Ultimate Data Science API Testing Tool

Key Takeaways

Hello all, my name is Krishna and welcome to my YouTube channel. So guys, today in this particular video I'm going to talk about this data science ultimate a API testing tool which is called as requestly. Now let's say if you're working as a data scientist or an AI engineer you know one of the most important task is basically to create machine learning endpoints uh where you may probably use a curl command or you may use any kind of tools you know for requesting that particular endpoint or let's say that you want to in integrate that endpoint with some of your applications at that point of time you know testing those endpoints can be very very much important along with that um when we talk about model inferencing what output you are specifically getting from the model. So in all the scenarios you try to create different configurations. You try to test this ML endpoints to make sure that you're getting the accurate result right and considering this you know this tool that is requestly definitely solves all this problem in just one unified application. Okay and that is what I'm actually going to talk about. U just a quick note uh this video is definitely sponsored by requestly but I have been using requestly uh for my workflows for developing AI agent applications you know making sure to test those kind of applications even uh when I want to probably inference deep learning models you know uh probably to see what kind of outputs I'm actually getting I'm actually using this okay so again u this is an amazing tool for you all so that's the reason why I'm making this particular video um and recent Recently if you see requestly right it has been um acquired by a company called as browser stack and uh you know because of this uh you know you can completely start this completely as a free and open source also. Okay. So in this particular video I will probably show you an example. I will take a amazing use case of YOLO 26 computer vision API you know where we will probably be able to do object detection and all and uh I will try to show you task like object detection segmentation classification with a specific endpoint. So that way you will be able to understand how powerful this specific tool is. Okay. So to start with what I'm actually going to do go over here. So this is u the GitHub that we have actually created with respect to the code. You need to clone this you know I've written all the code over here and uh the first thing what you need to do is that go to the requestly.com website. Okay go ahead and click on download. Okay so if you go ahead and click on download there are two options uh that you'll be seeing either you can add it as a chrome extension or a windows app. Okay whatever is your requirement you can go ahead with that. So I have installed already this in my Windows. So you can go ahead and click on download for Windows. And once you download it, it is just an exe file. You can double click it and you can start doing the installation. Okay. So here you can see after doing the installation, you'll be able to see it looks something like this. I have tested uh my YOLO application over here just to show it to you. But anyhow, I will show you completely from basics like how you can go ahead and create your collections, how you can go ahead and create your HTTP request and many more things. Okay. Now u we will take this entire project. We will clone it. And uh for this I'm actually going to use uh uh ID and uh you can use VS code, you can use Google anti-gravity. It is up to you. So I've cloned this entire thing. Uh now what you really need to do is that go ahead and create your uh you know environment. So I have already done this task because uh as being my audience right if you have seen my video I've shown this particular task a lot many times. But if you want to go ahead and just create a virtual environment, you just go ahead and write uv env. Okay. Once you write uv env since I'm going to use the uv package manager, it'll go ahead and create your environment which is called as venv. Okay. Now once you do this, I've already done that particular task. So I'll not repeat it. Then here you have something called as requirement.txt. We will be using all these specific libraries. uh and here you'll be also able to see fast API because with the help of fast API we'll be creating an endpoint for our uh you know all the features that we are implementing with uh Euro 26 uh you know computer vision APIs right for object detection segmentation and all right u so we are going to use fast API we are going to use uon you know and all these libraries will be installing it okay now if you go ahead and see simple yolo detector py so here is our entire class what we have basically written again I'm not going to go ahead in with this explaining each and every code because this kind of tutorials I have already uploaded in my YouTube channel okay but the thing that I'll be interested in is in my fast api_app py so here if you go ahead and see this here you'll be able to see that okay I have created a fast api application over here I have created a function called as get detector so this is my homepage in the homepage we are just going to display this hardcoded HTML Then there is also a health page and then here you can see API version detect uh version v1 detect okay so here we are basically taking a file we are detecting this you know and we are probably returning a response similarly there is one more endpoint which is called a segment we are going to use this okay uh inside this particular segment you'll be able to see that we are providing a model size and based on that we are getting some kind of response okay and similarly there is one more endpoint which is called as classify okay so three important function classify segment And here you can