Memoize admin data loader

main-multiple-administrations-#186762859
Gabe Farrell 2 years ago
parent f0024d019d
commit 04d61c0b1a

@ -1,7 +1,9 @@
module Dese
class Loader
@memo = Hash.new
def self.load_data(filepath:)
admin_data_values = []
@memo = Hash.new
CSV.parse(File.read(filepath), headers: true) do |row|
score = likert_score(row:)
next unless valid_likert_score(likert_score: score)
@ -35,16 +37,31 @@ module Dese
row["Admin Data Item"] || row["Item ID"] || row["Item Id"] || row["Item ID"]
end
# these three methods do the memoization
def self.find_school(dese_id:)
return @memo["school"+dese_id] if @memo.key? "school"+dese_id
@memo["school"+dese_id] ||= School.find_by_dese_id(dese_id.to_i)
end
def self.find_admin_data_item(admin_data_item_id:)
return @memo["admin"+admin_data_item_id] if @memo.key? "admin"+admin_data_item_id
@memo["admin"+admin_data_item_id] ||= AdminDataItem.find_by_admin_data_item_id(admin_data_item_id)
end
def self.find_ay(ay:)
return @memo["year"+ay] if @memo.key? "year"+ay
@memo["year"+ay] ||= AcademicYear.find_by_range(ay)
end
def self.create_admin_data_value(row:, score:)
school = School.find_by_dese_id(dese_id(row:).to_i)
school = find_school(dese_id: dese_id(row:))
admin_data_item_id = admin_data_item(row:)
admin_data_item = find_admin_data_item(admin_data_item_id:)
academic_year = find_ay(ay: ay(row:))
return if school.nil?
return if admin_data_item_id.nil? || admin_data_item_id.blank?
admin_data_value = AdminDataValue.find_by(academic_year: AcademicYear.find_by_range(ay(row:)),
school:,
admin_data_item: AdminDataItem.find_by_admin_data_item_id(admin_data_item_id))
admin_data_value = AdminDataValue.find_by(academic_year:, school:, admin_data_item:)
if admin_data_value.present?
admin_data_value.likert_score = score
admin_data_value.save
@ -52,9 +69,9 @@ module Dese
else
AdminDataValue.new(
likert_score: score,
academic_year: AcademicYear.find_by_range(ay(row:)),
academic_year:,
school:,
admin_data_item: AdminDataItem.find_by_admin_data_item_id(admin_data_item(row:))
admin_data_item:,
)
end
end

@ -28,15 +28,14 @@ namespace :data do
desc "load admin_data"
task load_admin_data: :environment do
original_count = AdminDataValue.count
Dir.glob(Rails.root.join("data", "admin_data", "dese", "*.csv")).each do |filepath|
puts "=====================> Loading data from csv at path: #{filepath}"
Dese::Loader.load_data filepath:
end
Dir.glob(Rails.root.join("data", "admin_data", "out_of_state", "*.csv")).each do |filepath|
jobs = Queue.new
Dir.glob(Rails.root.join("data", "admin_data", "dese", "*.csv")).each { |filepath| jobs << filepath }
Dir.glob(Rails.root.join("data", "admin_data", "out_of_state", "*.csv")).each { |filepath| jobs << filepath }
while filepath = jobs.pop(true)
puts "=====================> Loading data from csv at path: #{filepath}"
Dese::Loader.load_data filepath:
end
rescue ThreadError
puts "=====================> Completed loading #{AdminDataValue.count - original_count} admin data values"
end

Loading…
Cancel
Save