ritter 0.0.87 → 0.0.88

data/lib/twitter-text/extractor.rb

@@ -0,0 +1,328 @@
+# encoding: UTF-8
+
+class String
+  # Helper function to count the character length by first converting to an
+  # array.  This is needed because with unicode strings, the return value
+  # of length may be incorrect
+  def char_length
+    if respond_to? :codepoints
+      length
+    else
+      chars.kind_of?(Enumerable) ? chars.to_a.size : chars.size
+    end
+  end
+
+  # Helper function to convert this string into an array of unicode characters.
+  def to_char_a
+    @to_char_a ||= if chars.kind_of?(Enumerable)
+      chars.to_a
+    else
+      char_array = []
+      0.upto(char_length - 1) { |i| char_array << [chars.slice(i)].pack('U') }
+      char_array
+    end
+  end
+end
+
+# Helper functions to return character offsets instead of byte offsets.
+class MatchData
+  def char_begin(n)
+    if string.respond_to? :codepoints
+      self.begin(n)
+    else
+      string[0, self.begin(n)].char_length
+    end
+  end
+
+  def char_end(n)
+    if string.respond_to? :codepoints
+      self.end(n)
+    else
+      string[0, self.end(n)].char_length
+    end
+  end
+end
+
+module Twitter
+  # A module for including Tweet parsing in a class. This module provides function for the extraction and processing
+  # of usernames, lists, URLs and hashtags.
+  module Extractor extend self
+    # Remove overlapping entities.
+    # This returns a new array with no overlapping entities.
+    def remove_overlapping_entities(entities)
+      # sort by start index
+      entities = entities.sort_by{|entity| entity[:indices].first}
+
+      # remove duplicates
+      prev = nil
+      entities.reject!{|entity| (prev && prev[:indices].last > entity[:indices].first) || (prev = entity) && false}
+      entities
+    end
+
+    # Extracts all usernames, lists, hashtags and URLs  in the Tweet <tt>text</tt>
+    # along with the indices for where the entity ocurred
+    # If the <tt>text</tt> is <tt>nil</tt> or contains no entity an empty array
+    # will be returned.
+    #
+    # If a block is given then it will be called for each entity.
+    def extract_entities_with_indices(text, options = {}, &block)
+      # extract all entities
+      entities = extract_urls_with_indices(text, options) +
+                 extract_hashtags_with_indices(text, :check_url_overlap => false) +
+                 extract_mentions_or_lists_with_indices(text) +
+                 extract_cashtags_with_indices(text)
+
+      return [] if entities.empty?
+
+      entities = remove_overlapping_entities(entities)
+
+      entities.each(&block) if block_given?
+      entities
+    end
+
+    # Extracts a list of all usernames mentioned in the Tweet <tt>text</tt>. If the
+    # <tt>text</tt> is <tt>nil</tt> or contains no username mentions an empty array
+    # will be returned.
+    #
+    # If a block is given then it will be called for each username.
+    def extract_mentioned_screen_names(text, &block) # :yields: username
+      screen_names = extract_mentioned_screen_names_with_indices(text).map{|m| m[:screen_name]}
+      screen_names.each(&block) if block_given?
+      screen_names
+    end
+
+    # Extracts a list of all usernames mentioned in the Tweet <tt>text</tt>
+    # along with the indices for where the mention ocurred.  If the
+    # <tt>text</tt> is nil or contains no username mentions, an empty array
+    # will be returned.
+    #
+    # If a block is given, then it will be called with each username, the start
+    # index, and the end index in the <tt>text</tt>.
+    def extract_mentioned_screen_names_with_indices(text) # :yields: username, start, end
+      return [] unless text
+
+      possible_screen_names = []
+      extract_mentions_or_lists_with_indices(text) do |screen_name, list_slug, start_position, end_position|
+        next unless list_slug.empty?
+        possible_screen_names << {
+          :screen_name => screen_name,
+          :indices => [start_position, end_position]
+        }
+      end
+
+      if block_given?
+        possible_screen_names.each do |mention|
+          yield mention[:screen_name], mention[:indices].first, mention[:indices].last
+        end
+      end
+
+      possible_screen_names
+    end
+
+    # Extracts a list of all usernames or lists mentioned in the Tweet <tt>text</tt>
+    # along with the indices for where the mention ocurred.  If the
+    # <tt>text</tt> is nil or contains no username or list mentions, an empty array
+    # will be returned.
+    #
+    # If a block is given, then it will be called with each username, list slug, the start
+    # index, and the end index in the <tt>text</tt>. The list_slug will be an empty stirng
+    # if this is a username mention.
+    def extract_mentions_or_lists_with_indices(text) # :yields: username, list_slug, start, end
+      return [] unless text =~ /[@＠]/
+
+      possible_entries = []
+      text.to_s.scan(Twitter::Regex[:valid_mention_or_list]) do |before, at, screen_name, list_slug|
+        match_data = $~
+        after = $'
+        unless after =~ Twitter::Regex[:end_mention_match]
+          start_position = match_data.char_begin(3) - 1
+          end_position = match_data.char_end(list_slug.nil? ? 3 : 4)
+          possible_entries << {
+            :screen_name => screen_name,
+            :list_slug => list_slug || "",
+            :indices => [start_position, end_position]
+          }
+        end
+      end
+
+      if block_given?
+        possible_entries.each do |mention|
+          yield mention[:screen_name], mention[:list_slug], mention[:indices].first, mention[:indices].last
+        end
+      end
+
+      possible_entries
+    end
+
+    # Extracts the username username replied to in the Tweet <tt>text</tt>. If the
+    # <tt>text</tt> is <tt>nil</tt> or is not a reply nil will be returned.
+    #
+    # If a block is given then it will be called with the username replied to (if any)
+    def extract_reply_screen_name(text) # :yields: username
+      return nil unless text
+
+      possible_screen_name = text.match(Twitter::Regex[:valid_reply])
+      return unless possible_screen_name.respond_to?(:captures)
+      return if $' =~ Twitter::Regex[:end_mention_match]
+      screen_name = possible_screen_name.captures.first
+      yield screen_name if block_given?
+      screen_name
+    end
+
+    # Extracts a list of all URLs included in the Tweet <tt>text</tt>. If the
+    # <tt>text</tt> is <tt>nil</tt> or contains no URLs an empty array
+    # will be returned.
+    #
+    # If a block is given then it will be called for each URL.
+    def extract_urls(text, &block) # :yields: url
+      urls = extract_urls_with_indices(text).map{|u| u[:url]}
+      urls.each(&block) if block_given?
+      urls
+    end
+
+    # Extracts a list of all URLs included in the Tweet <tt>text</tt> along
+    # with the indices. If the <tt>text</tt> is <tt>nil</tt> or contains no
+    # URLs an empty array will be returned.
+    #
+    # If a block is given then it will be called for each URL.
+    def extract_urls_with_indices(text, options = {:extract_url_without_protocol => true}) # :yields: url, start, end
+      return [] unless text && (options[:extract_url_without_protocol] ? text.index(".") : text.index(":"))
+      urls = []
+      position = 0
+
+      text.to_s.scan(Twitter::Regex[:valid_url]) do |all, before, url, protocol, domain, port, path, query|
+        valid_url_match_data = $~
+
+        start_position = valid_url_match_data.char_begin(3)
+        end_position = valid_url_match_data.char_end(3)
+
+        # If protocol is missing and domain contains non-ASCII characters,
+        # extract ASCII-only domains.
+        if !protocol
+          next if !options[:extract_url_without_protocol] || before =~ Twitter::Regex[:invalid_url_without_protocol_preceding_chars]
+          last_url = nil
+          last_url_invalid_match = nil
+          domain.scan(Twitter::Regex[:valid_ascii_domain]) do |ascii_domain|
+            last_url = {
+              :url => ascii_domain,
+              :indices => [start_position + $~.char_begin(0),
+                           start_position + $~.char_end(0)]
+            }
+            last_url_invalid_match = ascii_domain =~ Twitter::Regex[:invalid_short_domain]
+            urls << last_url unless last_url_invalid_match
+          end
+
+          # no ASCII-only domain found. Skip the entire URL
+          next unless last_url
+
+          # last_url only contains domain. Need to add path and query if they exist.
