WEBVTT 00:00:00.001 --> 00:00:05.020 Hello and welcome to Python Bytes, where we deliver Python news and headlines directly to your earbuds. 00:00:05.020 --> 00:00:10.020 This is episode 259, recorded November 17th, 2021. 00:00:10.020 --> 00:00:11.400 And I'm Brian Okken. 00:00:11.400 --> 00:00:12.500 I'm Michael Kennedy. 00:00:12.500 --> 00:00:13.680 And I'm Renee Teet. 00:00:13.680 --> 00:00:15.620 Well, thanks, Renee, for joining us today. 00:00:15.620 --> 00:00:17.820 Can you tell us a little bit about who you are? 00:00:17.820 --> 00:00:20.980 I'm the director of data science at Helio Campus. 00:00:20.980 --> 00:00:26.160 And a lot of people know me as Data Science Renee or Becoming Data Sci on Twitter. 00:00:26.160 --> 00:00:29.560 So that's where a lot of people follow me. 00:00:29.760 --> 00:00:35.740 And then I started with, I had a podcast that's not actively recording, but it's called Becoming a Data Scientist Podcast. 00:00:35.740 --> 00:00:39.820 So some people listening probably know me from that as well. 00:00:39.820 --> 00:00:40.460 Cool. 00:00:40.460 --> 00:00:41.300 Yeah, awesome. 00:00:41.300 --> 00:00:42.680 You were doing a bunch of cool stuff there. 00:00:42.680 --> 00:00:45.360 And any chances of maybe going back to podcasting? 00:00:45.360 --> 00:00:47.100 It's definitely still open. 00:00:47.100 --> 00:00:50.380 I've always told myself this is a pause, not a stop. 00:00:50.380 --> 00:00:51.680 It's just an extended pause. 00:00:51.680 --> 00:00:53.400 So yes, hopefully I will get back to it. 00:00:53.400 --> 00:00:54.740 It's hard to keep going, isn't it? 00:00:54.740 --> 00:00:57.900 I mean, life gets in the way and then you get busy. 00:00:58.320 --> 00:01:02.980 I'm always so impressed with those of you that have hundreds of episodes very consistently recorded. 00:01:02.980 --> 00:01:05.240 Well, Brian makes me show up every week. 00:01:05.240 --> 00:01:09.540 So yeah, it definitely helps having a partner so that you can coerce each other in. 00:01:09.540 --> 00:01:10.200 That's right. 00:01:10.740 --> 00:01:14.300 Well, Michael, speaking of partners, want to tell us about something? 00:01:14.300 --> 00:01:16.060 Let's talk about some changes. 00:01:16.060 --> 00:01:17.920 Some PyPI changes. 00:01:17.920 --> 00:01:19.960 These come to us from Brian Skin. 00:01:19.960 --> 00:01:22.100 Thank you, Brian, for shooting this over. 00:01:22.540 --> 00:01:26.020 And it's a project by Bernay Gabor here. 00:01:26.020 --> 00:01:36.180 And if we pull this up, it says, have you ever wondered when did your Python packages, the packages in your environment that you have active or any given environment, how old are they? 00:01:36.180 --> 00:01:38.160 When were they last updated? 00:01:38.160 --> 00:01:40.020 Is there a version of them that's out of date? 00:01:40.020 --> 00:01:51.500 So I've been solving this by just forcing them to update using pip compile and the pip-tools stuff to just regenerate and reinstall the requirements files. 00:01:51.680 --> 00:01:55.740 But this is a way to just ask the question, hey, what's the status? 00:01:55.740 --> 00:02:00.200 And it wouldn't be an episode if we didn't somehow feature Will McGugan. 00:02:00.200 --> 00:02:03.920 So this is based on Rich, of course. 00:02:03.920 --> 00:02:05.980 So let's go check this thing out. 00:02:05.980 --> 00:02:11.900 So over here, if we go to the homepage, we get, as all projects should, a nice animation here. 00:02:11.900 --> 00:02:18.800 And if you look at it, you can just see type IPI dash changes and you specify the path to a Python in a virtual environment. 00:02:18.800 --> 00:02:23.200 So in this example, it's like PyPI changes, VNV slash bin slash Python. 00:02:23.200 --> 00:02:27.520 It does some thinking on the internet, caches some information about the packages. 00:02:27.520 --> 00:02:30.440 And it says, all right, you've got all these things installed. 00:02:30.440 --> 00:02:31.300 They're this version. 00:02:31.300 --> 00:02:35.620 Some of it will just say this was updated 10 months ago or a year and three days ago. 00:02:35.620 --> 00:02:41.220 Others, it'll say it was updated a year ago, but only six months on the latest version. 00:02:41.220 --> 00:02:42.580 It says remote such and such. 00:02:42.580 --> 00:02:44.460 That's the one you could install if you were to update it. 00:02:44.460 --> 00:02:55.960 So it's a real nice way to see, well, which ones are here that could be updated or even also sometimes it's interesting to know like, oh, this library, it doesn't have an update, but it's 10 years old. 00:02:55.960 --> 00:03:01.200 Maybe I should consider switching to a library that's a little more maintained and making progress. 00:03:01.200 --> 00:03:01.640 Right. 00:03:01.640 --> 00:03:02.420 What do you all think? 00:03:02.420 --> 00:03:03.180 That's handy. 00:03:03.180 --> 00:03:04.240 Cool, right? 00:03:04.240 --> 00:03:05.340 It is pretty neat. 00:03:05.680 --> 00:03:13.960 So, yeah, I've been playing with this today, installed it, checked it out, even pointed out that, you know, since yesterday, some things changed in one of my projects that I want to keep up to date. 00:03:13.960 --> 00:03:14.980 So I updated it. 00:03:14.980 --> 00:03:15.420 Yeah. 00:03:15.420 --> 00:03:16.100 So I like it. 00:03:16.100 --> 00:03:18.360 It's got a nice command line interface. 00:03:18.360 --> 00:03:24.260 You basically specify the Python that is in the environment that you want to check. 00:03:24.260 --> 00:03:28.580 That could either be the main Python or a environmental virtual environment Python. 00:03:28.940 --> 00:03:36.340 Like I said, you can control the caching because the first time it runs, it has to go get lots of information about each package that's installed and it's faster the second time. 00:03:36.340 --> 00:03:38.240 It also has some cool parallelism. 00:03:38.240 --> 00:03:41.480 So you can say number of jobs like --jobs. 00:03:41.480 --> 00:03:46.860 And by default, it runs 10 downloads in parallel as it's pulling this information in. 00:03:46.860 --> 00:03:48.580 But I guess you could go crazy there. 00:03:48.580 --> 00:03:50.600 So anyway, I thought this was pretty cool. 00:03:50.600 --> 00:03:52.480 It's a nice little thing to have. 00:03:52.480 --> 00:03:54.660 So I pipx installed this. 00:03:54.660 --> 00:03:58.800 It's perfect for pipx because it doesn't need to be in the project. 00:03:58.800 --> 00:03:59.540 It's testing. 00:03:59.540 --> 00:04:02.120 It just needs to be on your machine as a command. 00:04:02.120 --> 00:04:06.600 And then you point it at the environment, different environments, and it gives you reports on those environments. 00:04:06.600 --> 00:04:08.180 Yeah, I love pipx too. 00:04:08.180 --> 00:04:19.940 One of the things I want to note, just as I know a lot of package maintainers, having, I mean, it's worth checking things out if it's a really old packet, if it hasn't been updated for a long time. 00:04:19.940 --> 00:04:25.320 But some things are pure Python things that just do a little tiny thing and don't need updated very often. 00:04:25.320 --> 00:04:30.460 So it's not necessarily a bad thing that it's not updated, but it's an indicator. 00:04:30.460 --> 00:04:31.160 It's an indicator. 00:04:31.160 --> 00:04:31.580 Yeah. 00:04:31.580 --> 00:04:32.220 Of something. 00:04:32.220 --> 00:04:33.480 Yes, exactly. 00:04:33.480 --> 00:04:34.500 Let's see. 00:04:34.500 --> 00:04:39.800 Out in the live audience, Anthony Listerhate says, can the changes be exported to a text file? 00:04:39.800 --> 00:04:43.760 I haven't seen anything about that other than just, you know, piping it into a text file. 00:04:43.760 --> 00:04:47.980 And who knows what happens with all the color in there, but perhaps. 00:04:47.980 --> 00:04:48.600 Yeah. 00:04:48.600 --> 00:04:52.300 Renee, what do you do to manage your dependencies and all those kinds of things? 00:04:52.300 --> 00:04:55.460 Well, at work, we started using Docker for that. 00:04:55.460 --> 00:05:02.400 So we have a centralized Docker container that everyone on my team uses, and we make sure we have the same setup in there. 00:05:02.400 --> 00:05:10.060 So I'm not the one that directly manages it, but that's the solution that we've gone towards to make sure we're all on the same page. 00:05:10.060 --> 00:05:10.760 Oh, interesting. 00:05:10.760 --> 00:05:19.320 So you've got the Docker environment that has some version of Python set up with all the libraries you need pre-installed, and then you just use that to run, and that way you know it's the same. 00:05:19.680 --> 00:05:26.560 Yep. And then it's also nice because when we kind of move some of our projects into production, we can include that Docker container with it. 00:05:26.560 --> 00:05:28.960 So it will have whatever version it had at the time. 00:05:28.960 --> 00:05:37.780 So if for some reason it's not compatible with some later version we upgraded to, it still lives out there with the version of the tools that it had until we have a chance to update everything. 00:05:38.020 --> 00:05:50.560 One of the challenges that people have sometimes is they say, even though you've got some kind of version management, iProject.toml or requirements.txt or whatever, that doesn't necessarily mean that people actually install them the latest. 00:05:50.560 --> 00:05:52.260 So you could still be out of sync, right? 00:05:52.260 --> 00:05:57.280 So having the image that's constantly the same, constantly in sync, that's the kind of way to force it. 00:05:57.280 --> 00:06:01.780 I also want to give a quick shout out to this project, pip.dep.tree. 00:06:01.780 --> 00:06:04.600 Remember, Brian, we spoke about that before, which is pretty cool. 00:06:04.600 --> 00:06:12.300 And what it'll show you is this will show you the things you've directly installed versus the things that happen to be installed. 00:06:12.300 --> 00:06:18.640 So if we go back to this like animation here, you can see that it's got Flask, which is 202. 00:06:18.640 --> 00:06:21.460 But then it's got markup safe. 00:06:21.460 --> 00:06:22.600 It's got it's dangerous. 00:06:22.600 --> 00:06:24.300 Like nobody installed it's dangerous. 00:06:24.300 --> 00:06:26.800 That's a thing that was installed because of Flask. 00:06:26.800 --> 00:06:34.160 And so, for example, when I look at my environments, there were some things that were out of date, but they were out of date because they were pinned requirements. 00:06:34.160 --> 00:06:36.740 Of other things that I actually wanted to install. 00:06:36.740 --> 00:06:41.700 So, for example, example, doc opt and some other things are pinned to lower versions. 00:06:41.700 --> 00:06:45.160 And I can't really update those, but they'll show up as outdated. 00:06:45.160 --> 00:06:51.800 So you might pair this with some pip.dep.tree to see like what ones are you in control of and what ones are just kind of out there. 00:06:51.800 --> 00:06:52.720 That's pretty cool. 00:06:52.720 --> 00:06:53.640 That's that one. 00:06:53.640 --> 00:06:54.760 What do you got for us? 00:06:55.220 --> 00:07:00.840 Well, this is an interesting discussion about a possible change to future Python. 00:07:00.840 --> 00:07:04.440 Again, this is just stuff that people are discussing. 00:07:04.440 --> 00:07:06.820 It's nothing that's even decided on. 00:07:06.820 --> 00:07:12.180 But it's an idea of late bound arguments for Python. 00:07:12.180 --> 00:07:17.800 For late bound arguments for defaults. 00:07:17.800 --> 00:07:20.460 That's it for functions. 00:07:20.920 --> 00:07:22.080 Here's the idea. 00:07:22.080 --> 00:07:33.840 So we know that if you sign the default value for a function argument, that is bound at definition time. 00:07:33.840 --> 00:07:36.320 So when Python first goes and reads it. 00:07:36.320 --> 00:07:38.640 That seems fine. 00:07:38.640 --> 00:07:41.920 It's a weird thing about the namespace there, though. 00:07:41.920 --> 00:07:52.980 So what happens is if you have a variable foo, for instance, or a value foo, the value expression can be looked. 00:07:52.980 --> 00:07:56.200 You can look that up in the defining area. 00:07:56.200 --> 00:07:58.200 So the namespace where the function is defined. 00:07:58.200 --> 00:08:02.140 It's a little specific, but it causes some weirdness. 00:08:02.140 --> 00:08:04.440 It's not the namespace of the function. 00:08:04.440 --> 00:08:06.920 It's the namespace of the surrounding the function. 00:08:07.400 --> 00:08:21.140 The problem with that really is that, like, for instance, if we wanted to do something like a bisect function that took, you know, has a given an array and maybe an x value for the middle or something. 00:08:21.140 --> 00:08:22.800 We also have a high and low. 00:08:22.800 --> 00:08:25.760 We know the low index would be zero as a default. 00:08:25.760 --> 00:08:29.460 But what the high should be is should be the length of the array. 00:08:30.000 --> 00:08:35.480 And it's you can't do that because you can't reference the array as a default value. 00:08:35.480 --> 00:08:45.580 So that's what this kind of what this discussion is about is trying to figure out a way to possibly have an optional late binding of those values. 00:08:45.580 --> 00:08:55.280 And and in this specific case, it'd be very helpful to be able to late bind that value, like at the time that the function is called, not at the time that it's defined. 00:08:55.660 --> 00:09:02.080 And this was you want to take the first parameter and use it to set the default value of the subsequent parameter. 00:09:02.080 --> 00:09:02.680 Yeah. 00:09:02.680 --> 00:09:07.300 Like to say, like the length of the array is the default for length or something. 00:09:07.300 --> 00:09:14.760 And that's the that was was it Chris Angelico that suggested this. 00:09:14.860 --> 00:09:21.320 And the discussion actually got has some people even even Guido said, I'm not really opposed to it. 00:09:21.320 --> 00:09:23.020 Let's let's explore it a little bit. 00:09:23.020 --> 00:09:26.680 So there is some Chris is trying to do a proof of concept. 00:09:26.680 --> 00:09:30.840 There is some question about what the syntax was should be. 00:09:30.840 --> 00:09:35.200 So Chris suggested a equal colon. 00:09:35.200 --> 00:09:40.540 So like the reverse of the walrus operator, because apparently that's available. 00:09:40.540 --> 00:09:46.220 Another suggestion was equal greater than to kind of look like an arrow. 00:09:46.220 --> 00:09:49.520 But we already have like dash arrow to mean something else. 00:09:49.740 --> 00:09:52.280 So up in the air on the syntax. 00:09:52.280 --> 00:10:01.140 But anyway, one of the things I wanted to comment about is the in the article we're linking to. 00:10:01.140 --> 00:10:08.100 It says at first blush, Angelico's idea to fix this wart in Python seems fairly straightforward. 00:10:08.320 --> 00:10:12.080 But the discussion has shown that there are multiple facets to consider. 00:10:12.080 --> 00:10:12.500 Oh, yeah. 00:10:12.500 --> 00:10:16.680 And it's always tricky to add complexity or to the language. 00:10:16.680 --> 00:10:21.420 So, you know, the people in the steering council will take it. 00:10:21.420 --> 00:10:22.360 Think about it. 00:10:22.360 --> 00:10:22.600 Right. 00:10:22.600 --> 00:10:24.020 Under consideration. 00:10:24.020 --> 00:10:24.420 Yes. 00:10:24.420 --> 00:10:24.680 OK. 00:10:24.680 --> 00:10:26.400 Renee, what do you think about this? 00:10:26.400 --> 00:10:27.420 I'm going to be honest. 00:10:27.420 --> 00:10:29.160 It's going over my head a little bit. 00:10:29.160 --> 00:10:32.760 I don't consider myself like a real software developer. 00:10:32.760 --> 00:10:36.360 So I usually use Python for, you know, standard data science type of scripts. 00:10:37.260 --> 00:10:41.640 I'm trying to sit here thinking of a use case for this that I would use and not coming up with one. 00:10:41.640 --> 00:10:43.980 Yeah, I'm with you on that one as well. 00:10:43.980 --> 00:10:46.560 That doesn't mean it's a bad idea necessarily. 00:10:46.560 --> 00:10:55.060 Well, one use case would be to be able to set an empty list as a default value. 00:10:55.060 --> 00:10:58.460 You can't do that now because the list is bound. 00:10:58.460 --> 00:11:03.960 All calls to the function will get whatever the last function set it to. 00:11:03.960 --> 00:11:07.100 And that's a weirdness in Python, but we could probably fix that with this. 00:11:07.100 --> 00:11:07.620 Yeah. 00:11:07.620 --> 00:11:08.100 Yeah. 00:11:08.100 --> 00:11:14.540 That's what I was thinking as well, is if you pass immutable value as the default, then you're asking for trouble. 00:11:14.540 --> 00:11:14.840 Right. 00:11:14.840 --> 00:11:20.240 Because if it gets changed anywhere, then every subsequent call gets those changes applied to it. 00:11:20.240 --> 00:11:21.400 So that seems useful. 00:11:21.400 --> 00:11:24.880 This like sort of flowing one parameter into the next. 00:11:25.820 --> 00:11:27.980 I'm not sure it's worth the complexity. 00:11:27.980 --> 00:11:38.660 So, Renee, what I wanted to ask you was, as somebody who doesn't dive deep into the low levels of the language and like compiler parsing, all that kind of stuff, which is totally fine. 00:11:38.660 --> 00:11:40.480 That's like 99% of the people. 00:11:40.480 --> 00:11:44.060 How do you feel about these kinds of new features coming along? 00:11:44.060 --> 00:11:46.700 Are you like, oh, geez, now I got to learn the walrus operator. 00:11:46.700 --> 00:11:48.400 I got to learn pattern matching. 00:11:48.400 --> 00:11:49.360 I was fine. 00:11:49.360 --> 00:11:50.500 And now I got to deal with this code. 00:11:50.500 --> 00:11:50.960 What is this? 00:11:50.960 --> 00:11:52.480 Or do you see it as like, oh, awesome. 00:11:52.480 --> 00:11:53.160 Here's new stuff. 00:11:53.500 --> 00:11:57.260 Yeah, I mean, I guess it depends how much it really impacts my day-to-day work. 00:11:57.260 --> 00:12:05.260 If it's something that it's not impacting something I use frequently or it's kind of abstracted away from me or optional, then, you know, go ahead. 00:12:05.260 --> 00:12:13.000 But if it's something that some, you know, some features they roll out clearly have a wide-ranging impact and you have to go update everything. 00:12:13.000 --> 00:12:20.340 So I'm not great at keeping up with that, which is one reason that, you know, of course, you have to be so careful when you update to a new version. 