WEBVTT 00:05.230 --> 00:06.840 Hi everyone welcome back. 00:06.840 --> 00:11.720 So in this book you are going to be starting with the classification class which is really where all 00:11.740 --> 00:12.830 the heavy work goes. 00:12.840 --> 00:13.530 OK. 00:13.710 --> 00:16.710 So first of all we need to talk about a couple of things. 00:16.710 --> 00:24.090 OK so we'll going to first open Firefox and we're going to go to a link so we're going to write Tancer 00:24.120 --> 00:28.920 flow and then we're going to write in section v3. 00:28.950 --> 00:29.590 OK. 00:29.940 --> 00:31.660 So inception version 3. 00:31.770 --> 00:32.580 OK. 00:32.670 --> 00:35.010 When you write it you should see the first link here. 00:35.010 --> 00:41.820 Image recognition if you click on it and then go inside there you to see that you know they give you 00:41.820 --> 00:43.710 a model and then they talk to you. 00:43.800 --> 00:49.890 If you go down to use you as an API then you're going to see that there is a file called classify underscore 00:49.890 --> 00:51.210 image doppio. 00:51.450 --> 00:56.670 And it's very simple you're just going to go into this link so it's going to tell you go to this link 00:56.670 --> 00:59.740 start by cloning this links are going to click on this link. 01:00.000 --> 01:05.340 It's going to open a good repository with all the models of the answer flow and then what you're going 01:05.340 --> 01:11.840 to go is you're going to go to models tutorials image and then Image it. 01:11.880 --> 01:15.590 So we're going to go come here and then you're going to this is already on Mars. 01:15.600 --> 01:18.520 If you look at the or there's already models here. 01:18.540 --> 01:21.940 So what you're going to do is let me try to make this bigger. 01:22.020 --> 01:22.870 You're going to go. 01:22.920 --> 01:25.470 Let's see if we can make a slightly bigger Yes. 01:25.530 --> 01:26.440 There we go. 01:26.620 --> 01:28.560 I'm sure you're going to go into tutorials. 01:29.400 --> 01:35.960 And then we're going to go into image and then Image net and then you should see or classify underscore 01:36.020 --> 01:37.070 and Match.com. 01:37.220 --> 01:42.390 OK so we're going to click on this year we want to copy all of this file files are going to take this 01:42.390 --> 01:44.970 file from start go all the way up. 01:45.030 --> 01:50.110 Copy all this file because this is the file that we're going to be using to classify them. 01:50.150 --> 01:50.620 Great. 01:50.700 --> 01:54.040 So I'm going to copy this file and then I'm going to come here. 01:54.210 --> 01:57.530 Can I see this year and make a new document. 01:57.660 --> 02:07.070 Paste this and then control s go into desktop and then into misclassification web. 02:07.200 --> 02:11.210 And then I'm going to save it here as classify. 02:11.520 --> 02:18.640 So underscore image wise I've just downloaded this file and saved it here and that's it. 02:18.660 --> 02:23.820 OK so this model is supposed to be able to train an image that you give it. 02:23.820 --> 02:27.600 And I'm going to show you how you can use it in this video. 02:27.600 --> 02:28.200 OK. 02:28.470 --> 02:34.340 So first of all the thing that we were going to do is we just look. 02:34.370 --> 02:40.670 So first of all I want I want us to look at the file and just get a feel of how this is working. 02:40.710 --> 02:43.000 Before we actually use the file OK. 02:43.350 --> 02:48.780 So the first thing is when you call this file it first get some arguments OK. 02:48.810 --> 02:51.790 So you pass some arguments into this file. 02:51.810 --> 02:54.990 So the first argument is the model directory. 02:54.990 --> 02:58.460 So you need to download it where you download the model directory from. 02:58.470 --> 02:59.090 OK. 02:59.190 --> 03:01.730 So this model is what we're going to download. 03:01.740 --> 03:02.090 OK. 03:02.100 --> 03:09.150 So this model is if you go here if you go here you're going to find there is a right or in the first 03:09.150 --> 03:09.390 link. 03:09.390 --> 03:15.660 So if you go back you're in this link here we're going to find full data underscore your else. 03:15.660 --> 03:17.060 Let me just zoom in here. 03:18.240 --> 03:19.700 So go down. 03:19.860 --> 03:25.110 You're going to find that your ally here and this is where the model is going to come from. 03:25.260 --> 03:27.420 So we're going to copy this thing here. 03:27.540 --> 03:28.190 OK. 03:28.230 --> 03:34.740 Because we want to download the model locally and then we're going to go here and then paste and then 03:34.740 --> 03:35.770 click enter. 03:36.090 --> 03:38.670 Now it should tell you to save up to save the model. 03:38.670 --> 03:41.120 So the inception model is very big. 03:41.250 --> 03:47.000 It's actually if not super big race 85 megabytes or relatively big. 03:47.160 --> 03:48.610 So you're going to save the file. 03:48.630 --> 03:55.050 I already downloaded it so I won't download it again but once you download this file then you're going 03:55.050 --> 03:56.070 to go. 03:56.420 --> 04:00.590 So you're going to close that and then you're going to go to where you downloaded it. 04:00.630 --> 04:04.750 So in my case I've downloaded the model on the desktop so you can see it here. 04:04.760 --> 04:06.380 Inception here. 04:06.480 --> 04:07.770 So I'm going to cut this. 04:07.770 --> 04:10.970 And then I copy it going to image classification. 04:11.280 --> 04:16.780 And then what's so image classification wrongful there I guess. 04:16.890 --> 04:18.910 So image classification folder. 04:19.170 --> 04:21.800 And then I'm going to go into web and then I'm going to paste it here. 04:21.810 --> 04:23.010 So I'm going to leave it here. 04:23.040 --> 04:23.670 OK. 04:23.770 --> 04:27.350 So paste it paste this thing here and that's it. 04:27.360 --> 04:28.170 That's it for me. 04:28.170 --> 04:33.410 So I'm just going to leave the model here and you can see why I've left it here in a few seconds. 04:33.780 --> 04:35.190 OK so what is this doing. 04:35.400 --> 04:37.080 Let's go back to the arguments. 04:37.270 --> 04:39.010 Let's let's zoom out a bit. 04:39.150 --> 04:42.580 So first of all we need a model directory so it's selling. 04:42.630 --> 04:48.970 We need to tell tensor flow which model to use to predict what classification the images. 04:49.020 --> 04:54.960 So what model is going extensor flow going to use to predict whether this is an animal or a zebra or 04:54.960 --> 04:56.190 a giraffe or so on. 04:56.190 --> 04:56.980 OK. 04:57.390 --> 05:03.210 The second argument is image files so you need to give this file of your an argument which is image 05:03.210 --> 05:03.660 file. 05:03.810 --> 05:11.250 And this image files that you are to an image where where it can infer what or classify this image. 05:11.250 --> 05:13.440 This is the image that we want to classify. 05:13.440 --> 05:13.710 Right. 05:13.710 --> 05:16.480 This is the image that we're going to be sending out. 05:16.770 --> 05:19.870 And then also number of predictions. 05:19.920 --> 05:22.710 So how many predictions do you want. 05:22.710 --> 05:28.330 So for example if you remember when we were talking about the API that is going up to give you the top 05:28.330 --> 05:29.960 five predictions. 05:29.950 --> 05:36.690 So for example it might give you the first one is 80 percent the second one is 5 percent. 05:36.690 --> 05:40.530 The second 31st percent 2 percent 1 percent and so on. 05:40.530 --> 05:40.920 Right. 05:40.980 --> 05:47.880 So it's going to give you the percentages but in descending order of how positive it is off this prediction 05:48.840 --> 05:55.340 and then another thing is going to parse all these arguments and put it in flags. 05:55.350 --> 06:00.520 So it's going to put flet into it into flags and then it's going to say denser flow. 06:00.540 --> 06:02.980 Please run the function main. 06:03.030 --> 06:03.550 OK. 06:03.630 --> 06:05.520 So let's go back to the function menu. 06:05.520 --> 06:07.220 So this is the function main. 06:07.620 --> 06:12.370 So first of all it's going to be to call the function maybe download and extract. 06:12.450 --> 06:18.330 So this function what this is what it was what it's going to do is if you have the model this inception 06:18.330 --> 06:20.490 model then it's not going to do anything. 06:20.520 --> 06:21.760 It's just going to return. 06:22.020 --> 06:27.700 But if you don't there is going to use this data you are all here to download them all. 06:27.840 --> 06:29.740 And that is going to return. 06:29.760 --> 06:30.440 OK. 06:30.480 --> 06:33.820 So in other words even if I remove this. 06:33.840 --> 06:38.530 So let's go see you in Maine here even if I remove this line. 06:38.640 --> 06:40.050 This file would still work. 06:40.060 --> 06:40.250 Right. 06:40.260 --> 06:46.220 But I'm just going to leave now the second thing is going to find that you are EHLO that image. 06:46.220 --> 06:51.510 So it's going to get to you or of that image that you want to classify. 06:51.600 --> 06:55.070 And then it's going to call the function run in friends an image. 06:55.080 --> 06:55.600 OK. 06:55.650 --> 06:57.360 So it's going to inference. 06:57.390 --> 07:02.890 So if you have an image and we want to do some inference on you know what's inside of it. 07:02.910 --> 07:10.560 So if I go to run interference on and let's say if I go to the function it says run introns an image 07:10.560 --> 07:14.820 so it takes an argument or an image file and then it returns nothing. 07:14.900 --> 07:17.740 OK so what is going to do is do. 07:18.000 --> 07:20.430 It's going to do some machinery new stuff. 07:20.450 --> 07:26.970 So we're not going to go into too much details since everyone might know what machine learning is or 07:27.210 --> 07:27.840 how to use. 07:27.840 --> 07:32.080 Actually the answer to most people no machinery used but how to use underflow. 07:32.280 --> 07:38.670 But it's basically going to start predicting what this image might be and then if you come down here 07:38.670 --> 07:43.290 you're going to say top k is equal to predictions are X or it. 07:43.440 --> 07:47.080 So it's going to get to sort of prediction in descending order. 07:47.100 --> 07:54.390 So it gives you the best the best prediction for example 80 percent prediction a panda or a zebra. 07:54.510 --> 07:56.070 Then it's rancid. 07:56.070 --> 08:05.700 The number one right so cupcake is going to be two to sort the predictions that the model here predicted 08:06.150 --> 08:13.380 and then is going to go for each one of them is going to score is equal to prediction of ID and then 08:13.380 --> 08:22.360 it's going to pre-print the human strain so the human strain here is basically zebra or that's a giraffe 08:22.450 --> 08:24.980 or human or Angelina Jolie. 08:25.030 --> 08:30.000 So the mystery here is what this model thinks this image is off. 08:30.130 --> 08:33.880 And then the score is the percentage of how confident it is. 08:33.970 --> 08:41.290 So for example I'm 30 percent sure that this is Angelina Jolie or I'm 40 percent sure that this is bracket. 08:41.320 --> 08:41.920 OK. 08:42.070 --> 08:49.520 So we it's going to print what what it thinks the prediction is and how confident it is off this prediction. 08:49.520 --> 08:51.460 OK for each of the predictions. 08:51.480 --> 08:52.370 All right. 08:52.700 --> 08:54.450 Hopefully this makes the sense. 08:54.550 --> 09:01.670 So when we're writing the classify function what we're going to do is actually we want to get this top. 09:02.500 --> 09:05.400 Dictionary that has all the predictions right. 09:05.500 --> 09:10.720 So we got what we're going to do is we're going to you're right in this fight we're going to call this 09:10.720 --> 09:11.300 fight. 09:11.360 --> 09:11.750 OK. 09:11.800 --> 09:13.910 We're going to call this a subprocess. 09:14.140 --> 09:17.440 And given the image file and give it all the other stuff. 09:17.680 --> 09:25.420 And then this file is going to predict the what this image is and that it's going to write in this directory 09:25.420 --> 09:26.130 over here. 09:26.150 --> 09:29.930 It's going to write a file with a dictionary. 09:30.520 --> 09:38.170 Then once this happens in Abu Dhabi why I'm going to read this file which has the predictions and then 09:38.200 --> 09:40.580 I'm going to load it into a dictionary. 09:40.770 --> 09:43.020 So you're going to see what I mean by that. 09:43.030 --> 09:44.930 And if just a few seconds. 09:44.950 --> 09:46.640 So first of all let's write the class. 09:46.640 --> 09:50.710 So class classify resource. 09:50.770 --> 09:57.280 So what I did just was I wanted to make sure that you understand even a very high logic of what's happening 09:57.280 --> 09:58.110 in this file. 09:58.150 --> 10:03.230 So because we're going to come here and we're going to be hitting this area over here in a few minutes. 