The C++11 standard n3337 speaks of translation units (it is the view, by the compiler, of a source file, including preprocessing).
I don't think that it would be good idea to make a class of 50 functions and include that class even for calling a single function. But I wanted to be sure.
This is your opinion. The GCC compiler is coded in C++, and has lots of such classes. It is free software (about 10 to 20 millions lines of source code) and you can study and improve its source code.
The quality of a software project is more related to good development practices (e.g. the famous Joel Test) than to the number of methods in a given class. A given class can have hundred of methods and still be very readable. Look for example inside Qt source code. Read about cyclomatic complexity of call graphs.
The RefPerSys project has lots of classes and some of them have dozen of methods and some of these methods have a hundred of mostly sequential lines.
Some C++ functions of the helpcovid project are quite long but are mostly initializing functions.
I want to know how you expert guys organize your such bunch of functions?
My opinion is that proper documentation (including generated documentation by doxygen) and good, consistent and documented naming or coding conventions is much more important than the number of methods in a given class, or the number of statements implementing a given method.
Recent PHP implementations use techniques similar to compilation to some bytecode, so the size of source code does not matter that much for runtime performance (except for startup time). Time complexity of your code matters a lot more for performance.
BTW you could consider coding your website -if the web server is running Linux- using C (e.g. with libonion and sqlite, or as a plugin to Apache), C++ (e.g. with Wt), or Ocaml (with Ocsigen), or Guile (read first SICP then this) or Haxe. Be also aware of both CGI and FastCGI techniques.
Cybersecurity aspects and ease of maintenance are much more important for public web sites than raw performance. In particular, avoid SQL injections.
My suggestion is for you to publish the source code of your website (this paper explains why). Of course, keep most of its data private (to comply with GDPR regulations), and do backup it quite often.
Christian Queinnec published a lot of academic papers and books related to your concerns. Read about continuations and their relation to the WWW (mostly but not only HTTP, also SMTP, IMAP, DNS, WebSockets, Web services, AJAX, SOAP, JSONRPC, etc...).