00:12:20.660 --> 00:12:24.260 But, you know, I guess that's why people listen to podcasts like this. 00:12:24.260 --> 00:12:25.820 So you know it's potentially coming. 00:12:25.820 --> 00:12:26.720 So you're aware. 00:12:26.720 --> 00:12:27.360 I guess so. 00:12:27.360 --> 00:12:29.880 When it does come out, you're on top of it. 00:12:29.880 --> 00:12:32.220 But I don't have strong opinions. 00:12:32.220 --> 00:12:35.560 And what we worry about a lot in data science is the packages, right? 00:12:35.560 --> 00:12:40.240 So not the base Python, but the packages are constantly changing and the dependencies and the versions. 00:12:40.240 --> 00:12:45.120 So that does end up affecting us when it follows through to that level. 00:12:45.120 --> 00:12:48.880 Yeah, my concern is around teaching Python. 00:12:49.180 --> 00:12:55.740 Because every new syntax thing you put in makes it something that you potentially have to teach somebody. 00:12:55.740 --> 00:13:01.720 And maybe you don't have to teach newbies this, but they'll see it in code. 00:13:01.720 --> 00:13:03.900 So they should be able to understand what it is. 00:13:04.740 --> 00:13:12.940 But on the other hand, like things like you can do really crazy comprehensions, list comprehensions and stuff, but you don't have to. 00:13:12.940 --> 00:13:14.920 And most of the ones I see are fairly simple ones. 00:13:14.920 --> 00:13:17.360 So I don't think we should nix a phone. 00:13:17.360 --> 00:13:20.740 Nick shouldn't nix something just because it can be complicated. 00:13:20.740 --> 00:13:21.360 Anyway. 00:13:21.360 --> 00:13:22.220 Yeah. 00:13:22.220 --> 00:13:22.980 Indeed. 00:13:22.980 --> 00:13:23.260 Yeah. 00:13:23.260 --> 00:13:23.620 Good one. 00:13:23.620 --> 00:13:24.060 All right. 00:13:24.060 --> 00:13:24.880 Renee, you got the next one. 00:13:24.880 --> 00:13:25.560 All right. 00:13:25.560 --> 00:13:29.840 So speaking of data science packages, a lot of us use Pandas. 00:13:29.840 --> 00:13:34.120 So I wrote a book, which I'll come back to later, called SQL for Data Scientists. 00:13:34.960 --> 00:13:45.820 And since I wrote that and, you know, some people that have been learning data science in school or, you know, on the job haven't always used SQL or if they use it as kind of a separate process from their Python. 00:13:45.940 --> 00:13:48.980 So they started asking me, how do you use SQL alongside Python? 00:13:48.980 --> 00:13:54.060 So this is kind of beginner level, but also something that's just very useful. 00:13:54.060 --> 00:13:58.420 In the Pandas package, there's a read SQL function. 00:13:58.420 --> 00:14:00.840 And so you can read a SQL query. 00:14:00.840 --> 00:14:02.300 It runs the query. 00:14:02.300 --> 00:14:04.360 It's kind of a wrapper around some other functions. 00:14:04.360 --> 00:14:08.580 It will run the query and return the data set into your data frame. 00:14:08.580 --> 00:14:14.680 And so basically you're just running a query and the results become the data frame right in your notebook. 00:14:14.920 --> 00:14:17.780 So let's see some of my notes on here. 00:14:17.780 --> 00:14:20.800 So you can save your SQL as a text file. 00:14:20.800 --> 00:14:24.960 So you don't have to have the string in your actual notebook, which is sometimes useful. 00:14:24.960 --> 00:14:33.300 And then when you import it in from that Pandas data frame, that's where a lot of people do their data cleaning and feature engineering and everything like that. 00:14:33.300 --> 00:14:37.120 So you could just pull in the raw data from SQL and do a lot of the data engineering there. 00:14:37.120 --> 00:14:41.400 Sometimes I do feature engineering in SQL and then pull it in. 00:14:41.400 --> 00:14:43.600 So that's kind of up to each user. 00:14:43.760 --> 00:14:48.200 But you really just set up the connection in your database using a package like SQLAlchemy. 00:14:48.200 --> 00:14:58.720 So you have a connection to the database and you pass your SQL string either directly or from the file and the database connection and it returns a Pandas data frame. 00:14:59.180 --> 00:15:02.920 So I'm happy to talk a little bit more about how I use this at work. 00:15:03.620 --> 00:15:04.700 I think this is really good. 00:15:04.700 --> 00:15:12.540 You know, one of the things to do with Pandas is there's just so many of these little functions that solve whole problems. 00:15:12.540 --> 00:15:23.240 You know, it's like, oh, you could go and use requests to download some HTML and you could use beautiful soup to do some selectors and you could get some stuff and parse out some HTML. 00:15:23.500 --> 00:15:27.520 And then you could get some table information out and then convert that into a data frame. 00:15:27.520 --> 00:15:32.020 Or you could just say, read HTML table bracket zero or whatever. 00:15:32.020 --> 00:15:35.460 And boom, you have it like knowing about these, I think is really interesting. 00:15:35.660 --> 00:15:37.120 So it's cool that you highlighted this one. 00:15:37.120 --> 00:15:40.960 I actually just, on a side, I'm just literally like in an hour. 00:15:40.960 --> 00:15:45.260 So probably before this show ships, we'll ship this episode I did with Vex. 00:15:45.260 --> 00:15:49.280 You have about 25 Panda functions you didn't even know existed. 00:15:49.540 --> 00:15:52.060 And what's interesting is like this one wasn't even on the list. 00:15:52.060 --> 00:15:54.260 So good. 00:15:54.260 --> 00:15:55.400 So I'll highlight another one. 00:15:55.400 --> 00:15:56.600 Now, you know, exists. 00:15:56.600 --> 00:15:58.080 That's pretty cool. 00:15:58.080 --> 00:16:00.720 Let's see a couple of comments from the audience. 00:16:00.720 --> 00:16:01.400 Sam Morley. 00:16:01.400 --> 00:16:03.320 Hey, Sam says, Pandas is so amazing. 00:16:03.320 --> 00:16:07.380 I always find something too late that it has all of these IO functions. 00:16:07.380 --> 00:16:10.760 And then we have Paul Ansel. 00:16:10.760 --> 00:16:16.720 Hey, Paul says, do you have any recommendations on tutorials for how to create good SQLAlchemy selectables? 00:16:16.720 --> 00:16:18.980 This always feels like the scariest bit. 00:16:19.360 --> 00:16:20.740 I don't have any of that on hand. 00:16:20.740 --> 00:16:25.180 I'll try to find something later or I'll ask my Twitter following and see what they recommend. 00:16:25.180 --> 00:16:27.480 I don't have a good list of tutorials for that one. 00:16:27.480 --> 00:16:32.060 I can talk about, yeah, by selectable, he said he means connectable. 00:16:32.060 --> 00:16:35.900 So, yeah, I don't have a tutorial for that. 00:16:35.900 --> 00:16:36.940 There's a lot of documentation. 00:16:36.940 --> 00:16:40.620 And I know that SQLAlchemy can be a little mysterious sometimes. 00:16:40.620 --> 00:16:42.000 Maybe that's why it's alchemy. 00:16:42.000 --> 00:16:46.760 But, yeah, I will try to share that later on Twitter. 00:16:46.760 --> 00:16:47.980 Yeah, fantastic. 00:16:47.980 --> 00:16:48.660 All right. 00:16:48.680 --> 00:16:51.580 And Paul says, read clipboard is pretty great. 00:16:51.580 --> 00:16:52.180 Yeah. 00:16:52.180 --> 00:16:53.320 Yeah, very cool. 00:16:53.320 --> 00:16:54.680 Bunch of different things there. 00:16:54.680 --> 00:16:55.180 Yeah. 00:16:55.180 --> 00:16:58.800 So if you want me to walk through an example of how I use this at work, I'm happy to do that. 00:16:58.800 --> 00:16:59.920 Yeah, give us a quick example. 00:16:59.920 --> 00:17:00.600 Yeah. 00:17:00.740 --> 00:17:06.540 So at Helio Campus, one thing we do is we connect to a lot of different databases at universities. 00:17:06.540 --> 00:17:12.240 So the universities will have separate databases for admissions, enrollment, financial aid. 00:17:12.240 --> 00:17:12.740 Those are all separate systems. 00:17:12.740 --> 00:17:14.240 Those are all separate systems. 00:17:14.240 --> 00:17:16.580 And so we pull all that data into a data warehouse. 00:17:16.580 --> 00:17:22.080 And in SQL, we can combine that data. 00:17:22.080 --> 00:17:26.080 And so we can either use this to just read one of those tables directly. 00:17:26.700 --> 00:17:36.580 Or we can combine what I typically do is do a little bit of cleanup and feature engineering and narrow down my data set to the population that I want to run through my model in SQL. 00:17:36.580 --> 00:17:38.700 And then just pull those final results. 00:17:38.700 --> 00:17:41.980 And now I've got my data set with at least preliminary features. 00:17:41.980 --> 00:17:44.840 I might do some standardization and things like that in pandas. 00:17:44.840 --> 00:17:52.040 But I've got a pretty clean and subset of the data that I need right into my Jupyter notebook. 00:17:52.040 --> 00:17:53.160 Oh, that's fantastic. 00:17:53.160 --> 00:17:53.900 Pretty great. 00:17:53.900 --> 00:17:58.860 Yeah, I think definitely understanding SQL is an important skill for data scientists. 00:17:58.860 --> 00:18:04.500 And it's slightly different than for, say, like a web API developer, right? 00:18:04.500 --> 00:18:04.840 Absolutely. 00:18:04.840 --> 00:18:05.980 That's why I wrote the book. 00:18:05.980 --> 00:18:06.960 That's awesome. 