10:03.250 --> 10:04.300 OK. 10:04.480 --> 10:07.210 So the first thing is we have a class called classified. 10:07.210 --> 10:10.990 So what are we at what what does it support what method does it support. 10:10.990 --> 10:15.550 Post something on his own defined post self. 10:16.750 --> 10:19.680 And then first of all we want to get the posted data as usual. 10:19.690 --> 10:22.970 So you can your request that get your score. 10:23.080 --> 10:25.420 So that's the first thing. 10:25.460 --> 10:31.910 And then the second thing is we're to get the user name is equal to hosted data username. 10:32.050 --> 10:33.870 Right. 10:33.940 --> 10:41.680 And the password of password is equal to posted data and then password right. 10:42.080 --> 10:48.570 So now the third thing we need is you or else we're going to need your elders to posted data. 10:48.880 --> 10:53.110 But then this time is going to be you or L or write code. 10:53.110 --> 10:55.950 So now we have everything that we need. 10:56.170 --> 11:00.730 Now the first thing is well we need to verify the credentials of this user. 11:00.730 --> 11:01.520 OK. 11:01.630 --> 11:07.600 So for this I'm going to write a function just to make this it just do not clutter this area I'm going 11:07.600 --> 11:13.180 to write a function called verify credentials that is given a username and password and it tells me 11:13.180 --> 11:19.210 if there is any error or not and if there is an error it returns to me the return Jaisalmer that I am 11:19.210 --> 11:19.960 going to retrieve. 11:20.170 --> 11:30.360 So when I do return some error is equal to verify credentials username and password. 11:31.550 --> 11:32.420 OK. 11:32.880 --> 11:40.980 So if air is true so if there is an heir so the user name doesn't exist or your name doesn't match the 11:40.980 --> 11:43.890 password then this is going to be true. 11:43.980 --> 11:46.580 And this is going to have the return Jaser. 11:47.010 --> 11:49.300 But there is this is going to be false. 11:49.350 --> 11:50.690 And this is going to be none. 11:50.940 --> 11:51.620 OK. 11:51.900 --> 11:59.410 So I'm not to say if air then returned the sun if I returned Jason. 11:59.430 --> 12:00.100 OK. 12:00.420 --> 12:03.120 So so far I haven't written to verify credentials. 12:03.210 --> 12:04.760 Right so I can put here. 12:04.830 --> 12:10.090 I asked AG to do OK because I haven't done it. 12:10.160 --> 12:11.040 All right. 12:11.160 --> 12:16.590 So the second thing is we need to verify that the user had enough tokens right. 12:16.660 --> 12:23.550 So I say tokens is equal to users that's fine that's fine. 12:23.790 --> 12:31.110 And then I want to find the username with the username that is posted and then got access the first 12:31.110 --> 12:33.510 ones since there is only one of the Great. 12:33.660 --> 12:37.640 And then I'm also going to access the element tokens right. 12:38.620 --> 12:45.980 All right then I'm going to check if tokens is less than or equal to zero that what I want to do is 12:45.980 --> 12:49.650 to return an error of 0 3 not enough tokens right. 12:49.670 --> 12:53.570 You remember from here three or three out of tokens. 12:53.570 --> 12:59.840 So the way I'm going to do this instead of having to do a return Jasen equal to this and then status 13:00.200 --> 13:05.570 and then message right instead of having to do all this for every time I want to retune something I'm 13:05.570 --> 13:07.180 going to write the helper function. 13:07.400 --> 13:08.040 OK. 13:08.250 --> 13:15.810 So I'm going to say If tokens is less than or equal to zero then return Jaison if I. 13:16.310 --> 13:17.030 OK. 13:17.250 --> 13:23.320 And then I'm going to write a new function called let's see let's call it generate return to action. 13:23.430 --> 13:32.780 We generate return dictionary and he's going to have two arguments status and comment. 13:32.830 --> 13:33.320 OK. 13:33.370 --> 13:40.420 So for here I want us I want the status to be three of eight and for the comment to be let's say not 13:40.600 --> 13:43.470 enough tokens. 13:44.360 --> 13:44.960 Right. 13:45.130 --> 13:50.410 So what this is going to do is to generate a dictionary for me with this and this and this message and 13:50.410 --> 13:53.810 return it and then I'm going to just have to fight and return. 