data_collector 0.21.0 → 0.23.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 919f205875cd4bdd02d951b50fba36833743c96eb887d8b183bdb4fabcff62d2
-  data.tar.gz: 6feefeb606971a467dbf1cd47d39a3e76b5e23563fbd81c7be37f07f9c2f4bc2
+  metadata.gz: 1a6db4367b12a41101620cc818a1995868aab20b65157de8c19ea60c68aa90a0
+  data.tar.gz: 39a2cbcd665f85d71bda741a131c54d4f2871ae83ae63b86c9eed9f0d84e405a
 SHA512:
-  metadata.gz: d51898627739f4047234e617ed10e8686cbcdc76782ea9cdf78ed5c566ed18801de0938d0abe6d57d734297a066904043a29027d33b3de6ef35fbe255cf19c2b
-  data.tar.gz: cd05871afe0d9d9bcb698abb96292d7ed4780d41c3ddeb74db649abb7512f4c06f3991c11211cb2d574d7636de68a4fa1b23fa817943f894666fa64ff4aa31f8
+  metadata.gz: 5d7e6e243f9693603c4502b2f1bbe2fb64b8cf95718be257066940d6fafed24a8352ded805f285b617e25f5fbd703bf6e812dbd9f2b85d5c22cc143e7cdfdc1e
+  data.tar.gz: 73b0c03ef6ccb7855cfc1b6f424c543cc00fc79e182ccc2525c068f3b8d97535e4eaaaa66b0fd59225049d6a0dfb3dc55861319638dfb4c6df489ede6c1f9077

data/lib/data_collector/input/dir.rb

@@ -8,6 +8,17 @@ module DataCollector
         super
       end
 
+      def run(should_block = false, &block)
+        @listener.start unless running?
+        if block_given?
+          while should_block && !paused?
+            yield
+          end
+        else
+          sleep if should_block && running? && !paused?
+        end
+      end
+
       def running?
         @listener.processing?
       end
@@ -15,7 +26,10 @@ module DataCollector
       private
 
       def create_listener
-        @listener ||= Listen.to("#{@uri.host}/#{@uri.path}", @options) do |modified, added, _|
+        absolute_path = File.absolute_path("#{URI.decode_uri_component(@uri.to_s)}")
+        raise DataCollector::Error, "#{@uri.to_s} not found" unless File.exist?(absolute_path)
+
+        @listener ||= Listen.to(absolute_path, @options) do |modified, added, _|
           files = added | modified
           files.each do |filename|
             handle_on_message(@input, @output, filename)

data/lib/data_collector/input/generic.rb

@@ -4,7 +4,7 @@ module DataCollector
   class Input
     class Generic
       def initialize(uri, options = {})
-        @uri = uri
+        @uri = URI(URI.decode_uri_component(uri.to_s))
         @options = options
         @running = false
 

data/lib/data_collector/input.rb

@@ -28,6 +28,10 @@ module DataCollector
     def from_uri(source, options = {})
       source = CGI.unescapeHTML(source)
       @logger.info("Reading #{source}")
+      raise DataCollector::Error 'from_uri expects a scheme like file:// of https://' unless source =~ /:\/\//
+
+      scheme, path = source.split('://')
+      source="#{scheme}://#{URI.encode_uri_component(path)}"
       uri = URI(source)
       begin
         data = nil
@@ -37,11 +41,12 @@ module DataCollector
         when 'https'
           data = from_https(uri, options)
         when 'file'
-          if File.directory?("#{uri.host}/#{uri.path}")
-            raise DataCollector::Error, "#{uri.host}/#{uri.path} not found" unless File.exist?("#{uri.host}/#{uri.path}")
+          absolute_path = File.absolute_path("#{URI.decode_uri_component("#{uri.host}#{uri.path}")}")
+          if File.directory?(absolute_path)
+            #raise DataCollector::Error, "#{uri.host}/#{uri.path} not found" unless File.exist?("#{uri.host}/#{uri.path}")
             return from_dir(uri, options)
           else
-            raise DataCollector::Error, "#{uri.host}/#{uri.path} not found" unless File.exist?("#{uri.host}/#{uri.path}")
+            #            raise DataCollector::Error, "#{uri.host}/#{uri.path} not found" unless File.exist?("#{uri.host}/#{uri.path}")
             data = from_file(uri, options)
           end
         when /amqp/
@@ -75,6 +80,7 @@ module DataCollector
     end
 
     def from_https(uri, options = {})
+      uri = URI.decode_uri_component("#{uri.to_s}")
       data = nil
       if options.with_indifferent_access.include?(:logging) && options.with_indifferent_access[:logging]
         HTTP.default_options = HTTP::Options.new(features: { logging: { logger: @logger } })
@@ -154,7 +160,8 @@ module DataCollector
 
     def from_file(uri, options = {})
       data = nil
-      absolute_path = File.absolute_path("#{uri.host}#{uri.path}")
+      absolute_path = File.absolute_path("#{URI.decode_uri_component(uri.to_s)}")
+      raise DataCollector::Error, "#{uri.to_s} not found" unless File.exist?("#{absolute_path}")
       unless options.has_key?('raw') && options['raw'] == true
         @raw = data = File.read("#{absolute_path}")
         case File.extname(absolute_path)

data/lib/data_collector/version.rb

@@ -1,4 +1,4 @@
 # encoding: utf-8
 module DataCollector
-  VERSION = "0.21.0"
+  VERSION = "0.23.0"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: data_collector
 version: !ruby/object:Gem::Version
-  version: 0.21.0
+  version: 0.23.0
 platform: ruby
 authors:
 - Mehmet Celik
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2023-07-17 00:00:00.000000000 Z
+date: 2023-07-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport