Refactor the way the list of quran translations is generated

[assem-ch]

Current behavior, the translations file should be always there since its main resource.. after indexing translations, we run update_translations_indexed_list to get all translations from index and store them in the translation file. current problems involve that the file is essential and also should be tracked , and the operation of listing is very slow.

Right behavior, no file should be by default, if no file the list should be considered as empty dictionary. Translations should be added in the indexing phase not after indexing:

• Index translation x,
• If success, add translation x to translations.json
• First translation indexed should create the file