vore 0.2.8-x86_64-linux → 0.3.0-x86_64-linux

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 322dfb93cbfeae7bdf008ed00e0c7ac67affa00d787bd18fc483c3fe583c9215
-  data.tar.gz: d96c5a555ae20af0c4c967539256a528b8914c49ffe3322fc286d3f7684c1355
+  metadata.gz: 5fa0065651385809a53579488f9985cc3332197d4fa9818508859bb24274a16b
+  data.tar.gz: a0754878a08d651215dd5df40220e3002255ef65fedcb365e0605ff228db27fd
 SHA512:
-  metadata.gz: bc438087f76029e4af3a9a28f25ac35800b72022299c63ecb57b3ddfccc96e17c625cad4a02d7e7209d7ac2123efa235163eb5bfa74f75d7105592c83dbc32db
-  data.tar.gz: 2874256c99d57c2269e04b604e716c00f3eba0bb7c287f11a2b96f5f0a58bd812a9d1740da419839c0e1000e96bea36c8f4e4bc130700ddb07293c3c9ecb26e5
+  metadata.gz: 48298f9bf6de3e76b443a4ea46fa0adc687e4a39b90fc94ae6a97e0ce0153c6e983e292af0c8d6a53dcb44e94b8ecba94f8adc75fe5ce8cd7857e947ab113795
+  data.tar.gz: 404712a662f24fff36346998e791cfac7c9ac44661005209f0a2a65b38146f0233dfcdc9cfc115c5772340d310d085dbadbca8766201fad810843359ccff1c98

data/lib/vore/configuration.rb

@@ -5,5 +5,9 @@ module Vole
     DEFAULT_SANITIZATION_CONFIG = Selma::Sanitizer::Config::RELAXED.dup.merge({
       allow_doctype: false,
     })
+
+    DEFAULT_OPTIONS = {
+      delay: 3500,
+    }
   end
 end

data/lib/vore/crawler.rb

@@ -12,7 +12,7 @@ module Vore
 
     # Creates a crawler
     # denylist: Sets a denylist filter, allows a regexp, string or array of either to be matched.
-    def initialize(denylist: /a^/, sanitization_config: Vole::Configuration::DEFAULT_SANITIZATION_CONFIG)
+    def initialize(denylist: /a^/, sanitization_config: Vole::Configuration::DEFAULT_SANITIZATION_CONFIG, options: Vole::Configuration::DEFAULT_OPTIONS)
       @denylist_regexp = Regexp.union(denylist)
 
       @content_extractor = Vole::Handlers::ContentExtractor.new
@@ -20,6 +20,7 @@ module Vore
       ext = PLATFORM.include?("windows") ? ".exe" : ""
       @executable = File.expand_path([__FILE__, "..", "..", "..", "exe", "vore-spider#{ext}"].join(FILE_SEPERATOR))
       @parent_output_dir = "tmp/vore"
+      @options = options
 
       return if File.exist?(@executable)
 
@@ -31,7 +32,7 @@ module Vore
       @output_dir = "#{@parent_output_dir}/#{website.gsub(/[^a-zA-Z0-9]/, "_").squeeze("_")}"
       Vore.logger.info("Vore started crawling #{website}, outputting to #{output_dir}")
 
-      output = run_command(website, @output_dir)
+      output = run_command(website, delay: @options[:delay])
 
       Vore.logger.info("Vore finished crawling #{website}: #{output}")
 
@@ -86,14 +87,14 @@ module Vore
     #   crawl_site(site)
     # end
 
-    def run_command(website, output_dir)
+    def run_command(website, delay: 3500)
       %x(#{@executable} \
         --user-agent #{user_agent} \
-        --delay 3500 \
+        --delay #{delay} \
         --url #{website} \
         download \
         -t \
-        #{output_dir})
+        #{@output_dir})
     end
 
     def user_agent

data/lib/vore/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Vore
-  VERSION = "0.2.8"
+  VERSION = "0.3.0"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: vore
 version: !ruby/object:Gem::Version
-  version: 0.2.8
+  version: 0.3.0
 platform: x86_64-linux
 authors:
 - Garen J. Torikian