00:18:06.960 --> 00:18:07.540 Yeah, for sure. 00:18:07.540 --> 00:18:10.780 So on the API side, you kind of get something set up. 00:18:10.780 --> 00:18:15.120 You're very likely using an ORM like SQLAlchemy and you just connect it and go. 00:18:15.120 --> 00:18:18.500 And once you get it set, you can kind of forget about it and just program against it. 00:18:18.500 --> 00:18:20.440 As a data scientist, you're exploring. 00:18:20.440 --> 00:18:21.700 You don't totally know, right? 00:18:21.700 --> 00:18:25.900 You're kind of out there testing and digging into stuff and sorting and filtering. 00:18:25.900 --> 00:18:33.960 And yeah, I would say you probably need a better fluency with SQL as a data scientist a lot of times than as a web developer. 00:18:33.960 --> 00:18:37.620 Because I can just use an ORM or ODM and just kind of know what it's doing. 00:18:37.740 --> 00:18:45.220 Yeah, and it enables you to build your own data sets instead of relying on a data engineering trying to explain what you need and why and which fields you need. 00:18:45.220 --> 00:18:48.180 Now you could just do it yourself or, you know, add a field if you need it. 00:18:48.180 --> 00:18:51.600 You can do more sophisticated things like window functions. 00:18:51.600 --> 00:18:59.720 So yeah, I think knowing SQL is really a value add if you're looking to become a data scientist and, you know, putting yourself out there on the market. 00:18:59.720 --> 00:19:03.620 If you can do the whole pipeline end to end, it definitely makes you stand out. 00:19:03.620 --> 00:19:04.540 I would think so. 00:19:04.540 --> 00:19:05.140 All right. 00:19:05.140 --> 00:19:06.720 One thing to wrap up on this. 00:19:06.720 --> 00:19:10.780 Sam asks, can you configure SQLAlchemy to dump the raw queries that it runs? 00:19:10.780 --> 00:19:11.540 Yes. 00:19:11.540 --> 00:19:12.300 Yeah. 00:19:12.540 --> 00:19:15.980 In this case, you have the raw query in your function call. 00:19:15.980 --> 00:19:20.620 So I'm not actually using SQLAlchemy for that because I'm providing a query. 00:19:20.620 --> 00:19:21.340 Yeah, you just got a select statement, right? 00:19:21.340 --> 00:19:28.660 The problem with SQLAlchemy and data science is you have to, the structure of your models has to exactly match the structure of the data. 00:19:28.760 --> 00:19:35.040 And often I imagine you're just kind of dealing with the loose data and it doesn't make sense to take the time to like model it in classes. 00:19:35.040 --> 00:19:38.560 But for SQLAlchemy, you can just set echo equal true when you create the engine. 00:19:38.560 --> 00:19:46.000 And then everything that would get sent to the crossover to the database gets echoed as like DDL or SQL or whatever that it does. 00:19:46.000 --> 00:19:46.700 So yes. 00:19:46.700 --> 00:19:47.220 Good. 00:19:47.220 --> 00:19:47.380 Cool. 00:19:47.380 --> 00:19:47.820 For sure. 00:19:47.820 --> 00:19:48.360 Yeah. 00:19:48.360 --> 00:19:49.060 All right. 00:19:49.060 --> 00:19:51.660 Brian, want to tell us about our sponsor? 00:19:51.660 --> 00:19:52.820 Yeah, let's. 00:19:52.820 --> 00:19:57.940 I am pleased and happy that Shortcut is sponsoring the episode. 00:19:58.280 --> 00:19:59.120 So thank you, Shortcut. 00:19:59.120 --> 00:20:01.880 Formerly Clubhouse for sponsoring the episode. 00:20:01.880 --> 00:20:06.640 There are a lot of project management tools out there, but most suffer from common problems. 00:20:06.640 --> 00:20:14.160 Like it's too simple for an engineering team to use on several projects or it's too complex and it's hard to get started. 00:20:14.160 --> 00:20:16.080 And there's tons of options. 00:20:16.080 --> 00:20:19.360 And some of them are great for managers, but bad for engineers. 00:20:19.360 --> 00:20:21.440 And some are great for engineers and bad for managers. 00:20:21.440 --> 00:20:22.860 Shortcut is different. 00:20:22.860 --> 00:20:27.800 It's built for software teams and based on making workflows super easy. 00:20:27.800 --> 00:20:30.680 For example, keyboard friendly user interface. 00:20:30.680 --> 00:20:33.660 The UI is intuitive for mouse lovers, of course. 00:20:33.660 --> 00:20:39.860 But the activities that you use every day can be set to keyboard shortcuts if they aren't already. 00:20:39.860 --> 00:20:43.680 Just learn them and you'll start working faster. 00:20:43.680 --> 00:20:44.220 It's awesome. 00:20:44.220 --> 00:20:46.060 Tight VCS integration. 00:20:46.060 --> 00:20:49.940 So you can update task progress and commits with a commit or a PR. 00:20:49.940 --> 00:20:50.860 That's sweet. 00:20:51.300 --> 00:20:53.200 And iteration planning is a breeze. 00:20:53.200 --> 00:20:57.320 I like that there's a burndown and cycle time charts built in. 00:20:57.320 --> 00:21:00.560 They just are set up already for you when you start using this. 00:21:00.560 --> 00:21:02.360 So it's a pretty clean system. 00:21:02.360 --> 00:21:05.640 Give it a try at shortcut.com slash Python bytes. 00:21:05.640 --> 00:21:06.600 Yeah, absolutely. 00:21:06.600 --> 00:21:07.360 Thanks, Shortcut. 00:21:07.360 --> 00:21:08.540 For sponsoring this episode. 00:21:09.100 --> 00:21:10.880 Now, what have we got next here? 00:21:10.880 --> 00:21:11.720 Pidgin. 00:21:11.720 --> 00:21:12.700 I want to talk about Pidgin. 00:21:12.700 --> 00:21:17.940 So we already talked about Will McGooghan and Rich. 00:21:17.940 --> 00:21:24.000 So it's time to talk about Anthony Shaw so that we can complete our shoutouts we always seem to give over on the podcast. 00:21:24.480 --> 00:21:27.920 So I want to talk about Pidgin because I just interviewed Anthony Shaw. 00:21:27.920 --> 00:21:32.140 But more importantly, he just released Pidgin as 1.0. 00:21:32.140 --> 00:21:36.440 So Pidgin is a drop-in JIT compiler for Python 3.10. 00:21:36.440 --> 00:21:37.460 Let me say that again. 00:21:37.460 --> 00:21:39.800 A JIT compiler for Python. 00:21:39.800 --> 00:21:48.900 And there have been other speed-up type of attempts where people will like fork CPython and they'll do something inside of it to make it different. 00:21:48.900 --> 00:21:49.800 Think Cinder. 00:21:49.920 --> 00:21:55.180 There have been attempts to create a totally different but compatible one like PyPy. 00:21:55.180 --> 00:22:02.240 And they've worked pretty well, but they always have some sort of incompatibility or something. 00:22:02.240 --> 00:22:07.780 It would be nice if just the Python you ran could be compiled to go faster if you want it to be. 00:22:07.780 --> 00:22:09.280 So that's what this is. 00:22:09.280 --> 00:22:17.100 It uses a PEP whose number I forgot that allows you to plug in something that inspects the method frames before they get executed. 00:22:17.100 --> 00:22:26.580 And then instead of just interpreting that code, the bytecode as Python bytecode, it'll actually compile it to machine instructions. 00:22:26.580 --> 00:22:30.680 First to .NET intermediate instructions, intermediate language. 00:22:30.680 --> 00:22:34.880 And then those get JIT compiled to machine instructions that then run directly. 00:22:34.880 --> 00:22:39.440 Works on Linux, macOS, Windows, X64, and ARM64. 00:22:39.440 --> 00:22:42.260 So this is a pretty cool development. 00:22:42.740 --> 00:22:43.700 It is pretty cool. 00:22:43.700 --> 00:22:44.200 Yeah. 00:22:44.200 --> 00:22:49.860 So if we go over here and check it out, in order to use it, it has some requirements. 00:22:49.860 --> 00:22:50.960 You just pip install Pigeon. 00:22:50.960 --> 00:22:51.680 That's it. 00:22:51.680 --> 00:22:52.540 That's crazy, right? 00:22:52.540 --> 00:22:54.400 And then it has to be on 310. 00:22:54.400 --> 00:22:55.340 It can't be older than that. 00:22:55.340 --> 00:22:58.000 And you have to have .NET 6 installed. 00:22:58.000 --> 00:22:58.540 Okay. 00:22:58.540 --> 00:22:59.960 So that just got released. 00:22:59.960 --> 00:23:02.240 It's a good chance you don't have .NET 6 installed. 00:23:02.640 --> 00:23:07.780 But then once you set it up right, you can just say import Pigeon, Pigeon.enable at the startup of your app. 00:23:07.780 --> 00:23:10.800 And then it will look at all the methods and JIT compile them. 00:23:11.460 --> 00:23:16.600 So if you come down here, like he has an example of a half function that Anthony put up here. 00:23:16.600 --> 00:23:20.060 And when it first loads, it's not JIT compiled. 00:23:20.060 --> 00:23:24.580 But after that, you can go and say, if you run it, you can say disassemble this thing. 00:23:24.580 --> 00:23:31.060 And it'll show you basically assembly instructions of what would have otherwise been Python code. 00:23:31.060 --> 00:23:31.560 Wow. 00:23:31.560 --> 00:23:32.440 It's wild, right? 00:23:32.640 --> 00:23:34.560 So it's a little bit like Numba. 00:23:34.560 --> 00:23:44.600 It's a little bit like a tiny bit like Cython in the sense that it takes Python code, translates it into something else that then can be interoperated with, and then makes it go fast. 00:23:44.600 --> 00:23:46.400 So this is all well and good. 00:23:46.400 --> 00:23:49.800 If you're going to use it on the web, by default, it would be just fine. 00:23:49.800 --> 00:23:57.560 Except if you're hosting it, normally you host it in this supervisor process and then a bunch of forked off processes. 