           o.map! do |obj|
             deep_process(obj)

data/test/plugin/test_filter_email_obfuscate.rb

@@ -27,7 +27,8 @@ class EmailObfuscateFilterTest < Test::Unit::TestCase
           "field": "name3@example.museum"
         }
       },
-      "email_string": "Jane Doe <jane@example.name>, John Smith <john@example.name>"
+      "email_string": "Jane Doe <jane@example.name>, John Smith <john@example.name>",
+      "email_string_with_trailing_text": "Jane Doe <jane@example.name>, John Smith <john@example.name> trailing text is here"
     }
   end
 
@@ -46,7 +47,9 @@ class EmailObfuscateFilterTest < Test::Unit::TestCase
                  list1: ["user*@*******.***", "user*@*******.***"],
                  a: {nested:
                        {field: "name*@*******.******"}},
-                 email_string: "jane@*******.****, john@*******.****"}]
+                 email_string: "jane@*******.****, john@*******.****",
+                 email_string_with_trailing_text: "jane@*******.****, john@*******.****"}
+               ]
     assert_equal expected, d.filtered.map{|e| e.last}
   end
 
@@ -59,7 +62,9 @@ class EmailObfuscateFilterTest < Test::Unit::TestCase
                  list1: ["*****@*******.***", "*****@*******.***"],
                  a: {nested:
                        {field: "*****@*******.******"}},
-                 email_string: "****@*******.****, ****@*******.****"}]
+                 email_string: "****@*******.****, ****@*******.****",
+                 email_string_with_trailing_text: "****@*******.****, ****@*******.****"}
+               ]
     assert_equal expected, d.filtered.map{|e| e.last}
   end
 
@@ -72,12 +77,29 @@ class EmailObfuscateFilterTest < Test::Unit::TestCase
                  list1: ["user1@*******.***", "user2@*******.***"],
                  a: {nested:
                        {field: "name3@*******.******"}},
-                 email_string: "jane@*******.****, john@*******.****"}]
+                 email_string: "jane@*******.****, john@*******.****",
+                 email_string_with_trailing_text: "jane@*******.****, john@*******.****"}
+               ]
+    assert_equal expected, d.filtered.map{|e| e.last}
+  end
+
+  test "filter_name_substring" do
+    d = create_driver(CONF + %[mode name_substring])
+    d.run(default_tag: 'test') do
+      d.feed(sample_records)
+    end
+    expected = [{f1: "*******@example.net",
+                 list1: ["*****@example.com", "*****@example.org"],
+                 a: {nested:
+                       {field: "*****@example.museum"}},
+                 email_string: "Jane Doe <****@example.name>, John Smith <****@example.name>",
+                 email_string_with_trailing_text: "Jane Doe <****@example.name>, John Smith <****@example.name> trailing text is here"}
+               ]
     assert_equal expected, d.filtered.map{|e| e.last}
   end
 
   test "suffix whitelist" do
-    d = create_driver(CONF + %[suffix_whitelist [".example.com", "@example.com"]])
+    d = create_driver(CONF + %[suffix_whitelist [".example.com", "@example.com"]\nmode partial_name])
     sample_records = {
       "f1": "myEmail@example.net",
       "list1": [
@@ -102,6 +124,35 @@ class EmailObfuscateFilterTest < Test::Unit::TestCase
     assert_equal expected, d.filtered.map{|e| e.last}
   end
 
+  test "suffix whitelist mode name_substring" do
+    d = create_driver(CONF + %[suffix_whitelist [".example.com", "@example.com"]\nmode name_substring])
+    sample_records = {
+      "f1": "myEmail@example.net",
+      "list1": [
+        "user1@example.com",
+        "user2@subdomain.example.com"
+      ],
+      "a": {
+        "nested": {
+          "field": "name3@example.museum"
+        }
+      },
+      "email_string": "Jane Doe <jane@example.name>, John Smith <john@example.name> and some trailing text"
+    }
+    d.run(default_tag: 'test') do
+      d.feed(sample_records)
+    end
+    expected = [{f1: "*******@example.net",
+                 list1: ["user1@example.com", "user2@subdomain.example.com"],
+                 a: {nested:
+                       {field: "*****@example.museum"}},
+                 email_string: "Jane Doe <****@example.name>, John Smith <****@example.name> and some trailing text"}]
+
+
+    assert_equal expected, d.filtered.map{|e| e.last}
+  end
+
+
   private
 
   def create_driver(conf = CONF)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: fluent-plugin-email-obfuscate
 version: !ruby/object:Gem::Version
-  version: 0.0.4
+  version: 0.0.5
 platform: ruby
 authors:
 - JamesJJ
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2018-10-03 00:00:00.000000000 Z
+date: 2019-05-14 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler