What are __slots__ in Python? Optimize Your Code!

Key Takeaways

what is going on guys welcome back in this video today we're going to learn about slots in Python and how we can use them to reduce memory usage into to speed up our code so let us get right into [Music] ited all right so we're going to learn how to optimize our python code using slots and Slots is basically an attribute that we can Define in classes to explicitly state which attributes we expect our instances to have and this leads to faster exess time into to less memory usage uh for this we're going to go right into an example now one thing that we need to install first is a package that I introduced in the last video which is memory profiler just so we can track how much memory is actually being used for this you can use pip or pip 3 to install memory unor profiler this is just a package that allows us to track how much memory is being used in specific functions so what we're going to do now first is we're going to use one example we're going to code one example without using slots we're going to measure the memory usage and the time and then we're going to do the same thing with slots to see what changes so we're going to start here by importing statistics so that we can have multiple runs that we can take the mean from so that we have a more representative number um and then we're going to also say from memory profiler import memory usage and we're also going to import time it so time it and statistics are core python packages this is the one that we installed and now now I'm going to just Define a simple class called Point without slots and this class of course will not use slots we're just going to use basic python we're going to Define in a nit method we're going to have the parameters X and Y here in the Constructor and we're going to say self.x equals x self.y equal y um and that is basically now our class so by default we don't have a slots attribute so we just have have these attributes implicitly defined here so we say there is self.x self.y and so on and then we can go ahead and say p equals 1020 for example actually Point without slots print p.x p.y and you can see I get the values here now what this already creates in the background is an attribute called uncore uncore dict uncore uncore and this dictionary basically contains the variable names or the attribute names and the respective values this is what you get without using slots now to see what memory usage and performance we have here or speed we have here we're going to define a function create uh points and this function will basically just return uh a list comprehension so a list based on a list comprehension we're going to say points without slots and I I for I in range and we're just going to Define find uh a long range here let me just see what did I use here I used 1 million so I'm going to use 1 million here again uh and this is what we're going to track now so we create 1 million instances of these points here and we're going to do that 10 times so we're going to say m values equals an empty list time values equals an empty list and then we're going to say four placeholder in range 10 times what I want to do is I want to say the M usage without points uh is going to be equal to memory usage this is now from memory profiler uh and this function here we're going to pass the create points function uh which is going to be passed without calling it so we're passing it as an entity like this without the parenthesis uh and we're going to define the interval to be 0.01 and then basically we're going to say m values append and I only want to append the max value so max meem usage without points um I think we can also do that here with uh Max usage equals true but let's do it like this now uh just to be safe so we append this memory usage of this particular run to M values and now we're going to uh do the same thing with time so time without slots is going to be equal to time it time it and then Lambda um Point actually let's create a point here I'm going to say point is equal to point without slots one two and we're going to say point x point Y this is now the excess time so we're not actually tracking the time for the creation here we're using the or we're tracking the memory usage for the creation but we're tracking the excess time here so how how long does it take to access X and Y here uh and we want to do that a million times as well like this um and then we say again time values append time without slots and in the end what we want to print here is we want to print the mean so statistics mean of M values statistics mean of time values that is basically it so now let's run this let's see if I messed up anything or if we're going to get a number here and we should get the average um the average memory usage here so 235 and here we get 0.06 seconds it's quite fast uh but yeah this is the result now of this um benchmarking you could say now let's go ahead and copy this and create a second file let's call this Main 2py and here now we're going to do it with slot so we're going to say point with slots and what we're going to do here is we're going to define the following attribute the following class attribute uncore uncore slots uncore uncore equals and now basically here we pass a list or we Define a list of all the fields that we're going to use x and and you can already see here I defined X and Slots now the error message is gone here but I don't have Y which is why it says here points with slots object attribute Y is read only so I have to actually put it here to be able to set y to something so this is how it works now what you will see here and I'm going to comment this out for a second is when I go ahead and I say p equals point with slots 1020 and I try to print P doore uncore dictionary uncore uncore you can see that it doesn't have this attribute so basically using slots uh forbids this dictionary to be created this is how it basically say saves uh the access time and the the memory usage because it forbids or it denies dictionary and also weak references here uh to be to be used when we store values we Define them explicitly here the attributes in the slots class attribute so what we can do now is we can basically uncomment all this we can change this to with slots and we can change this to with slots everything else should stay the same and then we should be able to see a difference here so I'm going to run this and we should hopefully see that we're using less memory and the access time will probably be roughly the same so you can see first of all maybe I can run this here uh can I not run this at the same time how do I do that in py charm I'm not sure but you saw we had something like 120 something and here we had 230 something so the access time is let's just copy this values here these values here um like this and let's run this to compare the values here I think the access time is not that different in this case but the memory usage is definitely different you can see here it's a little bit faster but this could also just be some random fluctuation but this one here is definitely lower than this one so we use way less memory for the same functionality we don't lose any functionality the only thing is of course now we cannot dynamically just introduce new variables or new attributes uh if they're not defined in slots so that is of course a downside we can maybe increase the number here to 10 million to see if this changes something I mean of course it's going to change the run time but uh maybe we can see more of a difference here so maybe it now takes half a second to do that there you go so this is the value that we have for 10 million here and now let's run this with 10 million and maybe we're going to see some larger difference now here so this was the first value with slots and now hopefully we see something that's slower but we could also see something that's faster I'm not sure if that makes so much difference here there are of course scenarios in which this makes a lot of difference yeah you can see it actually takes some some longer if I increase this now to 100 million and to um 1 billion you would see even more so all these uh small differences here actually would or these here all these small differences here would actually make more of a difference as we go to larger and larger numbers but yeah this is how you can optimize your python code and your python classes using slots so that's it for today's video I hope you enjoyed it and hope you'll learned something if so let me know by hitting a like button and leaving a comment in the comment section down below and of course don't forget to subscribe to this Channel and hit the notification Bell to not miss a single future video for free other than that thank you much for watching see you in the next video and bye