+          if path
+            # last_url was not added. Add it to urls here.
+            urls << last_url if last_url_invalid_match
+            last_url[:url] = url.sub(domain, last_url[:url])
+            last_url[:indices][1] = end_position
+          end
+        else
+          # In the case of t.co URLs, don't allow additional path characters
+          if url =~ Twitter::Regex[:valid_tco_url]
+            url = $&
+            end_position = start_position + url.char_length
+          end
+          urls << {
+            :url => url,
+            :indices => [start_position, end_position]
+          }
+        end
+      end
+      urls.each{|url| yield url[:url], url[:indices].first, url[:indices].last} if block_given?
+      urls
+    end
+
+    # Extracts a list of all hashtags included in the Tweet <tt>text</tt>. If the
+    # <tt>text</tt> is <tt>nil</tt> or contains no hashtags an empty array
+    # will be returned. The array returned will not include the leading <tt>#</tt>
+    # character.
+    #
+    # If a block is given then it will be called for each hashtag.
+    def extract_hashtags(text, &block) # :yields: hashtag_text
+      hashtags = extract_hashtags_with_indices(text).map{|h| h[:hashtag]}
+      hashtags.each(&block) if block_given?
+      hashtags
+    end
+
+    # Extracts a list of all hashtags included in the Tweet <tt>text</tt>. If the
+    # <tt>text</tt> is <tt>nil</tt> or contains no hashtags an empty array
+    # will be returned. The array returned will not include the leading <tt>#</tt>
+    # character.
+    #
+    # If a block is given then it will be called for each hashtag.
+    def extract_hashtags_with_indices(text, options = {:check_url_overlap => true}) # :yields: hashtag_text, start, end
+      return [] unless text =~ /[#＃]/
+
+      tags = []
+      text.scan(Twitter::Regex[:valid_hashtag]) do |before, hash, hash_text|
+        match_data = $~
+        start_position = match_data.char_begin(2)
+        end_position = match_data.char_end(3)
+        after = $'
+        unless after =~ Twitter::Regex[:end_hashtag_match]
+          tags << {
+            :hashtag => hash_text,
+            :indices => [start_position, end_position]
+          }
+        end
+      end
+
+      if options[:check_url_overlap]
+        # extract URLs
+        urls = extract_urls_with_indices(text)
+        unless urls.empty?
+          tags.concat(urls)
+          # remove duplicates
+          tags = remove_overlapping_entities(tags)
+          # remove URL entities
+          tags.reject!{|entity| !entity[:hashtag] }
+        end
+      end
+
+      tags.each{|tag| yield tag[:hashtag], tag[:indices].first, tag[:indices].last} if block_given?
+      tags
+    end
+
+    # Extracts a list of all cashtags included in the Tweet <tt>text</tt>. If the
+    # <tt>text</tt> is <tt>nil</tt> or contains no cashtags an empty array
+    # will be returned. The array returned will not include the leading <tt>$</tt>
+    # character.
+    #
+    # If a block is given then it will be called for each cashtag.
+    def extract_cashtags(text, &block) # :yields: cashtag_text
+      cashtags = extract_cashtags_with_indices(text).map{|h| h[:cashtag]}
+      cashtags.each(&block) if block_given?
+      cashtags
+    end
+
+    # Extracts a list of all cashtags included in the Tweet <tt>text</tt>. If the
+    # <tt>text</tt> is <tt>nil</tt> or contains no cashtags an empty array
+    # will be returned. The array returned will not include the leading <tt>$</tt>
+    # character.
+    #
+    # If a block is given then it will be called for each cashtag.
+    def extract_cashtags_with_indices(text) # :yields: cashtag_text, start, end
+      return [] unless text =~ /\$/
+
+      tags = []
+      text.scan(Twitter::Regex[:valid_cashtag]) do |before, dollar, cash_text|
+        match_data = $~
+        start_position = match_data.char_begin(2)
+        end_position = match_data.char_end(3)
+        tags << {
+          :cashtag => cash_text,
+          :indices => [start_position, end_position]
+        }
+      end
+
+      tags.each{|tag| yield tag[:cashtag], tag[:indices].first, tag[:indices].last} if block_given?
