twords 0.1.2 → 0.1.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 3bd960588a5adf41656a50a83db3a70a527bd128
-  data.tar.gz: da855462537cc299c03e662223b073c528fb2263
+  metadata.gz: 0d9a3acb581e8d88cccecaa1f81755b24ccae3d8
+  data.tar.gz: e537d3ea87fec86ae57d06f3902b9b27a2ddfa89
 SHA512:
-  metadata.gz: 35a78833f558014514f74e06f9ae744add07bc16c0e109adfd6370cfe4bab07e189b2a45e08461e539ba501af3c39a7b26275f0294805f4589ca8f73b1de8f43
-  data.tar.gz: 98ea7895c7cec6745d498aa6b2ffb0164096801628d08589a3889732eb3ce54cec4e3c408fb79c248c5bfabf3232f464026fb1e95a3411f7ba6fe250176105c9
+  metadata.gz: c4adc74dfeb9c6aae7897aca0a8daf5eb519d3ac53cb826fee954b3ead3a38c462a5fc010c88ca28199012456759739ca0f3c1c544137e0558ca349b19fad53d
+  data.tar.gz: 2838be4f9d6f6a83fb8aab1cdee4c201d1665a545c595cd64e3fddde6a7c2e53bc263a44544ada42b531caaa12f94f8b5c1772d4f8d72cd13f8f89c3bdffad07

data/README.md

@@ -24,8 +24,8 @@ Or install it yourself as:
 
 ```ruby
 Twords.config do |config|
-  config.rejects  = %w[the for and a i of if]
-  config.range        = 14
+  config.rejects = %w[the for and a i of if]
+  config.range   = 14
   config.up_to { Time.now } # A time object to be lazy evaluated. The range is counted backward from here.
   
   config.twitter_client do |twitter|

data/lib/twords.rb

@@ -1,7 +1,7 @@
 # frozen_string_literal: true
 
-require 'date'
 require 'twitter'
+require 'uri'
 
 require 'twords/version'
 
@@ -27,10 +27,11 @@ require 'twords/version'
 # # => { "butts"=>35, "poo"=>32, "pups"=>28, ... }
 class Twords
   class << self
-    attr_reader :rejects, :range, :client, :up_to_block
+    attr_reader :rejects, :range, :client, :up_to_block, :include_hashtags, :include_uris,
+                :include_mentions
 
-    def config(&block)
-      class_eval(&block)
+    def config
+      yield self
     end
 
     def twitter_client(&block)
@@ -41,12 +42,35 @@ class Twords
       @rejects = args.flatten
     end
 
+    def include_hashtags=(boolean)
+      not_a_boolean_error(boolean)
+      @include_hashtags = boolean
+    end
+
+    def include_uris=(boolean)
+      not_a_boolean_error(boolean)
+      @include_uris = boolean
+    end
+    alias include_urls include_uris
+
+    def include_mentions=(boolean)
+      not_a_boolean_error(boolean)
+      @include_mentions = boolean
+    end
+
+    def not_a_boolean_error(boolean)
+      raise ArgumentError, 'argument must be a booolean value' unless is_a_boolean?(boolean)
+    end
+
+    def is_a_boolean?(other)
+      [true, false].include?(other)
+    end
+
     def range=(integer)
       @range = integer
     end
 
     def up_to(&time_block)
-      raise ArgumentError, 'object must respond to #call' unless time_block.respond_to?(:call)
       @up_to_block = time_block
     end
   end
@@ -54,7 +78,7 @@ class Twords
   attr_reader :screen_names, :words, :requests, :client
 
   def initialize(*screen_names)
-    @screen_names = screen_names
+    @screen_names = screen_names.flatten
     @words        = {}
     @requests     = 0
   end
@@ -67,10 +91,37 @@ class Twords
     @_range ||= self.class.range
   end
 
+  def rejects
+    @_rejects ||= self.class.rejects
+  end
+
   def audited?
     @audited
   end
 
+  def hashtag?(word)
+    return false if self.class.include_hashtags
+    word.match?(/#/)
+  end
+
+  def uri?(word)
+    return false if self.class.include_uris
+    word.match?(URI.regexp)
+  end
+
+  def mention?(word)
+    return false if self.class.include_mentions
+    word.match?(/@/)
+  end
+
+  def hashtags
+    /#/
+  end
+
+  def should_be_skipped?(word)
+    rejects.include?(word) || hashtag?(word) || uri?(word) || mention?(word)
+  end
+
   def sort_words
     words.sort { |a, b| b.last <=> a.last }
   end
@@ -84,9 +135,11 @@ class Twords
   def fetch_timeline(screen_name)
     return [] if screen_name.to_s.empty?
     @requests += 1
-    timeline = client.user_timeline(screen_name, count: 200)
+    timeline = client.user_timeline(screen_name, tweet_mode: 'extended', count: 200)
     return timeline if timeline.empty?
-    fetch_older_tweets(timeline, screen_name)
+    timeline = fetch_older_tweets(timeline, screen_name)
+    puts "Fetched #{screen_name}'s timeline"
+    timeline
   end
 
   def fetch_older_tweets(timeline, screen_name)
@@ -94,6 +147,7 @@ class Twords
     @requests += 1
     timeline += client.user_timeline(
       screen_name,
+      tweet_mode: 'extended',
       max_id: timeline.last.id - 1,
       count: 200
     )
@@ -107,15 +161,15 @@ class Twords
   end
 
   def age_of_tweet_in_days(tweet)
-    (self.class.up_to_block.call - tweet.created_at) / 60 / 60 / 24
+    (self.class.up_to_block.call.to_time - tweet.created_at) / 86400
   end
 
   def count_words
+    words.clear
     recent_tweets.each do |tweet|
-      tweet_with_full_text = fetch_tweet_with_full_text(tweet)
-      words_array = tweet_with_full_text.attrs[:full_text].downcase.split(' ')
+      words_array = tweet.attrs[:full_text].downcase.split(' ')
       words_array.each do |word|
-        next if self.class.rejects.include?(word)
+        next if should_be_skipped?(word)
         if words.has_key?(word)
           words[word] += 1
         else
@@ -125,16 +179,16 @@ class Twords
     end
   end
 
-  def fetch_tweet_with_full_text(tweet)
-    @requests += 1
-    client.status(tweet.id, tweet_mode: 'extended')
-  end
-
   def audit
     count_words unless audited?
     @audited = true
   end
 
+  def audit!
+    @audited = false
+    audit
+  end
+
   def recent_tweets_count
     @_recent_tweets_count ||= recent_tweets.count
   end

data/lib/twords/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 class Twords
-  VERSION = '0.1.2'
+  VERSION = '0.1.3'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: twords
 version: !ruby/object:Gem::Version
-  version: 0.1.2
+  version: 0.1.3
 platform: ruby
 authors:
 - M. Simon Borg
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2017-06-21 00:00:00.000000000 Z
+date: 2017-06-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: twitter