fluent-plugin-email-obfuscate 0.0.1 → 0.0.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 144ea941d820cf9397f8950c7393d9873f27306d
-  data.tar.gz: cbff42d1234a688b1d4da4e6bab9a4fe224c3b1a
+  metadata.gz: c8c95d15752cf1fbeeb6e933d0d42423b05ebaf1
+  data.tar.gz: 177b3fd14bf1c47b985aed8ac168e861006c0327
 SHA512:
-  metadata.gz: 190489016407ed6321b030ee8088fcdec9e68eae78194b08b79b85c0cd0f215b16494e0492181f76862f12e5d5141fe91cfa0ec8b8a47769596b1672fcae8bae
-  data.tar.gz: efe07cd21cd9144f187ef37e8bc1d34e1034731ec75f1bdda26703a5dad69b4de9f07d3b1411cae3570dfd08dc6580b5f25f390da0718d71ef521821a9a1164b
+  metadata.gz: ba7799b236bddd340016ffe52d09f4f7bc489c08d643fc173c0d2a615d9d2a64ef9c71cab921af779467e3acfb9fb33615a301860f7c2a8bf1e92e23d81fc9f6
+  data.tar.gz: 951b4832e624681add3ac22b0e9ca1e57d02bd2392e3cf3c822d56feda391710b2a889f364992dafbb67030975c3b9ac4e3bd597fef9382259bfec8ad29854fe

data/.gitignore

@@ -0,0 +1 @@
+**/*.gem

data/README.md

@@ -2,7 +2,10 @@
 
 [Fluentd](https://fluentd.org/) filter plugin obfuscate email addresses.
 
-This filter attempts to parse each field in the record, and will replace the "domain" part of the email address with asterisks.
+[![CodeFactor](https://www.codefactor.io/repository/github/JamesJJ/fluent-plugin-email-obfuscate/badge)](https://www.codefactor.io/repository/github/JamesJJ/fluent-plugin-email-obfuscate)
+[![Gem Version](https://badge.fury.io/rb/fluent-plugin-email-obfuscate.svg)](https://badge.fury.io/rb/fluent-plugin-email-obfuscate)
+
+This filter attempts to parse each field in the record, and will obfuscate all or part of any email addresses it finds, by replacing with the same number of asterisks.
 
 ### For example:
 
@@ -66,22 +69,39 @@ $ bundle
 
 ## Configuration
 
-_CURRENTLY THIS FILTER DOES NOT ACCEPT ANY CONFIGURATION_
-
-<!---
-You can generate configuration template:
+Use `email_obfuscate` filter.
 
 ```
-$ fluent-plugin-config-format filter email-obfuscate
+    <filter **>
+        @type email_obfuscate
+        mode <mode>
+        suffix_whitelist <list>
+    </filter>
 ```
 
-You can copy and paste generated documents here.
--->
+### 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 | `********@*******.***`
+
+_Note: `.` and `@` are never replaced_
+
+### suffix_whitelist (optional)
+
+A list of suffixes not to obfuscate. For example, with config `suffix_whitelist [".example.com", "@example.com"]` the result would be:
+
+Input                        | Action 
+:--                          | :--   
+`user@example.com`           | no change
+`user@subdomain.example.com` | no change
+`user@example.museum`        | obfuscated
 
 ## Todo
 
 * Add tests!
-* Support whitelists of domains not to be obfuscated
 * Support configuration of which fields to act upon
 * . . .
 

