Unless you're doing something
blatantly wrong or have a very specific disorder like coeliac, diet just doesn't have very much influence on health. There are a very wide range of diets that are more-or-less equally healthy, within a margin of error. Humans are highly adaptable omnivores that have evolved to survive and thrive on a broader range of foods than pretty much any other species. The data seems so mixed because the effect sizes of reasonable interventions are so small - a tiny signal drowned out by noise.
The entire problem is that most people in high- and middle-income countries are in fact doing something blatantly wrong - they are consistently eating vastly more calories than they use. Some of those people are ignorant of what 2000 to 2500 calories actually looks like, some are deluded, but a very large proportion know damned well that they're eating far too much and do it anyway.
The obesogenic environment that we now live in is partly due to the influence of the processed foods industry, but in large part it's simply a product of abundance. Before the late 20th century, it was simply inconceivable that poor people could afford to become morbidly obese. Agricultural productivity has improved beyond all recognition and the world is flooded with incredibly cheap food of all kinds.
We've spent the last few decades trying to push back against that with all manner of initiatives intended to endgender behavioural change, with very little success. It doesn't really matter what guidance we give people when they have shown a consistent inability or unwillingness to follow it.
If we're actually serious about the effects of diet on public health, I think there are only two credible options - extremely heavy-handed regulation, or the mass prescribing of GLP-1 receptor agonists. All of the other options are just permutations of "let's do more of the thing that hasn't worked".