probably see detect. Okay. So we will try to test this APIs in a proper way. And this is completely built with the help of fast API. And there is also one more uh endpoint that we have created / test just to see that whether we are getting a proper response with respect to the HTML or not. Okay. So these are the things and here we have used uicon to uh load the file over here. Okay. So now uh this is done. So here you can see unicorn fast API and we are calling this particular app and uh we are probably running this particular endpoint. Okay. Now I'll again go back to my command prompt. Now inside this command prompt the first thing that I'll do I will just go ahead and install this requirement.xt. In order to install it just write uv pip install minus r requirement.txt. Okay. So once you go ahead and write this you should be able to do it. Okay. And uh here you'll also be able to see that I've written all the instruction over here which you should be able to do it for fast API you can go with this manner for streamlate you can go ahead with this manner right and uh you know you just need to go ahead and run this. So what I will do I will just go ahead and um you know copy this your uon fast API app reload port 8000 and for this you know I will just go ahead and open my command prompt. Okay. Now inside this particular command prompt you can go ahead and even activate your virtual environment. So since I've already done this installation of the virtual environment you can actually go ahead and do it. Now in order to run this entire application I'm going to use the uv command that is uvcon fast api app col reload with port 8000. So instead of 8,000 I'll use at80. And where did I get this particular command? Again it is updated in the entire GitHub. So here you can see that in order to run the fast API app you should actually use this and you can change the port according to your like let's say if you're using 8,000 you can use that only. Okay. Now once I go ahead and execute this now here you'll be able to see that now my uon is running over here on this particular port and the my application has started. Now I'll click on this particular link. Okay. Once I click on this link you can see that I'm able to get this YOLO 26 API API documentation. If I go ahead and click on this, it'll go to /docs. Here are all my endpoints that I've actually created. Health. Okay. If you want to go ahead and just test out, you can just go ahead and click on try it out. And you can go ahead and click on execute. So once you do that, you'll be getting this specific information. Then you have detect uh you have segment classify test client. Okay. In the test client, if you go ahead and click on try it out and execute it, you will be able to get this HTML response. Okay. But our main aim is to probably make it much more efficient in terms of testing with the help of this specific tools. Okay. Now what I will do uh even though I've created some of the collections over here, I will show you step by step how we can go ahead and test all those endpoints. Okay. So first of all uh here there is an option of creating a collection. So let's say here in one of the collection I'll just say yolo test app. Okay. So this will basically be my collection name. Now this is how you go ahead and create a collection. Now along with that here you can see you can also provide some collection description. Okay. So let's say I will go ahead and write this is the YOLO test collection. So here you can go ahead and provide some of the information like description. You can say okay these are my API endpoints you know uh what all endpoints I have. Let's say one is /docs uh one is slash test you know /home uh like this all the endpoint information you can basically go ahead and write okay again the best way is that you go ahead and give chity ch will probably generate the entire content or you can use any other llm models okay so this is one of the good feature because if anybody refers it they should be able to get this description information and uh they'll be able to get the basic information about the endpoints that has been created. Okay. Now the next thing is that after you do this we need to go ahead and test some HTTP requests. Right. So some of the endpoints if you want to test. So let's say HTTP request I will go ahead and uh select the HTTP request. I'll name it home. Okay. Now you know in home uh what I have. Okay. So I will go ahead and write 127.0.0.1. Okay. So this will be 8,000 or 8080 because since I'm running over here and this will basically be my home. This is a get request. So definitely I'll go ahead and select this. I'll go ahead and click on send. So here you can see that entire web you know how it was showing in the browser is being displayed over here. So that basically means that HTML content is being displayed over here. You can also go ahead and you know test all the other applications. So here you can actually see when I go ahead and just see the homepage it looks something like this. Now along with this there are three main uh pages. Let's say I want to go ahead and test out this test client. So for that I've created another endpoint which is called a / test. I'll go ahead and click on send. So again the entire HTML is basically uploaded over here. Right? So this test uh for the homepage is very much simple and here you can also see the best part is that you're also able to load the HTML content and you're also able to view it right which is really really good. Okay. Now uh let's go ahead and see some of the uh other endpoints. So let's say one of the endpoint that I have is basically classify. Okay, classify or segment. Let's say I want to go ahead and give segment. Okay, segment over here. And this again I will go ahead and change my address. So it should be 127.0.1 colon 8080. And I think the endpoint uh address