WEBVTT 00:00.200 --> 00:01.520 Hey there Eden here. 00:01.520 --> 00:05.240 And in this video we're going to review the repository structure. 00:05.560 --> 00:07.800 Now we're going to organize our code. 00:07.800 --> 00:09.240 So it would be readable. 00:09.560 --> 00:13.000 It will be easier to maintain and to write testing. 00:13.000 --> 00:15.200 So our application will be more robust. 00:15.400 --> 00:22.440 And while this video series was inspired by the Link Chain Team's work implementing advanced RAC solution 00:22.440 --> 00:26.800 with Landgraf and their tutorial, I had to make some refactoring. 00:26.800 --> 00:35.080 Since the tutorial was more focused on a Jupyter notebook rather than a robust software that is supposed 00:35.080 --> 00:36.280 to run in production. 00:36.800 --> 00:43.600 Now, I always like to say that our repository structure should reflect our architecture, and in our 00:43.600 --> 00:49.120 case, it's going to be our graphs architecture, the nodes and edges that will be implementing. 00:49.760 --> 00:56.480 And most of our implementation is going to be in the graph package that will create and it will have 00:56.480 --> 00:57.840 the graph py file. 00:57.840 --> 01:02.360 And that's going to be where we're going to connect all of our nodes and edges. 01:02.600 --> 01:09.600 We're going to have the state py file, which is going to have the graph state object that will be modifying 01:09.640 --> 01:11.160 during the graph execution. 01:11.480 --> 01:19.080 We also want to have a const py file for constants that we want to use in the implementation, and we 01:19.080 --> 01:22.320 want to have a subpackage called nodes. 01:22.760 --> 01:27.720 And it's going to have all of the implementation of the nodes that are going to run. 01:28.040 --> 01:33.520 So in that package is going to have a lot of files where each file is going to be a node. 01:34.240 --> 01:41.400 And we're going to have the chains package where each file in the chains package is going to be a different 01:41.400 --> 01:47.520 chain, and it should correspond to the nodes that we're creating, because each node is eventually 01:47.520 --> 01:49.720 going to run a long chain chain. 01:50.120 --> 01:53.040 Now finally, we also want to be writing tests. 01:53.040 --> 01:57.920 So I created a test package and I created a test chains file. 01:57.920 --> 02:03.360 So here we're going to be implementing our test for the chains that we're going to write. 02:03.760 --> 02:10.150 And finally we also want to write an ingestion file which is going to be holding all of the logic that 02:10.150 --> 02:15.950 will download the information we want to index, and it's going to index it in the vector store. 02:16.310 --> 02:21.870 Now, before we begin and create all of this structure, I just want to note that this is something 02:21.870 --> 02:23.070 that works for me. 02:23.110 --> 02:27.670 I'm not claiming that this is the best project structure for a graph application. 02:27.670 --> 02:32.550 I'm just saying that this is something that works well for me, for my experience, and you'll see that 02:32.550 --> 02:38.550 once we develop this application, it would be very easy to start and to improve it and to make this 02:38.550 --> 02:40.630 application more robust with tests. 02:40.670 --> 02:44.630 Or we want to make it more complex with more nodes and edges. 02:44.830 --> 02:46.630 So this is just as an FYI. 02:48.670 --> 02:49.270 All right. 02:49.270 --> 02:50.430 Let's go to the code. 02:50.470 --> 02:53.870 Now notice that the branch is two project structure. 02:53.870 --> 02:56.550 So you can find here the final code for this video. 03:00.470 --> 03:03.510 Let's go and create now a new package. 03:03.510 --> 03:06.230 And we want to call this package graph. 03:10.830 --> 03:16.750 We want now to create a new file in graph package and we'll call this file state py. 03:19.750 --> 03:21.350 This will hold our graph state. 03:22.630 --> 03:27.470 Now let's go and create another Python file and we'll call it graph. 03:27.470 --> 03:31.590 And here we're going to write all the nodes and edges and connections between them. 03:32.390 --> 03:39.030 Finally we want to create a const file for constants that we're going to be using mainly names of nodes. 