00:23:57.960 --> 00:24:02.920 So there's a WSGI app configuration thing you can do as well somewhere in the docs. 00:24:02.920 --> 00:24:03.700 I'm not seeing it right now. 00:24:03.700 --> 00:24:11.300 But you basically allow it to push the pigeon changes on down into the worker processes, which is pretty cool. 00:24:11.300 --> 00:24:19.300 And it has a bunch of comparisons against PyPy, Piston, Numba, IronPython, et cetera, Nutka, and so on. 00:24:19.300 --> 00:24:21.200 Now, it's not that much faster. 00:24:21.200 --> 00:24:29.800 It is faster when you're doing more data science-y things, I believe, than if you're doing just a query against a database where you're mostly just waiting anyway. 00:24:29.800 --> 00:24:33.240 But still, I think this is promising, and it's really pretty early days. 00:24:33.240 --> 00:24:40.380 So the thing to look at is if there's optimizations coming along here somewhere in the docs. 00:24:41.220 --> 00:24:50.260 Anthony lists out the various optimizations he's put in so far, and really, it just needs more optimizations to make it faster still, which is pretty neat. 00:24:50.260 --> 00:24:51.300 I think that's pretty cool. 00:24:51.300 --> 00:24:57.740 One of the things that my first reaction was, oh, it's .NET only, so I have to use it on Windows. 00:24:57.740 --> 00:25:00.880 But that's not been that way for a long time. 00:25:00.880 --> 00:25:02.540 So .NET runs on just about everything. 00:25:02.540 --> 00:25:03.360 Yeah, exactly. 00:25:03.460 --> 00:25:05.120 It supports all the different frameworks. 00:25:05.120 --> 00:25:11.660 There's even this thing called live.trypigeon.com, where you can write Python code, like, over here on the left. 00:25:11.660 --> 00:25:13.580 And then you can say, compile it. 00:25:13.580 --> 00:25:17.860 And it will actually show you the assembly that it would compile to. 00:25:17.860 --> 00:25:20.740 And then here's the .NET intermediate language. 00:25:20.740 --> 00:25:23.880 I guess maybe they should possibly be switching orders here. 00:25:23.880 --> 00:25:27.080 Like, first it goes to IL, and then it goes to machine instructions through the JIT. 00:25:27.080 --> 00:25:31.140 But it shows you all the stuff that it's doing to make this work. 00:25:31.160 --> 00:25:35.980 You could even see at the bottom, there's, like, sort of a visual understanding of what it's doing. 00:25:35.980 --> 00:25:40.580 One of the things that's really cool that it does is, imagine you've got a math problem up here. 00:25:40.580 --> 00:25:47.940 Like, you're saying, like, X equals Y times Y plus Z times Z, or, you know, something like that. 00:25:47.940 --> 00:25:51.160 Like, each one of those steps generates an intermediate number. 00:25:51.160 --> 00:25:56.880 So, for example, Z times Z would generate, by default, a Python number. 00:25:56.880 --> 00:26:00.720 And then so would Y times Y, and then the addition, and finally you assign it. 00:26:01.120 --> 00:26:07.680 What it'll do is it'll say, well, okay, if those are two floats, let's just store that as a C float in the intermediate computation. 00:26:07.680 --> 00:26:09.200 And then that's as a C float. 00:26:09.200 --> 00:26:13.740 And so it can sort of stay lower level as it's doing a lot of computational type of things. 00:26:13.740 --> 00:26:15.580 So there's a bunch of interesting optimizations. 00:26:15.580 --> 00:26:16.580 People can check this out. 00:26:16.580 --> 00:26:17.980 I haven't had a chance to try it yet. 00:26:17.980 --> 00:26:19.640 I was hoping to, but haven't got there yet. 00:26:19.640 --> 00:26:22.120 Yeah, really interesting conversation you had with him, too. 00:26:22.120 --> 00:26:23.040 Oh, thanks. 00:26:24.040 --> 00:26:33.540 It's interesting timing to just get him to jump on this, like, right after he wrote the book on Python internals, CPython internals, to jump into this. 00:26:33.540 --> 00:26:36.840 Well, I guess he's working on it before, but still. 00:26:36.840 --> 00:26:40.120 Yeah, you definitely got to know CPython internals to do this. 00:26:40.120 --> 00:26:48.060 Renee, do you guys do anything to optimize your code with, like, Numba or Cython or anything like that? 00:26:48.060 --> 00:26:51.120 Or are you just running straight Python and letting the libraries deal with it? 00:26:51.120 --> 00:26:52.200 Yeah, not currently. 00:26:52.200 --> 00:26:59.340 We have a pretty good server and are working with relatively small data sets, you know, not millions of rows, for example. 00:26:59.340 --> 00:27:03.100 So for right now, we haven't gone in this direction at all. 00:27:03.100 --> 00:27:11.060 I can imagine this would also be really useful if you were a computer science student and trying to understand what's going on under the hood when you run these things. 00:27:11.060 --> 00:27:21.020 So it's interesting that it's for the people that aren't seeing the visual, you kind of have three columns here with the code side by side to kind of get a peek under the hood at what's going on there. 00:27:21.020 --> 00:27:23.180 But no, this isn't something I've used personally. 00:27:23.180 --> 00:27:24.080 Yeah, yeah. 00:27:24.080 --> 00:27:25.360 I haven't used it either. 00:27:25.360 --> 00:27:26.640 But like I said, I would like to. 00:27:26.640 --> 00:27:30.380 I think it's got the ability to just plug in and make things faster. 00:27:30.380 --> 00:27:33.100 And really, it is faster to some degree. 00:27:33.100 --> 00:27:35.040 Sometimes I think it's slower, sometimes faster. 00:27:35.040 --> 00:27:40.660 But the more optimizations the JIT compiler gets, the better it could be, right? 00:27:40.760 --> 00:27:54.340 So like if it could inline function calls rather than calling them, or it could, there's things like if it sees you allocating a list and putting stuff into it, it can skip some intermediate steps and just straight allocate that. 00:27:54.340 --> 00:28:01.160 Or if you're doing accessing elements by index out of the list, it can just do pointer operations instead of going through the Python APIs. 00:28:01.160 --> 00:28:06.700 There's a lot of hard work that Anthony's put into this, and I think it's pretty cool. 00:28:06.700 --> 00:28:08.400 Yeah, I haven't tried it. 00:28:08.400 --> 00:28:08.940 I would like to. 00:28:08.940 --> 00:28:10.000 Yeah. 00:28:10.360 --> 00:28:10.560 Cool. 00:28:10.560 --> 00:28:11.200 Indeed. 00:28:11.200 --> 00:28:11.580 All right. 00:28:11.580 --> 00:28:12.880 Brian, what do you got for us? 00:28:12.880 --> 00:28:18.120 Well, actually, before I jump to the next topic, I wanted to mention, I wanted to shoo into this last conversation. 00:28:18.120 --> 00:28:27.580 Brett Cannon just wrote an interesting article called Selecting a Programming Language Can Be a Form of Premature Optimization. 00:28:27.720 --> 00:28:32.940 And this is relevant to the conversation. 00:28:32.940 --> 00:28:40.620 He says if you think Python might be too slow, another implementation like Pigeon is like step three. 00:28:40.620 --> 00:28:47.680 So first prototype in Python, then optimize your data structures and algorithms and also profile it. 00:28:47.680 --> 00:28:52.260 And then try another implementation before you abandon Python altogether. 00:28:52.880 --> 00:28:59.380 And then, you know, you can do some late bindings like language bindings to connect to see if you need to or rust. 00:29:00.100 --> 00:29:07.440 But I think it ties in as like when would I choose Pigeon or PyPI over CPython. 00:29:07.440 --> 00:29:10.580 Well, it's step three, just to let people know. 00:29:10.580 --> 00:29:12.380 Step three. 00:29:12.380 --> 00:29:12.800 Got it. 00:29:12.800 --> 00:29:13.540 Step three. 00:29:13.940 --> 00:29:17.540 I wanted to do something more lighthearted, like use print for debugging. 00:29:17.540 --> 00:29:20.180 So I love this article. 00:29:20.180 --> 00:29:22.120 I am guilty of this. 00:29:22.120 --> 00:29:26.060 Of course, I use debuggers and logging systems as well. 00:29:26.060 --> 00:29:29.580 But I also throw print statements in there sometimes. 00:29:29.580 --> 00:29:31.700 And I'm not ashamed to say it. 00:29:31.700 --> 00:29:35.460 So Adam Johnson wrote tips for debugging the print. 00:29:35.460 --> 00:29:38.980 And there were a couple that with print, there were a couple that stood out to me. 00:29:38.980 --> 00:29:41.420 I really wanted to mention because I use them a lot. 00:29:41.780 --> 00:29:48.480 Even with logging, though, is use debug variables with f-strings and the equal sign. 00:29:48.480 --> 00:29:51.100 So this is brilliant. 00:29:51.100 --> 00:29:52.300 It's been in since 3.8. 00:29:52.300 --> 00:29:59.780 Instead of typing like print widget equals and then in a string and then the widget number or something, 00:29:59.780 --> 00:30:05.260 you can just use the f-strings and do the equal sign and it interpolates for you. 00:30:05.260 --> 00:30:06.720 Or it doesn't interpolate. 00:30:06.720 --> 00:30:07.980 It just prints it for you. 00:30:07.980 --> 00:30:08.700 So it's nice. 00:30:08.700 --> 00:30:09.460 I like that. 00:30:10.160 --> 00:30:12.700 The next one is, I love this. 00:30:12.700 --> 00:30:13.620 Use emojis. 00:30:13.620 --> 00:30:14.960 I never thought to do this. 00:30:14.960 --> 00:30:15.640 This is brilliant. 