taxonomy: improve taxonomy for products available in Croatia

[benbenben2]

updated taxonomies for products found in Croatia (include Serbian, Bosnian, Slovenian, German, etc. languages)

[codecov-commenter]

Codecov Report

Merging #8140 (19f79e8) into main (70a4926) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main    #8140   +/-   ##
=======================================
  Coverage   45.56%   45.56%           
=======================================
  Files         102      102           
  Lines       20279    20279           
  Branches     4632     4632           
=======================================
  Hits         9240     9240           
  Misses       9942     9942           
  Partials     1097     1097           

Impacted Files	Coverage Δ
lib/ProductOpener/Ingredients.pm	90.66% <ø> (ø)
tests/unit/additives.t	97.91% <ø> (ø)

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[benbenben2]

Why did you change this test? @stephanegigandet

• Long answer:

After adding so many entries in the taxonomies I encountered couple of errors regarding duplicated entries (for example, an entry in additives that was already in minerals).
Then, I worked on solving duplicated entries that were listed in perl unit tests (GitHub workflow).

In the logs there were also duplicates for entries not related to the changes I did (some fr and fi entries, mainly. I remember for vitamin c for these two languages that are in both additives and in vitamins)
And that was a challenge because some of those were also used in unit tests.
I tried to update unit tests according to the changes that I did to remove duplicates.
At some point there were failures on both unit test and perl unit test github workflows, or when it was not on one workflow it was failing on the other one. Since I could not find out a proper way to fix those,(1)(2) I rolled back the changes for entries not related to the changes I did.

Phosphate de calcium was in minerals and ingredients and that is good that you noticed that I forgot these two.

(1) could it be that one of the workflow (unit test) is related to the taxonomies files whereas the other workflow (perl unit test) is related to something else like ingredients.all ?
(2) why the duplicates that I created resulted into workflow failure whereas other duplicates (which appears in the logs at each pull request: checks, pull requests, perl unit tests, raw logs) are not resulting into failure?

- Short answer:

I messed up. I just fixed that.

- Rolled back changes

In any case, I rolled back changes as follow:

• removed comment In minerals:
  replaced:

 # commented because of duplicate with en:phosphate in ingredients.txt
# synonyms:fr:orthophosphate, phosphate

by:

synonyms:fr:orthophosphate, phosphate
- put back the test as it was before.

• removed comment In additives.t:
  replaced:
  ingredients_text => "huile de colza, phosphates de calcium, carbonate de calcium, citrates de potassium"
  by:
  ingredients_text => "huile de colza, orthophosphates de calcium, carbonate de calcium, citrates de potassium"

I guess other duplicates entries not related to the changes I did could be addressed in different pull request.
Same goes for schema description.

[sonarqubecloud]

Kudos, SonarCloud Quality Gate passed!   

0 Bugs
0 Vulnerabilities
0 Security Hotspots
0 Code Smells

No Coverage information
No Duplication information

[stephanegigandet]

@benbenben2 So for ingredients, it is expected to have duplicates, as some ingredients can be both minerals and additives for instance. When we concatenate them into ingredients.all, there will be duplicates. But we ignore duplicates for ingredients, in the sense that they won't make the tests fail.

[benbenben2]

I don't know, I misunderstood the logs after the failed tests. If you have time you can have a look at it (you can see the logs by scrolling here above, between "Merge branch 'main' into taxonomy_HR_ingredients_20230221" and "a69971f", if you click on the red cross, then "Details" next to "Pull Requests / Perl unit tests (pull_request)") and tell me what was wrong because I am a bit curious and would like to understand. If you don't have time let's forget about it.

Anyway, I applied changes since Yesterday. Let me know if something else need to be changed.

[stephanegigandet]

I don't know, I misunderstood the logs after the failed tests. If you have time you can have a look at it (you can see the logs by scrolling here above, between "Merge branch 'main' into taxonomy_HR_ingredients_20230221" and "a69971f", if you click on the red cross, then "Details" next to "Pull Requests / Perl unit tests (pull_request)") and tell me what was wrong because I am a bit curious and would like to understand. If you don't have time let's forget about it.

For this test, the error that makes the test fail is this one:

Errors in the ingredients_processing taxonomy definition:
ERROR - hr:pržene already is associated to en:fried - hr:pržene cannot be mapped to entry en:roasted
Errors in the ingredients_processing taxonomy definition at /opt/product-opener/lib/ProductOpener/Tags.pm line 924.

The corresponding code in Tags.pm explains a bit what happens:

		if ($errors ne "") {

			print STDERR "Errors in the $tagtype taxonomy definition:\n";
			print STDERR $errors;
			# Disable die for the ingredients taxonomy that is merged with additives, minerals etc.
			# Temporarily (hopefully) disable die for the categories taxonomy, to give time to fix issues
			unless (($tagtype eq "ingredients") or ($tagtype eq "categories")) {
				die("Errors in the $tagtype taxonomy definition");
			}
		}

[stephanegigandet]

Looks good, thank you!