Rewrite zip extraction code to use Buffer in upload.socketRoutes.ts

[greenscientist]

The code in packages/transition-backend/src/api/upload.socketRoutes.ts extract the zip file by read all the content into a variable in memory and then writing the content in a file.
We should be using buffers in this case, to not reach the variable size limit and save some memory.
We could also switch to a package which does it completely transparently.

[greenscientist]

Random blog post: https://blog.logrocket.com/best-methods-unzipping-files-node-js/

[davidmurray]

@greenscientist What if we instead just used the standard zip and unzip unix commands? All we are really doing in the code is zipping a directory or unzipping an archive to a directory. Nothing that can't be done with a one line zip or unzip call.
Plus, we would drop the dependency on a library, JSZip.
What do you think? cc @tahini @kaligrafy

[greenscientist]

@davidmurray, we usually don't know what is installed on the system. Also, the code can run on any OSes. We sould need to add this as a requirement in the installation.
the zip algoritgm is simple enough that it's easy to use in via a library.

[davidmurray]

good point
I'll rewrite the code to use https://www.npmjs.com/package/node-stream-zip then