00:30:15.640 --> 00:30:22.380 Throw emojis in your print statement so they pop out when you're debugging. 00:30:22.380 --> 00:30:23.260 Have you ever used emojis? 00:30:23.260 --> 00:30:23.400 Yeah, this is cool. 00:30:23.400 --> 00:30:26.180 I started using emojis in comments. 00:30:26.180 --> 00:30:27.080 Oh, okay. 00:30:27.080 --> 00:30:27.940 Comments. 00:30:27.940 --> 00:30:28.300 Nice. 00:30:28.680 --> 00:30:28.940 Yeah. 00:30:28.940 --> 00:30:33.120 So I'll put like the different emojis mean for me, I was doing some API stuff. 00:30:33.120 --> 00:30:35.780 So like, oh, this is the read only method here of an API. 00:30:35.780 --> 00:30:37.500 So I'll put a certain emoji up there. 00:30:37.500 --> 00:30:39.000 And this is one that changes data. 00:30:39.000 --> 00:30:40.540 So here's what I'll put there. 00:30:40.540 --> 00:30:42.860 Here's one that returns a list versus a single thing. 00:30:42.860 --> 00:30:45.320 So I'll put a whole bunch of those emojis and stuff. 00:30:45.600 --> 00:30:45.840 Yeah. 00:30:45.840 --> 00:30:49.520 Well, I mean, I used to do like a whole bunch of plus signs because they're easy to see. 00:30:49.520 --> 00:30:52.100 But an emoji would be way easier to see. 00:30:52.100 --> 00:30:53.100 Way more fun, man. 00:30:53.100 --> 00:30:53.740 Yeah. 00:30:53.740 --> 00:30:54.040 Yeah. 00:30:54.040 --> 00:30:55.120 I do this as well. 00:30:55.220 --> 00:30:59.640 I print all the time for debugging, especially in Jupyter Notebooks, because you don't always 00:30:59.640 --> 00:31:01.900 have the most sophisticated debugging tools in there. 00:31:01.900 --> 00:31:06.720 But being able to print and see what's going on as you go through each step of the notebook. 00:31:06.720 --> 00:31:11.100 And emojis are a great idea for that because it's so visual as you're scrolling through. 00:31:11.100 --> 00:31:14.120 You want to like they're showing there the X and the checkmark emoji. 00:31:14.120 --> 00:31:18.020 I like that for my little to-do lists in the comments that I leave for myself. 00:31:18.020 --> 00:31:18.120 Yeah, yeah. 00:31:18.120 --> 00:31:18.700 I have a thought. 00:31:18.700 --> 00:31:19.700 So I've done that as well. 00:31:19.700 --> 00:31:20.060 That's cool. 00:31:20.220 --> 00:31:26.780 Chris May out in the audience just put a heart sign, smiley face emoji as a response 00:31:26.780 --> 00:31:27.160 to this. 00:31:27.160 --> 00:31:27.640 Love it. 00:31:27.640 --> 00:31:28.380 Last thing. 00:31:28.380 --> 00:31:30.060 He's got like seven tips. 00:31:30.060 --> 00:31:36.280 The last tip I wanted to talk about was using rich and or specifically rich print, rich.print 00:31:36.280 --> 00:31:38.580 or pprint. 00:31:38.580 --> 00:31:44.920 So for pprint, you have to do from pprint install pprint or something, or unless you wanted to 00:31:44.920 --> 00:31:46.880 say it twice with pprint dot pprint. 00:31:46.880 --> 00:31:50.200 But pprint stands for pretty printing. 00:31:50.200 --> 00:31:55.100 And the gist of this really is the structures by default print horribly. 00:31:55.100 --> 00:32:00.660 If you just print like a dict or a set or something, it looks gross. 00:32:00.660 --> 00:32:04.080 But rich and pretty print make it look nice. 00:32:04.080 --> 00:32:07.200 So if you're debugging, printing with those and debugging, use that. 00:32:07.200 --> 00:32:08.500 So yeah. 00:32:08.500 --> 00:32:12.120 There's also exception handling stuff in there for it. 00:32:12.120 --> 00:32:14.520 And there's a lot of that kind of debugging stuff in rich. 00:32:14.520 --> 00:32:15.520 Yeah. 00:32:15.520 --> 00:32:17.300 Printing exceptions is great with that. 00:32:17.300 --> 00:32:22.880 I also wanted to say one of the reasons why I, one of the places where I use printing 00:32:22.880 --> 00:32:29.740 a lot for debugging is I print to print stuff in my, what I expect is going on when I'm 00:32:29.740 --> 00:32:30.600 writing a test function. 00:32:30.600 --> 00:32:33.340 So I'll often print out the flow, what's going on. 00:32:33.340 --> 00:32:40.380 The reason I do that is when if pytest for pytest, if a test fails, pytest dumps the standard 00:32:40.380 --> 00:32:40.620 out. 00:32:40.620 --> 00:32:43.940 So it'll dump all of your print statements from the failed procedure. 00:32:44.180 --> 00:32:47.260 So that's either the test under code or the test itself. 00:32:47.260 --> 00:32:49.360 If there's print statements, it gets dumped out. 00:32:49.360 --> 00:32:50.240 So that's helpful. 00:32:50.240 --> 00:32:50.800 Yeah. 00:32:50.800 --> 00:32:51.600 Nice. 00:32:51.600 --> 00:32:52.320 That's great. 00:32:52.320 --> 00:32:52.840 I love it. 00:32:52.840 --> 00:32:54.020 I use print statements a lot. 00:32:54.020 --> 00:32:55.780 My output is very verbose. 00:32:55.780 --> 00:32:58.600 You can see right in order what's happening. 00:32:58.600 --> 00:33:02.380 Sometimes a debugger helps, but sometimes it's time to just print. 00:33:02.380 --> 00:33:02.900 Yeah. 00:33:03.200 --> 00:33:05.680 Speaking of visual stuff, what's your last one here, Renee? 00:33:05.680 --> 00:33:06.300 Yeah. 00:33:06.300 --> 00:33:12.460 So in our line of work with data science, especially when you're providing the models as tools for 00:33:12.460 --> 00:33:17.780 end users that aren't the data scientists themselves, you really want the explainability is really 00:33:17.780 --> 00:33:18.120 important. 00:33:18.120 --> 00:33:23.980 So being able to explain why a certain prediction got the value it did, what the different inputs 00:33:23.980 --> 00:33:27.980 are, we're always working to make that more transparent for our end users. 00:33:28.440 --> 00:33:35.500 In our case, for example, we might be predicting which students might be at risk of not retaining 00:33:35.500 --> 00:33:36.260 at the university. 00:33:36.260 --> 00:33:37.880 So not being enrolled a year later. 00:33:37.880 --> 00:33:43.080 So what are the different factors, both overall for the whole population that are correlated with 00:33:43.080 --> 00:33:44.680 not being enrolled for a year? 00:33:44.680 --> 00:33:49.580 And for each individual student, what might be particular factors that, at least from the 00:33:49.580 --> 00:33:52.720 model's perspective, puts them at higher risk? 00:33:52.900 --> 00:33:59.600 So this package is called SHAP, and that stands for Shapley Additive Explanations. 00:33:59.600 --> 00:34:04.060 It was brought to my attention by my team member, Brian Richards at Helio Campus. 00:34:04.060 --> 00:34:08.060 And now we use it very frequently because it has really good visualizations. 00:34:08.060 --> 00:34:12.560 So these Shapley values, apparently they're from game theory. 00:34:12.560 --> 00:34:15.140 I won't pretend to understand the details of how they're generated. 00:34:15.680 --> 00:34:18.920 But you could think of it as like a model on top of your model. 00:34:18.920 --> 00:34:22.240 So it's additive and all the different features. 00:34:22.240 --> 00:34:26.400 If you see the visualization here, it's showing kind of a little waterfall chart. 00:34:26.400 --> 00:34:31.980 So some of the values that you think of a particular row that you're running through your algorithm, 00:34:31.980 --> 00:34:37.380 some of the values in that row are going to make the, if you're doing a classification model, 00:34:37.380 --> 00:34:40.120 some values might make you more likely to be in one class. 00:34:40.120 --> 00:34:42.740 Some might make you more likely to be in the other class. 00:34:42.740 --> 00:34:46.080 So you have these visuals of kind of the push and pull of each value. 00:34:46.080 --> 00:34:50.820 In this visual, we're seeing, you know, age is pushing a number to the right. 00:34:50.820 --> 00:34:52.760 Sex is pushing it to the left. 00:34:52.760 --> 00:34:55.980 I guess BP and BMI, that looks like a blood pressure. 00:34:55.980 --> 00:34:58.120 So it's got this like waterfall type of chart. 00:34:58.120 --> 00:35:00.340 And what it's actually doing is it's comparing. 00:35:00.340 --> 00:35:03.700 It starts with the expected value for the whole population. 00:35:03.700 --> 00:35:11.420 And then it's showing you where this particular record, each of the input values is kind of nudging 00:35:11.420 --> 00:35:15.360 that eventual prediction one in one direction or the other. 00:35:15.360 --> 00:35:22.480 So it's just nice visually to have those waterfall plots and to see which features are negatively 00:35:22.480 --> 00:35:27.420 or positively correlated with the, you know, the end result. 00:35:27.740 --> 00:35:30.340 And you could also do some cool scatter plots with this. 00:35:30.340 --> 00:35:37.280 So you can do the input value versus the shop value and have a point for every item in your population. 00:35:37.280 --> 00:35:39.500 So for in our, in our example, that would be students. 00:35:39.500 --> 00:35:45.200 So we can have a scatter plot of all the students and something like the number of cumulative credits 00:35:45.200 --> 00:35:46.760 that they have as of that term. 00:35:46.900 --> 00:35:52.640 And so you'll see the gradient of like from low credits to high credits, it's not usually linear. 00:35:52.640 --> 00:35:54.860 What are those kind of break points? 