WEBVTT 00:00.200 --> 00:00.640 Okay. 00:00.640 --> 00:02.080 Welcome back to the lab. 00:02.080 --> 00:07.360 Then we're at week two, day two, and we're going to run some imports and the load dot env. 00:07.680 --> 00:09.880 And now we're going to import account. 00:09.880 --> 00:11.640 This is the business logic class. 00:11.640 --> 00:14.760 Nothing to do with MCP servers yet I just want to show you this. 00:14.960 --> 00:19.520 When I import it it automatically creates a new SQL database with SQLite. 00:19.520 --> 00:21.680 Because that was in the code at the top. 00:21.680 --> 00:24.400 When you import, if it doesn't already exist, it creates one. 00:24.680 --> 00:30.640 And also if I get an account that doesn't exist, it creates a default account with $10,000 in it, 00:30.680 --> 00:31.480 which is kind. 00:31.600 --> 00:32.360 So there we go. 00:32.400 --> 00:33.440 This is all a simulation. 00:33.480 --> 00:33.760 Of course. 00:33.800 --> 00:37.760 It's just just to have something stored so we can watch trading happening. 00:37.920 --> 00:42.160 So here we go a new account called Ed with a fake balance of $10,000. 00:42.200 --> 00:45.640 No holdings, no trades, no time series. 00:45.880 --> 00:50.920 And we're now going to buy three shares of Amazon in this fake simulation. 00:51.120 --> 00:54.480 And we're going to pretend that I'm thinking of doing this years ago. 00:54.680 --> 01:01.100 And my my rationale for doing it is because this bookstore website looks promising. 01:01.580 --> 01:03.860 Oh, if only I had had that foresight. 01:04.140 --> 01:05.500 Uh, so there we go. 01:05.540 --> 01:08.980 We buy shares completed, and it gives some details of that. 01:09.020 --> 01:10.340 The balance, the strategy. 01:10.540 --> 01:15.700 And I should say, actually, there's some code in here that looks up the current share price of Amazon. 01:15.700 --> 01:18.340 And later we'll come back and I'll talk about how we do that. 01:19.020 --> 01:22.900 Then we are going to, uh, ask for a report of the account. 01:22.940 --> 01:28.060 And here it is with the strategy of the transactions, the rationale. 01:28.060 --> 01:33.060 And over at the end here, there is a total portfolio value and a profit and loss. 01:33.060 --> 01:38.260 And you'll notice that the total portfolio value is slightly less than the $10,000 it was originally. 01:38.260 --> 01:41.100 And there's a tiny negative, uh, profit and loss. 01:41.460 --> 01:47.660 And for those that care about this, that want to dig into this, uh, I have made a small assumption, 01:47.700 --> 01:54.220 a fixed assumption of, um, a fee that you pay every time that you trade in the form of what they call 01:54.260 --> 01:56.220 a a bid offer spread. 01:56.220 --> 01:57.420 And you'll see that in the code. 01:57.420 --> 02:03.030 It's just a very simplistic assumption to mean that every time you trade, you lose a little bit. 02:03.510 --> 02:03.990 Okay. 02:04.030 --> 02:04.830 So let's list. 02:04.870 --> 02:06.710 Transactions in our portfolio. 02:06.750 --> 02:08.510 We have three shares of Amazon at. 02:08.550 --> 02:09.390 That much each. 02:09.590 --> 02:12.390 And that is a quick demo of the business logic. 02:12.430 --> 02:19.070 The next step is to write is to is to use our MCP server and see if we can access this directly through 02:19.070 --> 02:20.310 the MCP server. 02:20.350 --> 02:20.910 All right. 02:20.910 --> 02:22.470 So how hard do you think it will be. 02:22.550 --> 02:24.270 It's going to be very easy. 02:24.590 --> 02:26.830 Uh so remember this params. 02:26.870 --> 02:29.870 Params is how you say what it is that you want to do. 02:29.910 --> 02:31.990 What what server do you want to run. 02:32.270 --> 02:35.230 And literally the command is you've. 02:35.470 --> 02:37.190 And the arguments is run. 02:37.190 --> 02:40.110 And then account server Pi. 02:40.390 --> 02:43.430 It's just going to spawn a process. 02:43.750 --> 02:44.910 It's going to do it itself. 02:44.910 --> 02:46.350 It's going to run this process. 02:46.350 --> 02:49.230 You've run account server dot Pi. 02:49.990 --> 02:50.550 Okay. 02:50.910 --> 02:56.270 Now I have this line async with MCP server studio as server. 02:56.270 --> 02:57.950 So what is MCP Server Studio. 02:57.950 --> 03:04.170 So this this is a class from OpenAI agents SDK. 03:04.530 --> 03:13.850 And it's basically some code which essentially creates an MCP client and attaches to an STD I o server 03:13.970 --> 03:15.330 that you want to run. 03:15.330 --> 03:21.250 It will it will spawn that server and then run it and allow us to do something with it. 03:21.250 --> 03:24.970 So all of that is simply in this one class. 