snapcrawl 0.5.1 → 0.5.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 83077ce26e7825a14948d6048ff92ee5cc15a17e3eeaae5687d891b68d83cd92
-  data.tar.gz: 15f264f683c08f9c0a2d379fa76e3dad1698a30808fb3694ccf334b45450c8ff
+  metadata.gz: 62a293da259afce5690315f27f2bbcd881e495a3d1b5344eb9ed9e2c46bd4a4d
+  data.tar.gz: d600fdbcd2344e5a19f853cbea67a0d8ad0c365a38d00aa4de8d02dd6e52e5b0
 SHA512:
-  metadata.gz: 67d1a62d2c24dbf675c7e83449782ed88e7de62122082033239bc9ce436566dea5f1bad5b442784f1a476d05d75f20a52ec17c3a0fd6bd3a3a4ef7dcf7f76d07
-  data.tar.gz: ab17a294bad674814dd1e47407f9da58ae9c81dd1becee0db9d57d944c41ea6727cad6123b03797bd238b061b3138d9c4c42b32bd7f384a04716aeb4d93ee670
+  metadata.gz: 3ebdb2355480bacd7f7a6faba264a31086e68c1864c692607fdb6fbc11df210eee17af936ab63305484ee46ac473d50b4033be11e995b51b9050b359c81dd906
+  data.tar.gz: 42a0a9f048fe9b5b1b04426d444710a256ccc8e9a914e3277f062c4ebf760d50a018c1f189e7b0cebced1c236f5d13ca56ab4abbf808a5ec4812bf9a754a9343

data/README.md

@@ -27,7 +27,7 @@ You can run Snapcrawl by using this docker image (which contains all the
 necessary prerequisites):
 
 ```shell
-$ alias snapcrawl='docker run --rm -it --volume $PWD:/app dannyben/snapcrawl'
+$ alias snapcrawl='docker run --rm -it --network host --volume "$PWD:/app" dannyben/snapcrawl'
 ```
 
 For more information on the Docker image, refer to the [docker-snapcrawl][3] repository.

data/lib/snapcrawl/config.rb

@@ -16,10 +16,10 @@ module Snapcrawl
         #        Config. The $logger is available, but it was not yet fully
         #        configured with log_level etc.
         if File.exist? file
-          # $logger.debug "loading config file %{green}#{file}%{reset}"
+          # $logger.debug "loading config file !txtgrn!#{file}"
           push file
         else
-          # $logger.debug "creating config file %{green}#{file}%{reset}"
+          # $logger.debug "creating config file !txtgrn!#{file}"
           create_config file
         end
       end

data/lib/snapcrawl/crawler.rb

@@ -7,7 +7,7 @@ module Snapcrawl
     attr_reader :url
 
     def initialize(url)
-      $logger.debug "initializing crawler with %{green}#{url}%{reset}"
+      $logger.debug "initializing crawler with !txtgrn!#{url}"
       
       config_for_display = Config.settings.dup
       config_for_display['name_template'] = '%%{url}' 
@@ -25,7 +25,7 @@ module Snapcrawl
   private
 
     def process_todo
-      $logger.debug "processing queue: %{green}#{todo.count} remaining%{reset}"
+      $logger.debug "processing queue: !txtgrn!#{todo.count} remaining"
 
       url, page = todo.shift
       done.push url
@@ -40,12 +40,12 @@ module Snapcrawl
         next if todo.has_key?(sub_page) or done.include?(sub_page)
         
         if Config.url_whitelist and sub_page.path !~ /#{Config.url_whitelist}/
-          $logger.debug "ignoring %{purple}%{underlined}#{sub_page.url}%{reset}, reason: whitelist"
+          $logger.debug "ignoring !undpur!#{sub_page.url}!txtrst!, reason: whitelist"
           next
         end
 
         if Config.url_blacklist and sub_page.path =~ /#{Config.url_blacklist}/
-          $logger.debug "ignoring %{purple}%{underlined}#{sub_page.url}%{reset}, reason: blacklist"
+          $logger.debug "ignoring !undpur!#{sub_page.url}!txtrst!, reason: blacklist"
           next
         end
 
@@ -56,7 +56,7 @@ module Snapcrawl
     def process_page(page)
       outfile = "#{Config.snaps_dir}/#{Config.name_template}.png" % { url: page.url.to_slug }
 
-      $logger.info "processing %{purple}%{underlined}#{page.url}%{reset}, depth: #{page.depth}"
+      $logger.info "processing !undpur!#{page.url}!txtrst!, depth: #{page.depth}"
 
       if !page.valid?
         $logger.debug "page #{page.path} is invalid, aborting process"
@@ -66,7 +66,7 @@ module Snapcrawl
       if file_fresh? outfile
         $logger.info "screenshot for #{page.path} already exists"
       else
-        $logger.info "%{bold}capturing screenshot for #{page.path}%{reset}"
+        $logger.info "!bldgrn!capturing screenshot for #{page.path}"
         save_screenshot page, outfile
       end
 