00:35:54.860 --> 00:36:02.840 And at what point are the values, you know, positively impactful to likelihood to retain or in the opposite direction? 00:36:02.840 --> 00:36:11.320 Of course, I'm glad that they put in this documentation, a whole, they have a whole section on basically correlation is not causation. 00:36:11.580 --> 00:36:14.740 And we're constantly having to talk to our end users about that. 00:36:14.740 --> 00:36:23.000 But, you know, if we say a student that, you know, lives in a certain town is, you know, potentially more likely to retain, 00:36:23.000 --> 00:36:28.840 maybe because of distance from campus, or maybe you have traditionally recruit a lot of students from that town. 00:36:28.840 --> 00:36:33.740 It doesn't mean that if you force someone to move to that town, they're more likely to stay at your institution. 00:36:33.740 --> 00:36:34.040 Right. 00:36:34.720 --> 00:36:37.600 So, yeah, correlation is not causation. 00:36:37.600 --> 00:36:44.920 And I'm going to switch over here to the visual, to something called a B-swarm plot that you can output this right in your Jupyter notebook, 00:36:44.920 --> 00:36:48.040 which is really handy when you develop a new model. 00:36:48.040 --> 00:36:51.940 And I'll try to describe this for people who are listening to the audio. 00:36:51.940 --> 00:36:55.760 It has along the x-axis a list of features. 00:36:56.400 --> 00:37:01.780 So you've got in this example on their website, age, relationship, capital gain, marital status. 00:37:01.780 --> 00:37:05.160 And then you see a bunch of dots going across horizontally. 00:37:05.160 --> 00:37:08.220 And there's areas where there's little clusters of dots. 00:37:08.220 --> 00:37:12.600 So what this is showing is the x-axis is the SHAP value. 00:37:12.600 --> 00:37:15.580 So what this SHAP package outputs. 00:37:16.100 --> 00:37:22.460 So you can see visually across, you know, what is the spread of the impact of this value. 00:37:22.460 --> 00:37:32.240 So if each dot is a person in this case, you see people all the way to the right, whatever their age was, positively impacted their eventual score. 00:37:32.240 --> 00:37:35.940 People all the way to the left, their age negatively impacted the score. 00:37:35.940 --> 00:37:39.700 And then each dot is a color that ranges from blue to red. 00:37:39.700 --> 00:37:45.420 So the blue ones are people with low age and the red ones are people with a high age. 00:37:45.480 --> 00:37:52.080 So you can see here, basically the higher the age, the more positive their eventual prediction. 00:37:52.080 --> 00:38:00.420 So just an interesting way to get both like a feature importance and see the distribution of the values within each feature. 00:38:00.420 --> 00:38:08.040 So it's just really helpful when you're doing predictive models, both for evaluating your own model and then eventually explaining it to end users. 00:38:09.000 --> 00:38:16.720 So would a wider spread mean that the feature is more useful or does it have any? 00:38:16.720 --> 00:38:17.260 Yeah. 00:38:17.260 --> 00:38:19.760 Especially if you can see a split in the numbers there. 00:38:19.760 --> 00:38:25.520 So you see in this example, relationship, you've got all the red ones to the right and all the blue ones to the left. 00:38:25.600 --> 00:38:31.820 That means that there's a clear relationship from this relationship field with the target variable. 00:38:31.820 --> 00:38:37.680 So there's a clear split where the low values are on one side and the high values are on the other side. 00:38:37.680 --> 00:38:48.060 And then, yeah, the spread means that if there's not a good example here, but sometimes you'll see like two clumps, two bee swarms spread apart with nothing in the middle. 00:38:48.520 --> 00:38:53.940 So that's when you have a really clear spread of the high impact group versus the low impact group. 00:38:53.940 --> 00:38:58.560 And if it's more narrow, that's less of an important variable. 00:38:58.560 --> 00:39:11.360 So you see if you look at the one that's sorted by max, here we go, absolute value of the shop value, the ones near the bottom for the population have less impact. 00:39:11.360 --> 00:39:18.260 Now there might be one person in there where that particular value was like the deciding factor of which class they ended up in. 00:39:18.260 --> 00:39:25.660 But for the population as a whole, there's less differentiation across these values than across the ones near the top of the list. 00:39:25.660 --> 00:39:30.220 Yeah, this is cool because that visualization of models is very tricky, right? 00:39:30.220 --> 00:39:32.780 And it's something like knowing why you got an answer. 00:39:32.780 --> 00:39:34.080 Now this looks very helpful. 00:39:34.080 --> 00:39:35.480 Yeah, it's really useful. 00:39:35.480 --> 00:39:40.260 And the visuals are so pretty by default, but then you can also pull those values into other tools. 00:39:40.260 --> 00:39:46.260 So for each feature in each row, you get a shop value. 00:39:46.260 --> 00:39:50.000 So you can write those back to the database and then pull those values into another tool. 00:39:50.000 --> 00:40:06.340 Like we use it in Tableau to highlight for each student, what are those really important features, either making them, you know, if you're not making them, it's not causation, but, you know, correlated with their more likely to retain or less likely to retain. 00:40:06.340 --> 00:40:10.160 So we might say, well, for this student, their GPA, that's the main factor. 00:40:10.160 --> 00:40:11.500 Their GPA is really low. 00:40:11.500 --> 00:40:13.840 Students with low GPAs tend not to retain. 00:40:13.840 --> 00:40:22.560 And so when the end user is looking at all of their values in, you know, in a table or some other kind of view, you can use the shop value to highlight. 00:40:22.560 --> 00:40:24.560 The GPA is the one you need to hone in on. 00:40:24.560 --> 00:40:26.280 The student is struggling academically. 00:40:26.420 --> 00:40:26.540 Right. 00:40:26.540 --> 00:40:28.980 Try to help them get some help with grades, for example. 00:40:28.980 --> 00:40:29.560 Cool. 00:40:29.560 --> 00:40:30.840 Yeah, this is a great find. 00:40:30.840 --> 00:40:31.360 Indeed. 00:40:31.360 --> 00:40:31.960 Indeed. 00:40:31.960 --> 00:40:32.320 Indeed. 00:40:32.320 --> 00:40:34.680 Brian, that brings us to the extras. 00:40:34.680 --> 00:40:35.520 Extras. 00:40:35.520 --> 00:40:36.460 Got any extras for us? 00:40:36.460 --> 00:40:37.140 I do. 00:40:37.140 --> 00:40:40.660 I've got one that was just a quick one. 00:40:40.660 --> 00:40:41.240 Let's see. 00:40:41.240 --> 00:40:42.060 Pull it up. 00:40:43.020 --> 00:40:49.080 Matthew Feigert mentioned on Twitter that pip index is a cool thing. 00:40:49.080 --> 00:40:52.240 And I kind of didn't know about it. 00:40:52.240 --> 00:40:53.020 So this is neat. 00:40:53.020 --> 00:40:57.440 So pip index is something you can take pip index. 00:40:57.440 --> 00:41:00.080 Well, specifically pip index versions. 00:41:00.080 --> 00:41:02.440 So pip index does a whole bunch of stuff. 00:41:02.660 --> 00:41:16.860 pip index versions will tell you, if you also give it a package, it'll tell you all the different versions that are available on PyPI and which one you have and which, you know, if you're out of date and stuff. 00:41:16.860 --> 00:41:25.800 But so for instance, if you're thinking about upgrading something and you don't know what to upgrade to, you can look to see what all is there, I guess. 00:41:25.800 --> 00:41:27.260 Or you want to roll it back. 00:41:27.260 --> 00:41:28.880 You're like, oh, this version is not working. 00:41:28.880 --> 00:41:30.880 I want to go back to a lower one. 00:41:31.020 --> 00:41:34.140 But if you're on 2.0, is it not 1.0? 00:41:34.140 --> 00:41:34.720 What is it, right? 00:41:34.720 --> 00:41:35.380 What do you go back to? 00:41:35.380 --> 00:41:37.500 And so this will like list all the available versions. 00:41:37.500 --> 00:41:44.140 Basically, this is a CLI version of the releases option in pypi.org. 00:41:44.140 --> 00:41:44.640 Right. 00:41:44.640 --> 00:41:48.380 And that's not like obvious how to get to on PyPI. 00:41:48.380 --> 00:41:50.780 But I know you can get to it. 00:41:50.780 --> 00:41:51.720 You can see all the releases. 00:41:51.720 --> 00:41:53.860 But by default, it doesn't show those. 00:41:53.860 --> 00:41:55.160 So this is pretty quick. 00:41:55.160 --> 00:41:55.640 Yeah. 00:41:55.640 --> 00:41:56.280 Pretty neat. 00:41:56.280 --> 00:41:57.080 Good one. 00:41:57.080 --> 00:41:57.500 Good one. 00:41:57.500 --> 00:41:58.500 Renee, how about you? 00:41:58.500 --> 00:41:59.000 Some extras? 00:41:59.000 --> 00:41:59.860 Sure. 00:41:59.860 --> 00:42:01.820 I wanted to make sure to mention my book. 00:42:01.820 --> 00:42:06.820 So just published and just out in Europe this week, actually, as a paperback, but it's been 00:42:06.820 --> 00:42:08.160 out since September in the US. 00:42:08.160 --> 00:42:12.540 SQL for Data Scientists, A Beginner's Guide for Building Data Sets for Analysis. 