13:53.820 --> 13:57.470 So instead of having to do it on multiple lines we can just do it on one line. 13:57.470 --> 13:57.760 Right. 13:57.790 --> 13:59.500 Makes the code much shorter. 13:59.830 --> 14:01.540 So there's also a to do here. 14:01.540 --> 14:06.840 So let's put a hash tag to do here because we still haven't written this function. 14:07.420 --> 14:11.690 So the next thing we're going to do is we want to get the image OK from the euro. 14:11.820 --> 14:13.710 So not to say R is equal to. 14:13.720 --> 14:16.330 So this is how we use requests of an s. 14:16.480 --> 14:19.920 So R is equal to your requests don't get. 14:20.020 --> 14:21.930 And then I'm going to put the U R L in there. 14:21.970 --> 14:22.630 OK. 14:22.900 --> 14:23.620 So I'm going to get. 14:23.620 --> 14:25.960 What's the content of this you are. 14:26.330 --> 14:31.350 And then I'm going to say returned J is equal to just an empty. 14:31.540 --> 14:38.570 So I'm just going to leave that as an empty dictionary for now that I'm going to say with open. 14:38.620 --> 14:46.300 So I'm going to open 10 that JP J Gates I'm going to just store a temporary image and then it's going 14:46.300 --> 14:52.340 to be W. B as F as efforts and then I'm going to do. 14:52.380 --> 14:53.570 F dot right. 14:53.600 --> 14:55.650 So I'm going to write it too. 14:56.050 --> 15:00.190 I want to write into f the content of R. 15:00.280 --> 15:07.750 OK so are has the image that I want then I'm going to open a new file called template that JP G. 15:07.870 --> 15:11.530 I'm going to write into the content of art which is the image itself. 15:11.560 --> 15:12.110 OK. 15:12.250 --> 15:18.000 So that's how simple it is to download an image in Python all right once I've done that then I'm going 15:18.000 --> 15:20.070 to do process is equal to. 15:20.070 --> 15:21.430 So I'm going to say process. 15:21.450 --> 15:27.860 I want to do a new process so I'm going to say some Prozess dot be open. 15:27.870 --> 15:38.040 So this is the syntax open and then Python classify underscore image Doppie y. 15:38.050 --> 15:39.730 So this is the file over here. 15:39.840 --> 15:47.220 So I'm going to open a new subprocess with this file Paice and classify underscore image not quite and 15:47.260 --> 15:52.430 that I'm going to pass the argument that the firewall. 15:52.470 --> 15:58.300 So the first one was models underscore or I think it was. 15:58.380 --> 15:59.220 Let's check. 15:59.220 --> 16:03.890 So it was modeled on your score directorates. 16:03.930 --> 16:08.200 No models it's a model underscore it or tree is equal to. 16:08.250 --> 16:14.680 And then I'm going to write dot because the font the the the the inception file as in this year this 16:14.680 --> 16:15.120 folder. 16:15.130 --> 16:27.190 So it's in the current folder and the image so the image file is in this folder and it's at dot slash 16:27.300 --> 16:29.580 10 and the GBG right. 16:29.590 --> 16:34.590 Because if you remember we downloaded the image and we store it in temporary GBG. 16:34.630 --> 16:39.590 So now what we have is the image file is stored in dot slash stamped GMP. 16:39.670 --> 16:40.300 OK. 16:41.730 --> 16:44.080 Now we're going to do. 16:44.080 --> 16:46.960 Process dot communicate. 16:47.210 --> 16:48.790 OK. 16:49.070 --> 16:50.200 Zero. 16:50.490 --> 16:51.900 And then we're going to process that. 16:51.930 --> 16:52.980 Wait. 16:53.490 --> 16:56.160 So we're going to wait until this process here is done. 16:56.180 --> 17:02.420 And this final and this subprocess writes the response Jason in a file. 17:02.430 --> 17:04.580 Let's say it's called text. 17:04.730 --> 17:05.670 Yes Steve. 17:05.790 --> 17:06.270 OK. 17:06.360 --> 17:07.680 So don't wait for it. 17:07.740 --> 17:12.630 And then now I want to get the of the the open the file that they have. 17:12.630 --> 17:13.110 OK. 17:13.190 --> 17:17.700 So I'm going to do with Open Text DXi. 17:17.760 --> 17:23.930 So what what writes this file is actually in this subprocesses over here that opens this file. 17:23.940 --> 17:27.660 So with open text that DST as. 17:27.810 --> 17:33.140 And then for example or let's say as G. 17:33.870 --> 17:41.000 And then we're going to do return J-Zone is equal to Jaison load and then G. 17:41.040 --> 17:41.510 OK. 17:41.550 --> 17:51.750 So I'm going to load the the the the dictionary that I stored into the the file text of the XTi And 17:51.750 --> 17:56.340 then finally once since I have the return adjacent here so I now have it. 17:56.340 --> 18:04.770 So once I exit out of all of this I'm just going to say Well first of all I want to return the return 18:04.770 --> 18:06.190 Jason to the user. 18:06.450 --> 18:11.180 But before I do this I need to take away one token from me because I used my service. 