@@ -76,7 +76,7 @@ module Snapcrawl
     def save_screenshot(page, outfile)
       page.save_screenshot outfile
     rescue => e
-      $logger.error "screenshot error on %{purple}%{underlined}#{page.path}%{reset} - %{red}#{e.class}%{reset}: #{e.message}"
+      $logger.error "screenshot error on !undpur!#{page.path}!txtrst! - !txtred!#{e.class}!txtrst!: #{e.message}"
     end
 
     def file_fresh?(file)

data/lib/snapcrawl/dependencies.rb

@@ -8,10 +8,10 @@ module Snapcrawl
       def verify
         return if @verified
 
-        $logger.debug 'verifying %{green}phantomjs%{reset} is present'
+        $logger.debug 'verifying !txtgrn!phantomjs!txtrst! is present'
         raise MissingPhantomJS unless command_exist? "phantomjs"
 
-        $logger.debug 'verifying %{green}imagemagick%{reset} is present'
+        $logger.debug 'verifying !txtgrn!imagemagick!txtrst! is present'
         raise MissingImageMagick unless command_exist? "convert"
 
         @verified = true

data/lib/snapcrawl/log_helpers.rb

@@ -1,57 +1,35 @@
+require 'colsole'
+
 module Snapcrawl
   module LogHelpers
+    include Colsole
+
     SEVERITY_COLORS = {
-      'INFO' => :blue,
-      'WARN' => :yellow,
-      'ERROR' => :red,
-      'FATAL' => :red,
-      'DEBUG' => :cyan
+      'INFO' => :txtblu,
+      'WARN' => :txtylw,
+      'ERROR' => :txtred,
+      'FATAL' => :txtred,
+      'DEBUG' => :txtcyn
     }
     
     def log_formatter
       proc do |severity, _time, _prog, message|
         severity_color = SEVERITY_COLORS[severity]
-
-        "%{#{severity_color}}#{severity.rjust 5}%{reset} : #{message}\n" % log_colors
+        line = "!#{severity_color}!#{severity.rjust 5}!txtrst! : #{message}\n"
+        use_colors? ? colorize(line) : strip_color_markers(line)
       end
     end
 
-    def log_colors
-      @log_colors ||= log_colors!
-    end
-
-    def log_colors!
-      colors? ? actual_colors : empty_colors
-    end
-
-    def actual_colors
-      {
-        red: "\e[31m", green: "\e[32m", yellow: "\e[33m",
-        blue: "\e[34m", purple: "\e[35m", cyan: "\e[36m",
-        underlined: "\e[4m", bold: "\e[1m",
-        none: "", reset: "\e[0m"
-      }
-    end
-
-    def empty_colors
-      {
-        red: "", green: "", yellow: "", 
-        blue: "", purple: "", cyan: "",
-        underlined: "", bold: "",
-        none: "", reset: ""
-      }
-    end
-
-    def colors?
-      if Config.log_color == 'auto'
-        tty?
-      else
-        Config.log_color
-      end
+    def use_colors?
+      @use_colors ||= (Config.log_color == 'auto' ? tty? : Config.log_color)
     end
 
     def tty?
       ENV['TTY'] == 'on' ? true : ENV['TTY'] == 'off' ? false : $stdout.tty?
     end
+
+    def strip_color_markers(text)
+      text.gsub(/\!([a-z]{6})\!/, '')
+    end
   end
 end

data/lib/snapcrawl/page.rb

@@ -55,13 +55,13 @@ module Snapcrawl
       response = cache.get(url) { HTTParty.get url }
 
       if !response.success?
-        $logger.warn "http error on %{purple}%{underlined}#{url}%{reset}, code: %{yellow}#{response.code}%{reset}, message: #{response.message.strip}"
+        $logger.warn "http error on !undpur!#{url}!txtrst!, code: !txtylw!#{response.code}!txtrst!, message: #{response.message.strip}"
       end
 
       response
 
     rescue => e
-      $logger.error "http error on %{purple}%{underlined}#{url}%{reset} - %{red}#{e.class}%{reset}: #{e.message}"
+      $logger.error "http error on !undpur!#{url}!txtrst! - !txtred!#{e.class}!txtrst!: #{e.message}"
       nil
 
     end
@@ -95,7 +95,7 @@ module Snapcrawl
       begin
         link = Addressable::URI.join(url, link).to_s.dup
       rescue => e
-        $logger.warn "%{red}#{e.class}%{reset}: #{e.message} on #{path} (link: #{link})"
+        $logger.warn "!txtred!#{e.class}!txtrst!: #{e.message} on #{path} (link: #{link})"
         return nil
       end
 

data/lib/snapcrawl/pretty_logger.rb

@@ -5,7 +5,7 @@ module Snapcrawl
     extend LogHelpers
 
     def self.new
-      Logger.new(STDOUT, formatter: log_formatter, level: Config.log_level)
+      Logger.new($stdout, formatter: log_formatter, level: Config.log_level)
     end
   end
 end

data/lib/snapcrawl/version.rb

@@ -1,3 +1,3 @@
 module Snapcrawl
-  VERSION = "0.5.1"
+  VERSION = "0.5.2"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: snapcrawl
 version: !ruby/object:Gem::Version
-  version: 0.5.1
+  version: 0.5.2
 platform: ruby
 authors:
 - Danny Ben Shitrit
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-03-14 00:00:00.000000000 Z
+date: 2021-02-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: colsole
@@ -166,7 +166,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.3
+rubygems_version: 3.2.3
 signing_key: 
 specification_version: 4
 summary: Crawl a website and take screenshots (CLI + Library)