added draft for impressionist(...) uniqueness
This commit is contained in:
parent
09bfa2f27e
commit
e63e835936
|
@ -12,11 +12,14 @@ module ImpressionistController
|
|||
base.before_filter :impressionist_app_filter
|
||||
end
|
||||
|
||||
def impressionist(obj,message=nil)
|
||||
def impressionist(obj,message=nil,opts={})
|
||||
unless bypass
|
||||
if obj.respond_to?("impressionable?")
|
||||
obj.impressions.create(create_statement({:message => message}))
|
||||
if unique_instance?(obj, opts[:unique])
|
||||
obj.impressions.create(associative_create_statement({:message => message}))
|
||||
end
|
||||
else
|
||||
# we could create an impression anyway. for classes, too. why not?
|
||||
raise "#{obj.class.to_s} is not impressionable!"
|
||||
end
|
||||
end
|
||||
|
@ -26,17 +29,11 @@ module ImpressionistController
|
|||
@impressionist_hash = Digest::SHA2.hexdigest(Time.now.to_f.to_s+rand(10000).to_s)
|
||||
end
|
||||
|
||||
def impressionist_subapp_filter(actions=nil, unique_opts=nil)
|
||||
def impressionist_subapp_filter(actions=nil,unique_opts=nil)
|
||||
unless bypass
|
||||
actions.collect!{|a|a.to_s} unless actions.blank?
|
||||
if (actions.blank? || actions.include?(action_name)) && (unique_opts.blank? || is_unique(unique_opts))
|
||||
if (!actions.blank? && !unique_opts.blank?)
|
||||
logger.info "Restricted to actions #{actions.inspect} and uniqueness for #{unique_opts.inspect}"
|
||||
end
|
||||
Impression.create(create_statement(
|
||||
:impressionable_type => controller_name.singularize.camelize,
|
||||
:impressionable_id=> params[:id]
|
||||
))
|
||||
if (actions.blank? || actions.include?(action_name)) && unique?(unique_opts)
|
||||
Impression.create(direct_create_statement)
|
||||
end
|
||||
end
|
||||
end
|
||||
|
@ -50,50 +47,53 @@ module ImpressionistController
|
|||
Impressionist::Bots::LIST.include? request.user_agent
|
||||
end
|
||||
|
||||
def is_unique(unique_opts)
|
||||
# FIXME think about uniqueness in relation to impressionable_id, impressionable_type and controller_name
|
||||
# is controller name redundant? does the controller name always have to match?
|
||||
default_statement = create_statement(
|
||||
:impressionable_type => controller_name.singularize.camelize,
|
||||
:impressionable_id=> params[:id]
|
||||
)
|
||||
statement = unique_opts.reduce({}) do |query, param|
|
||||
query[param] = default_statement[param]
|
||||
query
|
||||
end
|
||||
#logger.debug "Statement params: #{statement.inspect}."
|
||||
# always use impressionable type?
|
||||
statement[:impressionable_type] = controller_name.singularize.camelize
|
||||
#statement[:impressionable_id] = params[:id]
|
||||
return Impression.where(statement).size == 0
|
||||
def unique_instance?(impressionable, unique_opts)
|
||||
return unique_opts.blank? || impressionable.impressions.where(unique_query(unique_ops)).size == 0
|
||||
end
|
||||
|
||||
# creates a statment hash that contains default values for creating an impression (without
|
||||
# :impressionable_type and impressionable_id as they are not needed for creating via association).
|
||||
def create_statement(query_params={})
|
||||
def unique?(unique_opts)
|
||||
return unique_opts.blank? || Impression.where(unique_query(unique_opts)).size == 0
|
||||
end
|
||||
|
||||
# creates the query to check for uniqueness
|
||||
def unique_query(unique_opts)
|
||||
full_statement = direct_create_statement
|
||||
# reduce the full statement to the params we need for the specified unique options
|
||||
unique_opts.reduce({}) do |query, param|
|
||||
query[param] = full_statement[param]
|
||||
query
|
||||
end
|
||||
end
|
||||
|
||||
# creates a statment hash that contains default values for creating an impression via an AR relation.
|
||||
def associative_create_statement(query_params={})
|
||||
query_params.reverse_merge!(
|
||||
:controller_name => controller_name,
|
||||
:action_name => action_name,
|
||||
:user_id => user_id,
|
||||
:request_hash => @impressionist_hash,
|
||||
:session_hash => session_hash,
|
||||
:ip_address => remote_ip,
|
||||
:ip_address => request.remote_ip,
|
||||
:referrer => request.referer
|
||||
)
|
||||
end
|
||||
|
||||
def session_hash
|
||||
# # careful: request.session_options[:id] encoding in rspec test was ASCII-8BIT
|
||||
# # that broke the database query for uniqueness. not sure how to solve this issue
|
||||
# # seems to depend on app setup/config
|
||||
# str = request.session_options[:id]
|
||||
# # probably this isn't a fix: request.session_options[:id].encode("ISO-8859-1")
|
||||
# logger.debug "Encoding: #{str.encoding.inspect}"
|
||||
request.session_options[:id]
|
||||
# creates a statment hash that contains default values for creating an impression.
|
||||
def direct_create_statement(query_params={})
|
||||
query_params.reverse_merge!(
|
||||
:impressionable_type => controller_name.singularize.camelize,
|
||||
:impressionable_id=> params[:id]
|
||||
)
|
||||
associative_create_statement(query_params)
|
||||
end
|
||||
|
||||
def remote_ip
|
||||
request.remote_ip
|
||||
def session_hash
|
||||
# # careful: request.session_options[:id] encoding in rspec test was ASCII-8BIT
|
||||
# # that broke the database query for uniqueness. not sure if this is a testing only issue.
|
||||
# str = request.session_options[:id]
|
||||
# logger.debug "Encoding: #{str.encoding.inspect}"
|
||||
# # request.session_options[:id].encode("ISO-8859-1")
|
||||
request.session_options[:id]
|
||||
end
|
||||
|
||||
#use both @current_user and current_user helper
|
||||
|
|
|
@ -57,7 +57,8 @@ describe PostsController do
|
|||
end
|
||||
|
||||
it "should recognize referrer uniqueness" do
|
||||
controller.stub!(:referrer).and_return("http://somehost.someurl.somdomain/some/path")
|
||||
@request.env['HTTP_REFERER'] = 'http://somehost.someurl.somdomain/some/path'
|
||||
#controller.stub!(:referer).and_return("http://somehost.someurl.somdomain/some/path")
|
||||
controller.impressionist_subapp_filter(nil, [:referrer])
|
||||
controller.impressionist_subapp_filter(nil, [:referrer])
|
||||
Impression.should have(@impression_count + 1).records
|
||||
|
|
Loading…
Reference in New Issue