+      tags
+    end
+  end
+end

data/lib/twitter-text/hash_helper.rb

@@ -0,0 +1,21 @@
+module Twitter
+  module HashHelper
+    # Return a new hash with all keys converted to symbols, as long as
+    # they respond to +to_sym+.
+    #
+    #   { 'name' => 'Rob', 'years' => '28' }.symbolize_keys
+    #   #=> { :name => "Rob", :years => "28" }
+    def self.symbolize_keys(hash)
+      hash.dup.symbolize_keys!
+    end
+
+    # Destructively convert all keys to symbols, as long as they respond
+    # to +to_sym+. Same as +symbolize_keys+, but modifies +self+.
+    def self.symbolize_keys!(hash)
+      hash.keys.each do |key|
+        hash[(key.to_sym rescue key) || key] = hash.delete(key)
+      end
+      hash
+    end
+  end
+end

data/lib/twitter-text/hit_highlighter.rb

@@ -0,0 +1,86 @@
+module Twitter
+  # Module for doing "hit highlighting" on tweets that have been auto-linked already.
+  # Useful with the results returned from the Search API.
+  module HitHighlighter extend self
+    # Default Tag used for hit highlighting
+    DEFAULT_HIGHLIGHT_TAG = "em"
+
+    # Add <tt><em></em></tt> tags around the <tt>hits</tt> provided in the <tt>text</tt>. The
+    # <tt>hits</tt> should be an array of (start, end) index pairs, relative to the original
+    # text, before auto-linking (but the <tt>text</tt> may already be auto-linked if desired)
+    #
+    # The <tt><em></em></tt> tags can be overridden using the <tt>:tag</tt> option. For example:
+    #
+    #  irb> hit_highlight("test hit here", [[5, 8]], :tag => 'strong')
+    #  => "test <strong>hit</strong> here"
+    def hit_highlight(text, hits = [], options = {})
+      if hits.empty?
+        return text
+      end
+
+      tag_name = options[:tag] || DEFAULT_HIGHLIGHT_TAG
+      tags = ["<" + tag_name + ">", "</" + tag_name + ">"]
+
+      chunks = text.split(/[<>]/)
+
+      result = []
+      chunk_index, chunk = 0, chunks[0]
+      chunk_chars = chunk.to_s.to_char_a
+      prev_chunks_len = 0
+      chunk_cursor = 0
+      start_in_chunk = false
+      for hit, index in hits.flatten.each_with_index do
+        tag = tags[index % 2]
+
+        placed = false
+        until chunk.nil? || hit < prev_chunks_len + chunk.length do
+          result << chunk_chars[chunk_cursor..-1]
+          if start_in_chunk && hit == prev_chunks_len + chunk_chars.length
+            result << tag
+            placed = true
+          end
+
+          # correctly handle highlights that end on the final character.
+          if tag_text = chunks[chunk_index+1]
+            result << "<#{tag_text}>"
+          end
+
+          prev_chunks_len += chunk_chars.length
+          chunk_cursor = 0
+          chunk_index += 2
+          chunk = chunks[chunk_index]
+          chunk_chars = chunk.to_s.to_char_a
+          start_in_chunk = false
+        end
+
+        if !placed && !chunk.nil?
+          hit_spot = hit - prev_chunks_len
+          result << chunk_chars[chunk_cursor...hit_spot] << tag
+          chunk_cursor = hit_spot
+          if index % 2 == 0
+            start_in_chunk = true
+          else
+            start_in_chunk = false
+          end
+          placed = true
+        end
+
+        # ultimate fallback, hits that run off the end get a closing tag
+        if !placed
+          result << tag
+        end
+      end
+
+      if chunk
+        if chunk_cursor < chunk_chars.length
+          result << chunk_chars[chunk_cursor..-1]
+        end
+        (chunk_index+1).upto(chunks.length-1).each do |index|
+          result << (index.even? ? chunks[index] : "<#{chunks[index]}>")
+        end
+      end
+
+      result.flatten.join
+    end
+  end
+end