03:39.510 --> 03:43.030 And under graph we want to create a new sub package. 03:43.030 --> 03:44.790 And we want to call it nodes. 03:44.950 --> 03:47.710 And it's going to have all of our nodes implementation. 03:48.590 --> 03:51.270 So each node is going to be a file in this package. 03:51.630 --> 03:55.270 And every node is eventually going to run a chain. 03:55.590 --> 03:58.510 So I want to create a new change package. 03:58.510 --> 04:04.350 And here every file is going to be a new chain that should correspond to our nodes package. 04:04.830 --> 04:08.190 And we're going to be running tests for those chains. 04:08.190 --> 04:11.350 So let's create another sub package under chains. 04:11.350 --> 04:13.230 And we want to call it tests. 04:13.950 --> 04:17.390 And I want here to add now a new file. 04:17.390 --> 04:20.070 And I'm going to call it test chains. 04:26.190 --> 04:29.590 And this is going to have all of the testing implementation. 04:30.270 --> 04:35.350 And I have to add that the naming convention here is very important because we're going to be leveraging 04:35.350 --> 04:37.910 Pytest to orchestrate and run our tests. 04:37.910 --> 04:43.350 And Pytest is going to be looking for directories, which starts with tests, and it's going to look 04:43.350 --> 04:46.870 for test files with the prefix of test. 04:47.790 --> 04:52.750 Let's now add a dummy test to this project under Test chains. 04:53.070 --> 04:57.470 So I'm going to define a new function I'm going to call it test foo. 04:58.790 --> 05:01.870 And it's going to not return anything. 05:03.070 --> 05:06.390 And here I simply want to assert that one is equal to one. 05:06.390 --> 05:07.790 So this test should pass. 05:08.190 --> 05:12.030 And now I want to test that everything is working. 05:12.030 --> 05:14.510 And Pytest is able to run the tests. 05:14.710 --> 05:20.430 I will write in the terminal pytest dot to tell Py.test to run from the current directory, which is 05:20.430 --> 05:25.590 the root directory of the project is to display from stdout. 05:26.150 --> 05:32.150 Dash V is the verbose flag, which is going to show the tests that we ran in pytest, and we can see 05:32.150 --> 05:34.830 that Pytest ran the test as we wanted. 05:34.830 --> 05:43.070 It ran test foo from the test chains file, and now let's configure our runner to run our tests through 05:43.070 --> 05:44.430 the PyCharm runner. 05:44.910 --> 05:49.190 Let's go to the top right to Edit configurations. 05:49.710 --> 05:56.390 Here we want to click the plus button and to add Python tests and to select Pytest. 05:56.670 --> 06:04.270 And here we want to give here the um the directory to be the root directory here this is the script 06:04.270 --> 06:04.710 path. 06:04.990 --> 06:09.670 And the parameters is going to be dot dash dash v like we saw earlier. 06:10.470 --> 06:18.590 And now if we'll take a look we have here test chains and we can simply um run it. 06:18.590 --> 06:19.910 So let's go and run it. 06:19.910 --> 06:26.980 And we can see we ran the test successfully and if we want, we can even see the exact execution. 06:26.980 --> 06:31.340 So this is going to be convenient when we're going to be writing and running our tests. 06:33.020 --> 06:33.580 Cool. 06:33.580 --> 06:36.540 So we're almost done with the project structure. 06:36.980 --> 06:43.340 Now we want to create another file under our root directory for our vector store ingestion. 06:44.700 --> 06:46.700 And we'll call it ingestion. 06:50.700 --> 06:53.020 And that's pretty much it. 06:53.220 --> 06:56.660 We've now created the structure for our repository. 06:56.980 --> 07:01.900 And in the next video series we're going to be implementing all of those files. 07:01.900 --> 07:05.140 And we're going to be adding more files to this structure. 07:05.700 --> 07:11.740 And if you want to compare your code or to even download it, you can go and check out the two project 07:11.740 --> 07:16.420 structure branch, where we have all of what we saw in this video there. 07:17.740 --> 07:23.420 In the next video we'll be implementing our vector store ingestion to chroma DB.