The importwith my local changes took over 5h to complete. Restarted migeld and ch.oddb services
Looking againg for products:
Search via bin/admin for the root case:
migel> Migel::Model::Product.all.find_all{|x| x.pharmacode.eql? '2244775'}.collect{|x| [x.odba_id, x.pharmacode, x.ean_code, x.migel_code]}
-> [[32437, "2244775", "4046445600104", "05.04.05.00.1"], [3081, "2244775", "4046445107702", "05.04.03.00.1"]]
migel> Migel::Model::Migelid.find_by_migel_code('05.04.03.00.1').products.find_all{|x| x.pharmacode.eql? '2244775'}
-> []
migel> Migel::Model::Migelid.find_by_migel_code('05.04.05.00.1').products.find_all{|x| x.pharmacode.eql? '2244775'}
-> [#<ODBA::Stub:27666240#32437 @odba_class=Migel::Model::Product @odba_container=27482100#3287>]
The old wrong product (odba_id 3081) is still laying around. Corrected the delete loop to find delete this kind of products, too. Improved logging of deleted items to show odba_id, pharmacode, ean_code and migel_code. Reloading the database. Restarting the import. Will check the result once the import has finished.
My preliminary patch see: Attach:delete-bauerfeind.2.txt
The main idea is simple:
I could see in the importer.rb how Masa handled the case of a product changing its migel_code, as each product belongs to a migelid (as seen in lib/migel/model/product.rb). Also I could not find in the spec test any test case where the migelid of a product is changed. I looked at spec/model/product_spec.rb and spec/util/importer_spec.rb.
The import finished without a problem. After restarting migeld and ch.oddb all searches work as expected! Pushed commit Delete all Bauerfeind products before importing
I must adapt the VCR download to return a zipped transfer.dat with the same content as the unzipped file before.
Took me some time to create a new simple zip file for the transfer.dat. Done with the following snippet which reads in a manually zipped transfer.zip
c.before_record(:zurrose) do |i|
if /pillbox/i.match(i.request.uri)
puts "#{Time.now}: #{__LINE__}: URI was #{i.request.uri}"
tmp_zip = File.join(Oddb2xml::SpecData, 'vcr', 'transfer.zip')
i.response.body = IO.binread(tmp_zip)
i.response.headers['Content-Length'] = i.response.body.size
end
end
Also returning empty XLSX files when opening an Person/Betrieb XLSX file fails.
Must investigate failures like
Oddb2xml::Builder when -o for fachinfo is given when -f dat is given should contain the correct values fo CMUT from transfer.dat
Failure/Error: expect(line[60..65]).to eq '000495'
expected: "000495"
got: "000496"
Pushed commit Fixed spec tests. Workaroung for XLSX problem
Creating a test for the following https://mail.google.com/mail/u/3/#inbox/156b030e0bf93119. Must test
Made test for splitting the e-mail. This works fine. Working on reading the articles (plain ascii) and the custormer (ISO-8859-1).
Adding the test for customers showed that I did not strip a leading space when importing from the CSV file. Nice!
Pushed commits:
Working on testing an incoming order. As these mail get dispatched and markes as read, I am unable to retrieve the original mail message as I did for the customer. I see that the TestPopMission has many skip. I think the best way is to convert bbmb to use the gem mail (instead of rmail) and convert all tests accordingly. I think the PopMission is only used by Vetoquinol.