uh we will go ahead and probably go ahead and see from our code. So let's say I will go ahead and write over here. So classify is /v1 segment slash I can use this particular endpoint. Okay. So here you can see I've put this. So this is there. This should be a post request in the body. you'll be able to see that uh I I should definitely be able to so if you see over here right so let's say yolo test over here I have created one right in the body you can give a multi-art form data where two parameters we are accepting in our code one is file and one is model size model sizes medium small and high uh that those are hardcoded over here okay so I will go back to my segment uh instead of segment I can also rename this to classify since uh I'm trying to do the classify over here. Okay. And then you should be able to see over here. So file. Okay. This should be a file. I will select a file. Let's say one of the file that I really want is like a dog cat image. You can probably download it from anyone any places. So I have downloaded some. So this is one good dog uh image that I have. uh other than that as I said the other parameter which I have already noted it out is nothing but model size okay so here I will go ahead and specify model size okay and this should be medium okay now this is there my and this should basically be a post request okay I did not save it now I'll go ahead and write it out 1270.1/80 okay/ slash whatever the API that we have actually seen API v1 classify API v1 classify so it should be colon not at80 colon / ai v1 uh classify okay so this is there in the body we have also saved this so I'll go ahead and save it so that this request is saved I'll go ahead and click on send okay once I go ahead and click on send let's see whether this will be able to probably give us some kind of response for the initial ial request it is definitely going to take some time because uh you know it is going to probably load the model and then probably give you the output. So here you can see success true classify it is labeled cyberian husky confidence this this this is there and this is the best part you know this image data that you're seeing this is very important okay and what is this image data I will just show you I'll copy this and I'll write image base wait 64 convert or decode this specific image so I'll Go ahead and paste it over here. You can decode this base. This is a base 64 uh data that you have and we'll try to decode this into the image. Okay. So here you'll be able to see that Siberian husky and all it is able to classify which is good, right? And uh this is the entire information that I have. So this is my classify. I will also go ahead and create one more HTTP request. Let's say this should be for segment. Okay, segment. And for segment also uh what I will do? I'll save this first of all. So in the classify here you are seeing this URL in the segment it should be slash segment okay and again this should be a post request in the body I have multififor data one is image this should be a file I will go ahead and upload the file let's say u that same image we will go ahead and use it okay we'll add one more which is nothing but model size okay model size and this should be sorry a text and let's say I'm going to use a medium you can also use small it is up to you and now again since we going to use slash segment I will go ahead and click on send and here you'll be able to see body file field required okay sorry I just made one simple mistake instead of this image right I have to probably go ahead and use file okay now I'll go ahead and save it now let's go ahead and click on send now this should definitely be able to segment it and this based on the segment you'll be able to see the box that you'll be able to see right the coordinates of that particular box right all those information is basically over here you can just see that how easy it is basically uh you know to test it out and again you want to go ahead and convert this you can just take this specific B 64 and you'll be able to convert this into an image so let's go ahead and write it over here and here you should be able to see some boxes see dog 83 which is good Right. So this is done. Uh I will go again back to my request and there is some more amazing feature. Okay. Let's say I go to this particular classify and uh inside this particular classify there is also an option to create environment variables. Okay. So let's say this is a classify. I will just go ahead and copy this or let's say this is the common part of the URL that I have. Okay. I will go to my YOLO test app collection name. I will go ahead and click on variables. Okay, I'll add a variable name. Let's say this is my URL. The type is string, number, boolean, secret, whatever you want. And I'll enter the value. So this is my initial value 127.0.0.1 u uh you know this is there. Okay, this is just like uh we are setting it up uh you know just to make sure that uh we have this. Okay, so this is my initial value and I also mentioned the local value is this. Okay, now this is my URL. Okay, let's say I will go ahead and again create one more environment variable. This is let's say home. Now you know in home the URL should be till here, right? Like let's say this is my home. Okay, and this is my home, right? Like just like my home date. I'll save this environment variable. Now the best part about this is that this environment variable can be used directly as a placeholder instead you writing the entire URL. Right? So if I now go to home right and let's say I will go ahead and specify over here instead of this I will just use a placeholder okay and the placeholder will basically be home okay now this home placeholder is basically specifying the entire it is referring the entire URL like what we have mentioned in our environment variables okay now how do you go ahead