From 003b17fd44158a353ffbc89ed3c7039178da4c50 Mon Sep 17 00:00:00 2001
From: Joel McCoy <joelmccoy@gmail.com>
Date: Sat, 27 Dec 2014 18:38:02 -0500
Subject: [PATCH] Make banned_terms global, overwrite Model#valid_tweet?, first
 pass at bootstrapping

---
 boodoo.rb | 55 ++++++++++++++++++++++++++++++++++++++++++++++++++++++-
 bots.rb   | 28 +++++++++++++++++++---------
 2 files changed, 73 insertions(+), 10 deletions(-)

diff --git a/boodoo.rb b/boodoo.rb
index 7eb93d3..95c76a1 100644
--- a/boodoo.rb
+++ b/boodoo.rb
@@ -22,15 +22,40 @@ module Ebooks::Boodoo
     array_splitter ||= / *[,;]+ */
     value.split(array_splitter).map(&:strip)
   end
+
+  def make_client
+    Twitter::REST::Client.new do |config|
+      config.consumer_key = @consumer_key
+      config.consumer_secret = @consumer_secret
+      config.access_token = @access_token
+      config.access_token_secret = @access_token_secret
+    end
+  end
 end
 
+## Retweet check based on Really-Existing-RT practices
 class Ebooks::TweetMeta
   def is_retweet?
-    tweet.retweeted_status? || !!tweet.text[/^[RM]T[: ]/i]
+    tweet.retweeted_status? || !!tweet.text[/[RM]T ?[@:]/i]
+  end
+end
+
+class Ebooks::Boodoo::Model < Ebooks::Model
+  def valid_tweet?(tokens, limit)
+    tweet = NLP.reconstruct(tokens)
+    found_banned = $banned_words.any? do |word|
+      re = Regexp.new("\\b#{word}\\b", "i")
+      re.match tweet
+    end
+    tweet.length <= limit && !found_banned && !NLP.unmatched_enclosers?(tweet)
   end
 end
 
 class Ebooks::Boodoo::BoodooBot < Ebooks::Bot
+  @required_fields = ['consumer_key', 'consumer_secret',
+                      'access_token', 'access_token_secret',
+                      'bot_name', 'original']
+
   # A rough error-catch/retry for rate limit, dupe fave, server timeouts
   def catch_twitter
     begin
@@ -83,4 +108,32 @@ class Ebooks::Boodoo::BoodooBot < Ebooks::Bot
     end
   end
 
+  def has_model?
+    File.exists? @model_path
+  end
+
+  def has_archive?
+    File.exists? @archive_path
+  end
+
+  def get_archive!
+    @archive = Archive.new(@original, @archive_path, make_client).sync
+  end
+
+  def make_model!
+    log "Updating model: #{@model_path}"
+    Ebooks::Boodoo::Model.consume(@archive_path).save(@model_path)
+    log "Loading model..."
+    @model = Ebooks::Boodoo::Model.load(@model_path)
+  end
+
+  def can_run?
+    missing_fields.empty?
+  end
+
+  def missing_fields
+    @required_fields.select { |field|
+      !send(field).nil? && !send(field).empty?
+    }
+  end
 end
diff --git a/bots.rb b/bots.rb
index 1830165..c444a25 100644
--- a/bots.rb
+++ b/bots.rb
@@ -23,7 +23,7 @@ class UserInfo
 end
 
 class CloneBot < BoodooBot
-  attr_accessor :original, :model, :model_path, :auth_name
+  attr_accessor :original, :model, :model_path, :auth_name, :archive_path, :archive
   # alias_method :oauth_token, :access_token
   # alias_method :oauth_token_secret, :access_token_secret
   def configure
@@ -45,6 +45,7 @@ class CloneBot < BoodooBot
     # Array fields are CSV or SSV
     @blacklist =        parse_array(SETTINGS['BLACKLIST'])
     @banned_terms =     parse_array(SETTINGS['BANNED_TERMS'])
+    $banned_terms =     @banned_terms
     @special_terms  =   parse_array(SETTINGS['SPECIAL_TERMS'])
 
     # Fields parsed as Fixnum, Float, or Range:
@@ -63,10 +64,20 @@ class CloneBot < BoodooBot
     @attempts = 0
     @followers = []
     @following = []
+    @archive_path = "corpus/#{@original}.json"
+    @model_path = "model/#{@original}.model"
     # @have_talked = {}
 
-    # load model file
-    load_model!
+    if can_run?
+      get_archive! unless has_archive?
+      make_model! unless has_model?
+    else
+      missing_fields.each {|missing|
+        log "Can't run without #{missing}"
+      }
+      log "Heroku will automatically try again immediately or in 10 minutes..."
+      Kernel.exit(1)
+    end
   end
 
   def top100; @top100 ||= model.keywords.take(100); end
@@ -90,10 +101,11 @@ class CloneBot < BoodooBot
       follow_parity
     end
 
-    # TODO: This throws a weird error.
-    #       Probably don't need it anyway?
-    # @auth_name ||= twitter.user.screen_name
-    # log "Logged in as #{auth_name}"
+    scheduler.interval @refresh_model_interval do
+      log "Refreshing archive/model..."
+      get_archive!
+      make_model!
+    end
   end
 
   def on_direct_message(dm)
@@ -210,8 +222,6 @@ class CloneBot < BoodooBot
   end
 end
 
-
-
 CloneBot.new(SETTINGS['BOT_NAME']) do |bot|
   # CloneBot#configure does everything!
   bot