More intelligent removing of language variants

[pavel-karatsiuba]

Right now with the parameter --withoutLanguageVariants script is not created zim for languages with variants. This is not good because with this parameter we will never see for example simulations for the Chinese language.

For languages with variants, I propose to use language with the biggest count of simulations. For example, in the result, we have 3 Chinese zim files: Chinese (Hong Kong) contains 33 simulations, Chinese (Simplified) contains 221 simulations and Chinese (Traditional) contains 223 simulations.
My proposition is not to remove all 3 but at least leave Chinese (Traditional) which contains the biggest count of simulations and rename it to cn without language variants addition.
The next list is shown languages with variants and count of simulations. Also, a list containing the same language but without variants. So which parameter all languages with arguments will be removed.

"ar": {
    "name": "Arabic",
    "count": 194
},
"ar_MA": {  
    "name": "Arabic (Morocco)",
    "count": 96
},
"ar_SA": {
    "name": "Arabic (Saudi Arabia)",
    "count": 119
},
"zh_HK": {
    "name": "Chinese (Hong Kong)",
    "count": 33
},
"zh_CN": {
    "name": "Chinese (Simplified)",
    "count": 221
},
"zh_TW": {
    "name": "Chinese (Traditional)",
    "count": 223
},
"en": {
    "name": "English",
    "count": 224
},
"en_CA": {
    "name": "English (Canada)",
    "count": 1
},
"en_GB": {
    "name": "English (United Kingdom)",
    "count": 1
},
"ku": {
    "name": "Kurdish",
    "count": 41
},
"ku_TR": {
    "name": "Kurdish (Turkey)",
    "count": 63
},
"pt": {
    "name": "Portuguese",
    "count": 170
},
"pt_BR": {
    "name": "Portuguese (Brazil)",
    "count": 224
},
"es": {
    "name": "Spanish",
    "count": 224
},
"es_CO": {
    "name": "Spanish (Colombia)",
    "count": 17
},
"es_MX": {
    "name": "Spanish (Mexico)",
    "count": 208
},
"es_PE": {
    "name": "Spanish (Peru)",
    "count": 224
},
"es_ES": {
    "name": "Spanish (Spain)",
    "count": 48
},`

[kelson42]

@pavel-karatsiuba Agree. But please harcode this (basically zh_CN -> zh), so if for some reason we want to not follow that rule for a specific language, then we can easily change it.