Fix problem with dese scraper lumping in 2021-22 data as 2022-23 data.

Deleted unused csvs.  Turned off puts statements in admin loader.
Remove old, now unused admin data loader class.
mciea-main
rebuilt 3 years ago
parent c42b569e78
commit f1022728fa

@ -7,7 +7,7 @@ module Dese
include Dese::Enrollments
attr_reader :filepaths
def initialize(filepaths: [Rails.root.join('data', 'admin_data', 'dese', '5D_2_enrollments.csv'),
def initialize(filepaths: [Rails.root.join('data', 'admin_data', 'dese', 'enrollments.csv'),
Rails.root.join('data', 'admin_data', 'dese', '5D_2_age_staffing.csv')])
@filepaths = filepaths
end

@ -1,20 +1,24 @@
require 'csv'
require "csv"
module Dese
class Loader
def self.load_data(filepath:)
admin_data_values = []
CSV.parse(File.read(filepath), headers: true) do |row|
score = likert_score(row:)
unless valid_likert_score(likert_score: score)
school = School.find_by_dese_id(row['DESE ID']) || School.new(name: 'Non-MCIEA school',
dese_id: row['DESE ID'])
puts "Invalid score: #{score}
for school: #{school.name}
admin data item #{admin_data_item(row:)} "
# school = School.find_by_dese_id(row['DESE ID']) || School.new(name: 'School not in consortium',
# dese_id: row['DESE ID'])
# puts "Invalid score: #{score}
# for school: #{school.name}
# admin data item #{admin_data_item(row:)} "
next
end
create_admin_data_value(row:, score:)
admin_data_values << create_admin_data_value(row:, score:)
end
AdminDataValue.import(admin_data_values.flatten.compact, batch_size: 1_000, on_duplicate_key_update: :all)
end
private
@ -24,7 +28,7 @@ module Dese
end
def self.likert_score(row:)
likert_score = (row['Likert Score'] || row['LikertScore'] || row['Likert_Score']).to_f
likert_score = (row["Likert Score"] || row["LikertScore"] || row["Likert_Score"]).to_f
round_up_to_one(likert_score:)
end
@ -34,15 +38,15 @@ module Dese
end
def self.ay(row:)
row['Academic Year'] || row['AcademicYear']
row["Academic Year"] || row["AcademicYear"]
end
def self.dese_id(row:)
row['DESE ID'] || row['Dese ID'] || row['Dese Id']
row["DESE ID"] || row["Dese ID"] || row["Dese Id"]
end
def self.admin_data_item(row:)
row['Admin Data Item'] || row['Item ID'] || row['Item Id']
row["Admin Data Item"] || row["Item ID"] || row["Item Id"]
end
def self.create_admin_data_value(row:, score:)
@ -53,13 +57,14 @@ module Dese
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))
school:,
admin_data_item: AdminDataItem.find_by_admin_data_item_id(admin_data_item_id))
if admin_data_value.present?
admin_data_value.likert_score = score
admin_data_value.save
nil
else
AdminDataValue.create!(
AdminDataValue.new(
likert_score: score,
academic_year: AcademicYear.find_by_range(ay(row:)),
school:,

@ -35,7 +35,7 @@ module Dese
browser.goto(url)
selectors.each do |key, value|
next unless browser.option(text: value).present?
return unless browser.option(text: value).present?
browser.select(id: key).select(text: value)
end

File diff suppressed because it is too large Load Diff

File diff suppressed because it is too large Load Diff
Loading…
Cancel
Save