03:25.250 --> 03:30.850 And as I think I mentioned before, this is actually relatively new in OpenAI agents SDK. 03:30.850 --> 03:36.690 And as of just recently, you would have to do that last part yourself of writing the code that spawns 03:36.690 --> 03:37.370 the server. 03:37.610 --> 03:40.970 And fear not, we will do that anyway, just so you see that side of it. 03:40.970 --> 03:44.610 But you don't need to because OpenAI agents SDK has written that for you. 03:44.810 --> 03:49.570 So this is the client and spawns that service. 03:49.850 --> 03:56.410 And then we are just going to call list tools, just as we did for fetch, for all of the other tools 03:56.410 --> 03:57.850 that we used last time. 03:57.850 --> 03:58.850 This is the same code. 03:58.850 --> 04:00.370 It's just three lines. 04:00.390 --> 04:04.030 we run it, we get the MCP tools and here they are. 04:04.070 --> 04:08.550 Get balance, get holdings, buy shares, sell shares with a description. 04:08.550 --> 04:13.910 And then you'll see if we scroll along that there are going to be all of the different bits and pieces 04:13.910 --> 04:14.670 to it. 04:14.670 --> 04:23.950 So if we now just remind ourselves if we go to the accounts server, you'll see here is our MCP server. 04:23.950 --> 04:30.470 And so get balance right here was one of the one of the decorated functions from our server. 04:30.470 --> 04:36.230 And if we go back here you'll see that get balance with that description is exactly what has been surfaced 04:36.230 --> 04:36.750 here. 04:36.950 --> 04:40.830 And so it really is that straightforward okay. 04:40.830 --> 04:43.790 So we've seen how easy it is to list the tools. 04:43.870 --> 04:45.110 Now let's use them. 04:45.310 --> 04:49.070 So it really is going to be again very straightforward. 04:49.350 --> 04:54.750 Uh writing MCP servers is a little bit of a few extra decorators here and there. 04:54.790 --> 04:56.470 Using them is a cinch. 04:57.030 --> 05:00.430 So here here's uh, some instructions or a system prompt. 05:00.490 --> 05:04.090 you are able to manage an account for a client and answer questions about the account. 05:04.130 --> 05:04.850 My request. 05:04.850 --> 05:05.850 My name is Ed. 05:05.890 --> 05:07.450 My account is under the name Ed. 05:07.490 --> 05:09.490 What's my balance and my holdings? 05:09.650 --> 05:11.490 Model GPT four mini. 05:11.730 --> 05:12.690 That sets us up. 05:12.970 --> 05:13.970 And so here it is. 05:13.970 --> 05:17.730 This is the code with MCP server studio. 05:17.770 --> 05:24.370 That is the OpenAI agents SDK thing passing in the parameters of our server, caching the tools. 05:24.370 --> 05:27.930 Why not uh, as uh, as MCP server? 05:28.250 --> 05:32.850 We don't need to do this, particularly for one call like this, especially to to show you how simple 05:32.850 --> 05:33.690 the code looks. 05:34.090 --> 05:38.570 Uh, so with an we create an agent called account manager. 05:38.570 --> 05:44.290 We give it the system, prompt the model, and we pass in the server, the MCP server that we want to 05:44.330 --> 05:48.090 use, which is r MCP server, our very own homemade one. 05:48.090 --> 05:52.730 And then we put in a trace and we say run a run with the agent. 05:52.730 --> 05:56.050 And we pass in our request and we'll display the final output. 05:56.050 --> 05:57.930 The question is is this going to work. 05:57.930 --> 05:58.650 What it's going to do? 05:58.650 --> 06:04.820 If it works, it's going to create an MCP client, which will then spawn an MCP server. 06:04.820 --> 06:10.860 It will basically call you've run account Server.py launching that process. 06:10.860 --> 06:16.180 It will talk to that process using standard input and output, and it will collect the tools. 06:16.180 --> 06:22.340 It will pass those tools into our agent by constructing the JSON gumpf that will go in. 06:22.580 --> 06:28.740 And when the LM says it wants to use a tool, then this code will then run that using standard input 06:28.740 --> 06:32.220 output to ask our spawned process to run that tool. 06:32.340 --> 06:35.780 And then the results will come back and be given to the LM which will complete. 06:35.780 --> 06:41.980 So there is a lot of stuff happening backwards and forwards that's all been elegantly wrapped into this. 06:42.060 --> 06:42.780 Will it work? 06:42.780 --> 06:49.500 Will we find out our holdings in natural language from asking GPT four or mini based on our very own 06:49.540 --> 06:50.140 server? 06:50.460 --> 06:51.460 Yes we will. 06:51.740 --> 06:55.340 Your current balance is 9470. 06:55.420 --> 06:58.980 In terms of holdings, you have three shares of Amazon. 06:59.420 --> 07:01.140 That is mission accomplished.