Support external layer + Delete source folder when using custom layer + Fix multiple functions + Lower redundancy

[magicyoda]

Support for external layer (e.g. Klayers)

• I added 4 possible "public_layer_X_arn" input variables.
• Those allow the script to add external / public layer to the target functions.
• I required those in order to load "Klayers" dependencies: https://github.com/keithrozario/Klayers/tree/master/deployments/python3.8/arns

Delete custom_layer source folder after layer creation

• When using a "custom_layer", the script picks a source folder and create a layer out of it.
• As the source folder has now been made available as a new customer layer, I added the deletion of this source folder in order to avoid code redundancy: "rm -rf ${INPUT_CUSTOM_LAYER_4_PATH}" on each custom_layer_X_path function.

Fix code for multiple functions

• During my test, the multiple function code was not working properly. There was no "directory" named after the function in which the script shall enter. I removed therefore the "cd "${1}"" and "cd .." on function 'publish_function()'

Lower operation redundancy

• The function "publish_function()" was finally completely removed so that the zip "code.zip" is only created and deleted once and not each time for each target function.

[magicyoda]

@seattleagainstslavery After reviewing your web scraper code, I realise that this Pull Request would break it because of how your "multiple function" structure was designed. It is therefore not recommended to perform this full pull request. Only first and second point could be interesting for you.