and set up the environment variable very simple just go to the collection name go ahead and see the variables over here and here you can specify any number of environment variables with the entire value. Okay. Now, if I go to home and if I just go ahead and click on send again, see, same thing is basically happening. Now, in the classify also, right, I will directly use the entire URL. No need to go ahead and write, which is again a pain, right? I don't need to go ahead and write the entire URL over here. I'll just go ahead and write URL and I'll go ahead and send. See, I'm getting the entire output. So, this is really amazing feature uh specifically with respect to the placeholder that you will be able to see over here. And uh here uh when you're giving the multiart form data also right I don't have to probably worry about base 64 image and all I'll directly upload the image and this tool will be able to do the entire task easily. Okay. Now similarly you can also go ahead and create one more HTTP request that is specifically for detect. Okay. Detect. Okay. Now inside the detect again I will uh this should be a post request. I will remove this and I will use the URL. Okay. And I'll set / detect. Okay. And in the body again it should be a multi- form data. I'll use the file name. File should be over here. Let's select the file. You can select any number of images you know or different images that you want. And here I will be using model size. Model size will be again medium. I'll be using medium or small whatever you want. Okay. And I've given this. Now if I go ahead and click on send, let's see. So I've got this. Now it is being able to detect the box, the couch. So this is basically an object detection task. Now if I go ahead and just let's say convert this into a normal image. Let's see what we'll be able to get. I think we should be able to get it in a better way. Okay. So I will just go ahead and open this B 64 and I'll click. Okay. Now I think it should be able to detect the result of the B 64 will appear here. Okay. I have to remove the quotes. See now it is able to detect the dog. It is able to detect the couch. This is good right? So all these things are basically happening and we have also seen that okay we can also create an environment variable which is a very good task altogether right uh I've shown you how to create an API collection how to and if you want to create a new collection just go ahead and click on new create a collection over here click on collection give the collection name give the collection information okay so collection is done API HTTP request is done I've also spoke about how to create create environment variables so these are the different routes books in my runner. Uh so out of this let's start with the easiest one to do the testing. Okay. So right now I will be attaching few test scripts. So first of all let's select the route. So by default all the routes are selected. I will be just unselecting them and let's start with the easiest one which is the get one. So let me just do a save. Okay. Now let's go into this route which is the home route. Now inside this as you can see I have a pre- request section and the post response section. Okay. So right now in the prerequest let's keep it blank otherwise I can keep some test cases but for now I easily focus on the post response here. So let me generate some test script seals in EI. So here you can see so here I'm writing two test cases for this request and check status 200. So let's wait. Okay. And here I can see all right upset. So now this is the script and we have saved it. Okay. Now when I go to my runner. Okay. And if I try to execute only this script run you can see that. Okay. So both the test cases status code is 200 and content type hta uh the content type header is text htm. So both the things that can be verified from here. All right. Now similarly if I want to try to add for the detect route in the detect route if I want I can add something with respect to my prerequest also. Now this information that I have added here. So simply you can see here some time stamp and request ID. Okay. These variables that I am setting and I'm just printing. So console.log. Now this information will be available in the browser console. So right now I won't be opening the browser but I'm much more worried about this post response part. Okay, in the post response. So let me just uh get my test script. So here you can see uh now these are the variables check that I am doing right now. Okay, like if script uh script executed and then we'll be printing this. So let me just add more uh generate uh two more test cases because the status 100 is already done. I will accept this one. Okay. So more test cases that I have added even like uh this here also I have added. So this can be checked from the browser but for my collection runner I can directly go here. Okay. So this is my collection here inside the runner. Let me select detect for now. Okay. And if I try to uh run this one, you will see that okay in both the conditions like get yellow stamp. Okay. Now the yellow test app as well as /detect and / in both the cases the test case have successfully passed. So similarly here you can add more test cases for your other routes. But I hope uh you got an idea about definitely working with requestly uh a very handful uh ultimate API testing tools definitely you can actually use for different different use cases. Anyhow all the information regarding this like the code everything you will be able to test it out. I will provide all the information in the description of this particular video. Uh but yes uh thank you uh requestly for sponsoring this particular video. Definitely a good tool to probably go ahead and explore it. So yeah, this was it from my side. I'll see you in the next video. Have a great day. Thank you all. Take care. on