18:11.280 --> 18:22.890 So when I do users that update users update and then inside here we're going to do an user name and 18:22.890 --> 18:24.880 then it's going to be username. 18:24.960 --> 18:31.350 And what I'm going to update them with is I'm going to set his tokens to be one less than what he has 18:31.380 --> 18:33.980 and the new dollar set. 18:34.020 --> 18:38.380 And then hear him say tokens are going to be tokens. 18:38.400 --> 18:39.970 The variable minus one. 18:40.050 --> 18:42.020 So now he has one less pair. 18:42.030 --> 18:44.860 So you can see that we already got the tokens right. 18:44.910 --> 18:47.440 So he has one less token to worry about. 18:47.490 --> 18:48.760 OK. 18:49.030 --> 18:54.810 And now once we have done that we just returned to him the response which is generated by this file 18:54.810 --> 18:55.680 over here. 18:55.950 --> 19:02.910 Now let's go to this file and now we want to change this file to reflect what what this users actually 19:03.630 --> 19:04.860 is expecting right. 19:04.860 --> 19:10.940 So how do we write this returned Jason into the final text of the XTi. 19:11.070 --> 19:16.340 So we're going to go up here and then you can see here that there is a top Caze equal to predictions 19:16.360 --> 19:19.150 done on X or it slides a number addiction. 19:19.170 --> 19:19.600 OK. 19:19.800 --> 19:25.130 So after that I'm going to write return Jason is equal to an empty dictionary ok. 19:25.620 --> 19:34.690 Then inside this loop here I'm going to say return J-Zone human unstring is equal to a score. 19:34.710 --> 19:37.630 OK so for each you ministering I'm going to give it a score. 19:37.860 --> 19:38.670 All right. 19:40.000 --> 19:42.580 And then once I'm done with this process. 19:42.580 --> 19:49.960 So once I've generated this return Jason what I'm going to do is I'm going to store that or dump that 19:50.140 --> 19:51.560 into the text. 19:51.790 --> 19:52.530 64. 19:52.630 --> 20:05.310 So I'm going to do with Open Text dx d as F and then I'm going to do Jaison that dump and then return 20:05.310 --> 20:05.530 J. 20:05.530 --> 20:10.480 So I'm going to return this I'm going to dump this returned Jason and that I filled with that with the 20:10.480 --> 20:17.650 results and then I'm going to do Also a half into after the file text of DST and that's it. 20:17.650 --> 20:26.730 So we've just added a couple a few lines and then basically it is written into a file called texta DXi. 20:26.890 --> 20:33.390 Then from here what happens is after I run this command this command finishes so after post is dead. 20:33.400 --> 20:41.220 Wait what i do is I open the file text Dickstein that this file here has written and then I get visé 20:41.270 --> 20:41.740 Sunfire. 20:41.750 --> 20:49.250 So this way I've sent it there result from this classify image file into my file apt up yours. 20:49.330 --> 20:50.450 OK. 20:50.830 --> 20:56.950 So hopefully this made sense I wasn't and wasn't you know too difficult to understand. 20:56.950 --> 20:59.070 Now there's a couple of things that we need to do first. 20:59.110 --> 21:03.830 We have a couple of to do is right so there's a few functions that we haven't done. 21:03.910 --> 21:06.690 So the first one is to verify credentials. 21:06.700 --> 21:10.370 So why do you fine verify credentials. 21:10.390 --> 21:13.690 We have a username and password. 21:14.050 --> 21:14.790 OK. 21:17.940 --> 21:24.540 OK so how is the function going to look like someone to say first if not use or exists or it doesn't 21:24.540 --> 21:27.680 exist then what are we going to do. 21:27.720 --> 21:32.950 We're going to return generate return dictionaries. 21:32.990 --> 21:33.830 I'm going to. 21:33.990 --> 21:39.630 So I'm going to return this generate return dictionary and when I generate a new dictionary with the 21:39.630 --> 21:49.680 air feel one and the message invalid username username and then I'm also going to return true because 21:49.680 --> 21:51.940 remember this function returns two things. 