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 = "0.0.1"
+  spec.version = "0.0.3"
   spec.authors = ["JamesJJ"]
   spec.email   = ["jj@fcg.fyi"]
 
@@ -19,6 +19,7 @@ Gem::Specification.new do |spec|
   spec.executables   = files.grep(%r{^bin/}) { |f| File.basename(f) }
   spec.test_files    = test_files
   spec.require_paths = ["lib"]
+  spec.required_ruby_version = '>= 2.3.0'
 
   spec.add_development_dependency "bundler", "~> 1.14"
   spec.add_development_dependency "rake", "~> 12.0"

data/lib/fluent/plugin/filter_email_obfuscate.rb

@@ -20,33 +20,82 @@ module Fluent
     class EmailObfuscateFilter < Fluent::Plugin::Filter
       Fluent::Plugin.register_filter("email_obfuscate", self)
 
+      config_param :mode, :string, default: 'partial_name',
+        desc: <<-DESC
+'full' will replace all characters.
+'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.
+DESC
+
+      config_param :suffix_whitelist, :array, default: [],
+        desc: <<-DESC
+List of email suffixes not to obfuscate e.g. ['@example.com', '.example.com'] would result in:
+user@example.com => unchanged
+user@subdomain.example.com => unchanged
+user@example.net => obfuscated
+DESC
+
+      def configure(conf)
+        super
+
+        if conf.has_key?('mode')
+          raise ConfigError, "'mode' must be one of: domain_only, full, partial_name" unless
+            ['domain_only', 'full', 'partial_name'].include?(conf.dig('mode'))
+        end
+      end
+
+      def hide_partial(str)
+          case
+          when str.length < 5
+            len = str.length
+          when str.length < 11
+            len = (str.length / 2) + 2
+          else
+            len = (str.length / 3) + 4
+          end
+          str[0, len] + '*' * (str.length - len)
+      end
+
       def obfuscate(str)
-        str.match(/^([^@]+)(@.+)$/){|m| m[1] + m[2].tr('@.a-zA-Z0-9','@.*') } || str
+        strmatch = str.match(/^([^@]+)(@.+)$/) { |m|
+           case @mode
+           when 'domain_only'
+             m[1] + m[2].tr("@.a-zA-Z0-9", "@.*")
+           when 'full'
+             m[1].gsub(/./, '*') + m[2].tr("@.a-zA-Z0-9", "@.*")
+           else
+             hide_partial(m[1]) + m[2].tr("@.a-zA-Z0-9", "@.*")
+           end
+        }
+        if strmatch.nil?
+          str
+        elsif @suffix_whitelist.select{ |a| str.downcase.end_with?(a.downcase)}.empty?
+          strmatch
+        else
+          str
+        end
       end
 
       def deep_process(o)
-
         case o
         when String
           os = o.scan(/<([^<>]+@[^<>]+)>/)
-          o = os.length.zero? ? obfuscate(o) : deep_process(os).join(', ')
-
+          o = os.length.zero? ? obfuscate(o) : deep_process(os).join(", ")
         when Array
           o.map! do |obj|
             deep_process(obj)
           end
-
         when Hash
           o.each_pair do |key, value|
             o[key] = deep_process(value)
           end
-
         end
         o
       end
 
       def filter(tag, time, record)
-         deep_process(record)
+        deep_process(record)
       end
     end
   end

data/tool/build_and_push_to_rubygems.sh

@@ -0,0 +1,7 @@
+#!/bin/bash
+
+rm -f ./*.gem
+gem build fluent-plugin-email-obfuscate.gemspec && \
+  gem push fluent-plugin-email-obfuscate-*.gem
+rm -f ./*.gem
+

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: fluent-plugin-email-obfuscate
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.0.3
 platform: ruby
 authors:
 - JamesJJ
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2018-05-09 00:00:00.000000000 Z
+date: 2018-05-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -79,6 +79,7 @@ executables: []
 extensions: []
 extra_rdoc_files: []
 files:
+- ".gitignore"
 - Gemfile
 - LICENSE
 - README.md
@@ -87,6 +88,7 @@ files:
 - lib/fluent/plugin/filter_email_obfuscate.rb
 - test/helper.rb
 - test/plugin/test_filter_email_obfuscate.rb
+- tool/build_and_push_to_rubygems.sh
 homepage: https://github.com/JamesJJ/fluent-plugin-email-obfuscate
 licenses:
 - Apache-2.0
@@ -99,7 +101,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: '0'
+      version: 2.3.0
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="