When coding in Rust it's common to have chaining functions like a().b().c().... A single function could contain more than 50~60 LOC (more than one screen on my laptop) with several these code blocks after formatting by rustfmt. (Sometimes more than 100 LOC.) Long function body seems hard to read.
Do you wrap chaining function like this ONLY to write less code in the call side?
Whether or not the implementation is using chaining specifically isn't something I think about. If the function is useful, make it, otherwise don't. Functions vs methods are, in general, pretty orthogonal to that.
I agree with @erelde ; if I think a particular chain would be common, or if I want to describe specifically what it is doing, I would at least consider putting it in a function. But I wouldn't do it for aesthetic/readability reasons; if the chain is too long I assume rustfmt will put it on separate lines to make it readable.
rustfmt is supposed to be the standard for all Rust code, but the decisions it made are far from unanimous. Deviation from the default formatting is often discouraged, but in its current state, authors cannot always be blamed for it. When in doubt, use your own judgment as well as machine-made ones to improve on the code quality.