To what extent can advisors edit our writing?
Do know that there are a lot of moving parts in this dynamic: your and your advisor's personality, interaction, language proficiency, etc. factor in as well.
In term of convention on how much and advisor edit the student's work, there is none. Some like to be more hands on, treating it as a collaboration; some like to be more hands off, treating it as a consultation. Some may render more help (justified or not) if the thesis language is not the student's primary language.
Instead of "how can I change the advisor," I'd suggest focusing on how to "get the most learning out of this situation." Not all good researchers are good language teachers, they could be writing well and yet unable to explain why they wrote that way. So, some active deciphering would be needed.
First, put all emotion aside (I know, having your writing mutilated does not feel good, but still put that aside) and try to figure out the motive of each edit. Usually, they fall into a few major categories:
Grammatical: tenses, plural/singular, etc.
Clarity: unaccounted pronouns, run-on sentences that detach the subject and the action, etc.
Contextual: the language was right, but the concept was wrong. Maybe due to misinterpretation of some results, derailing the discussion.
Stylistic: the work makes sense overall, but not on par with professional convention. For example, using future/present tense in Methods rather than past tense. Using APA format instead of Vancouver, etc.
Idiosyncratic: these are more like individual preferences that are either "my mentor's mentor's mentor did it that way, so I do it the same way," pet peeves, or habit stemming from misinformed grammatical rules.
Based on the prevalence of each type, you can decide what improvements may be called for. Grammar, clarity, and style are often well-documented in textbook and professional manuals.
For idiosyncratic types, I usually let those stay as I believe there are more than one way to say the right thing, just some people prefer their "righter" way. But of course, if the changes are harmful, feel free to discuss or edit it to avoid that issue.
If you found most of them are contextual. Then it may be better to change the writing submission protocol. For example, before sending in the full text, generate a writing outline. The outline should include bullets of your interpretation in short sentences, and the bullets should be grouped by paragraph. This would help the advisor spot any misunderstanding earlier, and the list would also prime the advisor about your layout. This could temper their surprise when they see the full text not laid out as how they envisioned.
If this analysis gets too difficult, prepare an original version and a clean edited version from your advisor, and consult a professional editor who is experienced in your field. Basically, ask for an analysis on the two piece and how can your version be improved.
It's your advisor's job to prepare you for a professional career. That should include helping you learn to write in the style and up to the standards of your discipline.
Your thesis is a good place to learn to write better - even though it is "just a thesis". It is still yours. Your advisor's writing critique improves this piece of work and future work of yours.
When I edit a student's thesis, my intention is to make suggestions, not demands, even when the editing was done by writing my comments in red all over a printed copy of the student's draft. Of course, some of my comments are corrections, which a student would obviously accept, but some are matters of judgment, and a student might well disagree. I'd discuss the disagreement with the student, and until now (28 Ph.D. students so far) we've always reached an agreement. In at least one case, the final version of the thesis contained a proof which, though correct, was so far from optimal (in my opinion) that I wrote down a better proof for my own records.
In principle, there could be a real problem if the student absolutely insists on doing things one way (because it is, after all, the student's thesis) while the adviser absolutely insists on another way (and won't approve the thesis if it's done the student's way). Fortunately, I've never heard of such a situation in real life.