00:42:12.540 --> 00:42:18.540 So I mentioned earlier, you know, I wrote this book because I think a lot of students coming 00:42:18.540 --> 00:42:25.360 out of data science programs or people who are coming from maybe statistics background that 00:42:25.360 --> 00:42:28.580 are in data science might not have experience pulling the data. 00:42:28.580 --> 00:42:32.680 So in class, a lot of times you're given a clean spreadsheet to start with when you're building 00:42:32.680 --> 00:42:33.580 your predictive model. 00:42:33.580 --> 00:42:37.580 Then you get to the job and, you know, you sit down the first day and they say, great, build 00:42:37.580 --> 00:42:38.080 us a model. 00:42:38.080 --> 00:42:39.460 And you say, well, where's the data? 00:42:39.780 --> 00:42:41.880 It's in a raw form in the database. 00:42:41.880 --> 00:42:43.880 So you have to build your own data set. 00:42:43.880 --> 00:42:47.340 So that's what the purpose of the book is, to kind of get you from that point of when you 00:42:47.340 --> 00:42:52.260 have access to raw data to exploring and building your data set so that you can run it through 00:42:52.260 --> 00:42:53.280 your predictive model. 00:42:53.280 --> 00:42:56.100 So on the screen there, you see my website. 00:42:56.100 --> 00:43:00.280 And for people on the audio, it's SQL for datascientist.com. 00:43:01.020 --> 00:43:04.800 And you can go to different chapters on the website. 00:43:04.800 --> 00:43:07.480 And I have some example SQL, and you can also run it. 00:43:07.480 --> 00:43:11.220 So there's a SQLite database in the browser here. 00:43:11.220 --> 00:43:17.060 And so you can actually copy and paste some of the SQL on the page, click execute, and it 00:43:17.060 --> 00:43:18.500 shows up in a table down here. 00:43:18.500 --> 00:43:20.160 You can edit it and rerun it. 00:43:20.160 --> 00:43:24.100 So you get a little bit of practice with this, using the database in the book. 00:43:24.100 --> 00:43:24.560 Neat. 00:43:24.560 --> 00:43:25.360 Yeah. 00:43:25.360 --> 00:43:26.320 Cool book. 00:43:26.320 --> 00:43:28.220 And wow, SQLite in the browser. 00:43:28.220 --> 00:43:28.800 Very neat. 00:43:28.800 --> 00:43:29.780 Thank you. 00:43:29.780 --> 00:43:30.220 Yeah. 00:43:30.220 --> 00:43:30.880 Awesome. 00:43:31.000 --> 00:43:32.920 It's a book that definitely should exist. 00:43:32.920 --> 00:43:33.440 All right. 00:43:33.440 --> 00:43:37.160 Really quickly, I'm going to do a webcast with Paul Everitt. 00:43:37.160 --> 00:43:39.180 I haven't seen Paul in the audience today. 00:43:39.180 --> 00:43:39.840 Paul, where are you? 00:43:39.840 --> 00:43:40.740 No, I'm not sure. 00:43:40.740 --> 00:43:41.720 He might be working. 00:43:41.720 --> 00:43:46.220 But on November 23rd, I'm going to be doing a webcast around PyCharm. 00:43:46.220 --> 00:43:50.680 I've updated my PyCharm course with all sorts of good stuff. 00:43:50.680 --> 00:43:54.240 I haven't quite totally announced it yet because there's a few things I'm waiting, slightly 00:43:54.240 --> 00:43:59.160 more stable versions to come out of JetBrains to finish some of their data science tools, 00:43:59.160 --> 00:43:59.520 actually. 00:43:59.940 --> 00:44:01.480 And then I'll talk more about it. 00:44:01.480 --> 00:44:04.000 But we're doing a webcast in about a week or so. 00:44:04.000 --> 00:44:05.340 So that should be a lot of fun. 00:44:05.340 --> 00:44:07.820 And yeah, come check it out. 00:44:07.820 --> 00:44:10.200 Watch Paul and me make the code go. 00:44:10.200 --> 00:44:11.700 Two days before Thanksgiving. 00:44:11.700 --> 00:44:13.300 Yes, indeed. 00:44:13.300 --> 00:44:13.960 All right. 00:44:13.960 --> 00:44:15.160 That brings us to our joke. 00:44:15.160 --> 00:44:16.320 I need a joke. 00:44:16.660 --> 00:44:21.560 And the joke is a response to something that you posted on Twitter. 00:44:21.560 --> 00:44:27.040 Really appreciate my foresight of using lots of stuff comment as the message in Git commit. 00:44:27.040 --> 00:44:31.460 Well, it actually confused me because I did a Git rebase main and it said applying lots of 00:44:31.460 --> 00:44:31.760 stuff. 00:44:31.760 --> 00:44:37.260 And I thought it was a feature of Git rebase and it just happened to be my commit message. 00:44:38.260 --> 00:44:38.620 Yeah. 00:44:38.620 --> 00:44:40.120 It's like, oh, Git's gotten real casual. 00:44:40.120 --> 00:44:41.040 It's a lot of stuff. 00:44:41.980 --> 00:44:47.820 So Francois Voron said, time for a classic XKCD link here. 00:44:47.820 --> 00:44:48.420 Yeah. 00:44:48.420 --> 00:44:48.760 Oh, yeah. 00:44:48.760 --> 00:44:49.420 Yeah. 00:44:49.640 --> 00:44:55.860 And so this is like the commit history throughout the project as you get farther into it. 00:44:55.860 --> 00:45:01.380 So it starts out with very formal, proper comments like created main loop and timing control. 00:45:01.380 --> 00:45:04.640 The next commit is enabled config file parsing. 00:45:04.640 --> 00:45:09.800 And then starts to miscellaneous bug fixes and then code additions and edits. 00:45:09.800 --> 00:45:12.140 And then a branch, more code. 00:45:12.140 --> 00:45:16.220 You have code, just eight letter A's. 00:45:16.220 --> 00:45:18.400 It comes back with screaming. 00:45:18.880 --> 00:45:19.320 Exactly. 00:45:19.320 --> 00:45:22.080 Just A-D-K-F-J-S-L-K. 00:45:22.080 --> 00:45:23.220 Just a bunch of home row. 00:45:23.220 --> 00:45:25.700 And then my hands are typing words. 00:45:25.700 --> 00:45:27.040 And then just hands. 00:45:27.040 --> 00:45:33.400 And the title is, as a project drags on, my Git commit messages get less and less informative. 00:45:33.400 --> 00:45:34.480 We've all been there, right? 00:45:34.480 --> 00:45:34.900 Yeah. 00:45:34.900 --> 00:45:35.580 Yes. 00:45:35.580 --> 00:45:39.540 It happens to me with branch names too. 00:45:39.540 --> 00:45:44.320 Because if I'm working on one feature and push part of it and then I go and I'm still working 00:45:44.320 --> 00:45:46.920 on it, I like to use a new branch name. 00:45:47.400 --> 00:45:51.840 And I just, I can't, it's hard to come up with good branch names for a feature. 00:45:51.840 --> 00:45:52.700 I'm branching. 00:45:52.700 --> 00:45:54.760 Exactly. 00:45:54.760 --> 00:45:59.440 I try, I try to be more formal on the branches at least so I know I can delete them later. 00:45:59.440 --> 00:46:04.540 And so when I'm working on projects that are mostly just me, I'll create a GitHub issue 00:46:04.540 --> 00:46:08.820 and then create the branch name to be like a short version of the issue title and the issue 00:46:08.820 --> 00:46:09.340 number. 00:46:09.760 --> 00:46:13.520 So then when I commit back in, I can just look at the branch name and put a hash that 00:46:13.520 --> 00:46:17.200 number and it'll tag it in the commit on the issue and GitHub. 00:46:17.200 --> 00:46:17.760 Yeah. 00:46:17.760 --> 00:46:23.260 If I'm working with someone like a team, I might put like my name slash branch name. 00:46:23.260 --> 00:46:26.920 And then actually in some of the tools like source tree, you have like little expando widgets 00:46:26.920 --> 00:46:28.760 around that on the branches. 00:46:28.760 --> 00:46:32.500 So you can say these are Michael's branches and these are Renee's branches and so on. 00:46:32.740 --> 00:46:32.920 Yeah. 00:46:32.920 --> 00:46:34.340 We got into the habit of doing that too. 00:46:34.340 --> 00:46:37.440 It helps a lot to see right up front whose branch is this? 00:46:37.440 --> 00:46:38.020 Yeah. 00:46:38.020 --> 00:46:39.420 I can get, can get out of control. 00:46:39.420 --> 00:46:39.660 Right. 00:46:39.660 --> 00:46:40.220 All right. 00:46:40.220 --> 00:46:41.920 A quick couple follow-outs, Brian. 00:46:41.920 --> 00:46:44.840 Anthony says, the book looks great, Renee. 00:46:44.840 --> 00:46:45.640 I'll check it out. 00:46:45.640 --> 00:46:46.160 Great. 00:46:46.160 --> 00:46:46.620 Thank you. 00:46:46.620 --> 00:46:48.480 Chris May likes it as well. 00:46:48.480 --> 00:46:49.280 It's a great book idea. 00:46:49.280 --> 00:46:50.480 I do like it. 00:46:50.480 --> 00:46:53.360 Especially when I keep working long after I should have gone home. 00:46:53.360 --> 00:46:54.660 No, this is the joke. 00:46:54.660 --> 00:46:55.200 This is me. 00:46:55.200 --> 00:46:58.500 Especially after I keep working long after I should have gone home and gone home. 00:46:58.500 --> 00:46:59.240 Yeah, absolutely. 00:46:59.840 --> 00:47:04.720 And Sam, oops, forgot to stage this as a common message in my repositories. 00:47:04.720 --> 00:47:06.080 Nice. 00:47:06.080 --> 00:47:06.440 Indeed. 00:47:06.440 --> 00:47:07.900 So, cool. 00:47:07.900 --> 00:47:09.560 It was a fun episode. 00:47:09.560 --> 00:47:10.860 Thanks, Renee, for coming on. 00:47:10.860 --> 00:47:12.100 Yeah, thanks for having me. 00:47:12.100 --> 00:47:12.480 It's fun. 00:47:12.480 --> 00:47:15.260 Don't get to dive into Python too often. 00:47:15.260 --> 00:47:17.320 I mean, I'm using the same type of things over and over. 00:47:17.320 --> 00:47:19.760 So it's nice to see what's new and what's on the horizon. 00:47:19.760 --> 00:47:20.500 Awesome. 00:47:20.500 --> 00:47:20.800 Yep. 00:47:20.800 --> 00:47:21.400 Thanks, Mike. 00:47:21.400 --> 00:47:22.080 Thanks, Brian. 00:47:22.080 --> 00:47:22.660 Bye. 00:47:22.660 --> 00:47:22.720 Bye. 00:47:22.720 --> 00:47:22.740 Bye. 00:47:22.740 --> 00:47:22.800 Bye. 00:47:22.800 --> 00:47:23.400 Bye.