The question stemmed from that thread, actually, where you suggested to follow what the dockerfile was doing.
I took a peek inside and it looked to me like a series of commands which could be translated to a bash script for bare metal installation, hence my question on converting it.
My current situation is that I installed pihole a long time ago via yvelon / pi-hole · GitLab but with the last update it half-borked the installation because the FTL and "main" pihole stayed at the "custom" version and the web part updated at v6.
My fix was to manually download the official v6 FTL because I figured that now that Docker image is using Alpine, the newer FTL builds are probably compatible "out of the box", which was the case.
I then had to redo part of the configuration via the web interface and it is now working.
The only thing that I don't really like much is that now I have the "base" pihole stuck at commit 0c6b70fa of yvelon / pi-hole · GitLab and the web interface and FTL at the latest "official" v6 and everytime I update I have to manually update the FTL by downloading it and replacing it (which I can definitely script it when running pihole -up, but still).
I don't know what yvelon has done with his pihole fork, but now that "official" FTL and webUI are working out of the box with Alpine, how hard would it be to make it completely compatible?