I question the premise... somewhat. Yes, some patterns show up in lots of works. Yes, there can be benefit to practicing patterns out of context. But I'd focus on a small number of them. Scales and arpeggios are definitely compositional "building blocks" that can be found in many pieces in many genres. Most folks recognize that practicing scales and arpeggios, in addition to "real pieces," is a good use of time; in fact, I think few would ever recommend that you skip them. So if you're looking for a "quick regimen," I like Carl Flesch's Scale System as a sort of "daily multivitamin": For a given key, it gives plain scales, arpeggiations of various chords, scales in "broken thirds" (zig-zag pattern), chromatic scale, and then some of the same things while playing double stops, in thirds, sixths, octaves, tenths, and harmonics. (I also have some preferred ways of abbreviating this regimen, when I want to devote 5 minutes out of a 30-minute session to scales, instead of 30 minutes out of 2 hours!)
The language about "memorization" is an interesting angle in this question. If you mean that you have trouble remembering which note comes after which in an entire song, I challenge the notion that learning abstracted "building blocks" will help much. There are so many ways to put notes together that there's little overlap from one song to another—aside from scales and arpeggios—and what there is might be such small blocks ("up a fifth, down a step") that they don't help much in memorizing the overall picture. If the problem is remembering how songs go, there are better tips that are handled in other questions.
But that doesn't mean you shouldn't practice these abstractions. You mention "muscle memory"; really, it's a misleading bit of semantics to use the word "memory" both for that phenomenon and for remembering how a song goes. Neurologically, they happen in different places. I can tell you right now the plot of Hamlet, or tell you my mailing address, or walk up a flight of stairs. The plot of Hamlet is something I "remember," but I'm actively and consciously thinking to put together my words about it. My mailing address is something I've repeated so often that I can use rote memory, even the way I explain the spelling and the timing and inflection with which I say it are probably patterned and happen without active thought. But walking up stairs is different from them both: it's never involved cognition or the speech center of my brain; my vague layman's knowledge is that it comes from the brain stem. This movement isn't "memorized," at least not via the same cognitive mechanism as my mailing address; it's practiced and perfected over my first few years of life. This is what scales, arpeggios, and etudes help with, and yes, it's worth the work, but the benefit is in agility and accuracy, not retention of compositional structure.
Which is to say, yes, you should practice patterns. Note, there are "musical" patterns—concepts based on abstract theory, like "a sixth" or "a diminished triad," but there are also physical patterns. A diminished triad starting on open D string involves contracting your hand as you lower your 4th finger for the A flat, but one starting on E a whole step higher involves expanding your hand as your first finger moves from the E to the B flat, moving over one string and reaching back by a half step. The same musical concept, but completely different muscular reality.
And the best way to practice these muscular realities is by augmenting your practice of "real pieces" with etudes. There's a large body of these pieces—"studies"—written not for concert but to work on some particular focus. Besides left hand, some work on certain bow strokes, or the control of bow speed during dynamic changes, etc. Whatever your skill level, there are etudes to match, from some Wohlfahrt works for beginners, up to virtuosic studies by Wieniawski, Ysaÿe, etc. This syllabus is not a complete list, but a good starting point; if you can find pieces on it that are at your current level, you can try titles from the corresponding "Technical studies" section. Note, for many of these etude books, the version that you'd want to use today is old enough to be public domain and can be found for free on IMSLP.