21:52.020 --> 21:54.790 It returns a dictionary and true right. 21:54.810 --> 21:56.780 The air here is going to be true. 21:57.190 --> 22:02.290 OK next I'm also going to do the correct password correct. 22:02.380 --> 22:06.120 Underscore password is equal to verify. 22:07.010 --> 22:07.770 Password. 22:07.820 --> 22:13.910 So I'm gonna do a new function called the verify password that I'll give the user name user name and 22:13.910 --> 22:19.360 the password and he's going to tell me whether this user name matches the password or not. 22:19.530 --> 22:22.630 Then unless if not correct password. 22:22.790 --> 22:23.820 Then what do I do. 22:24.020 --> 22:25.450 Well I return. 22:25.730 --> 22:33.800 Generate return dictionary of 3 0 2 and then invalid password. 22:35.540 --> 22:38.120 And also true because there's an error here. 22:38.380 --> 22:39.100 OK. 22:39.380 --> 22:45.740 What is the password is correct and the user exists then that means that I there's no errors I return 22:45.830 --> 22:47.360 none and false. 22:47.480 --> 22:50.600 So zero errors and none came. 22:50.600 --> 22:53.360 The dictionary is just not great because we don't need it. 22:53.360 --> 22:54.780 There is no errors. 22:54.920 --> 22:55.610 Right. 22:55.880 --> 23:00.920 But then we've used here the generator your own dictionary and verify password so you need to write 23:00.920 --> 23:01.380 them. 23:01.600 --> 23:04.530 OK so the first one is the verify password. 23:04.580 --> 23:17.720 So you find verify password we take in a username and password and then if not user exist username then 23:17.720 --> 23:20.320 we're going to return false. 23:20.540 --> 23:26.000 Right because this this user name doesn't exist so the verify verifying the password failed. 23:26.990 --> 23:33.540 Also then we're going to hash get the hashed password so hashed password is equal to users not find. 23:33.800 --> 23:34.940 OK. 23:34.940 --> 23:39.410 And then inside the phone fine are you going to get their username the correct user. 23:39.680 --> 23:46.730 So user name user name and then for the hashed function I want to get the passwords I'm going to do 23:46.760 --> 23:48.740 zero and then password. 23:48.800 --> 23:49.040 Right. 23:49.040 --> 23:52.630 This is the hash password not the plain text. 23:53.060 --> 24:02.030 After I've done that I'm going to say if be script write the script that hash password and then password 24:02.120 --> 24:09.020 not encode UTF 8 UTF 8 bits 7. 24:09.020 --> 24:14.280 So UTF 8 with hash password. 24:14.320 --> 24:20.150 So this should be hashed password is equal to hash password. 24:20.530 --> 24:22.100 Then I'm going to return true. 24:22.300 --> 24:22.840 OK. 24:22.840 --> 24:28.520 So we're going to return true in this case because the passwords hash is matched match. 24:28.660 --> 24:30.420 Else I'm going to return false. 24:30.430 --> 24:31.960 So this is not correct. 24:32.040 --> 24:34.480 The Richards-Ross or it. 24:34.480 --> 24:40.860 So now we need another five to five on charities that generate return dictionary ok it's around right. 24:40.870 --> 24:45.040 Define return or generate return dictionary. 24:45.040 --> 24:54.630 So it's defined generate return dictionary and it's going to take a status and a message. 24:54.680 --> 24:55.100 Right. 24:55.330 --> 25:06.070 So I'm going to say return Jaison is equal to status is status and the message as the message right 25:07.410 --> 25:08.190 and the. 25:08.260 --> 25:13.000 And we just return this return return Jason and that's it. 25:13.000 --> 25:18.310 So we generate just using the status of the message we automatically in one line generate this Jason 25:18.310 --> 25:21.100 File here without having to waste a lot of lines. 25:22.090 --> 25:25.950 So that should be it for the classification. 25:25.960 --> 25:27.880 Right so we've done everything. 25:27.970 --> 25:34.330 And then finally we have the refill function which we'll do in the next video games are going to stop 25:34.330 --> 25:38.000 here and in the next video we're going to pick up from where we stop. 25:38.100 --> 25:38.620 OK. 25:38.800 --> 25:40.580 So until the next video recording.