WEBVTT 00:00.160 --> 00:01.560 Hey there Eden here. 00:01.560 --> 00:08.120 And in this video, we're going to be setting up our environment before we implement the documentation 00:08.120 --> 00:08.680 helper. 00:09.160 --> 00:14.040 So we're going to start by cloning the starting branch with all the boilerplate code that we need. 00:14.080 --> 00:20.840 And we're going to set up our env file which is going to holding the environment variables for our OpenAI 00:20.960 --> 00:24.360 API key and our pinecone API key. 00:25.600 --> 00:25.880 Okay. 00:25.920 --> 00:30.240 So in the repository we want to be cloning the branch one. 00:30.240 --> 00:31.000 Start here. 00:31.680 --> 00:36.000 So I'm going to go and copy this code this URL. 00:36.240 --> 00:43.560 And I'm going to go to the terminal I'm going to write git clone paste the URL dash b one. 00:43.560 --> 00:44.280 Start here. 00:44.520 --> 00:48.200 And this will clone the repository in the beginning branch. 00:48.440 --> 00:51.920 So we'll start by listing the current working directory of our project. 00:52.160 --> 01:00.480 Okay, so now it's time to go to pinecone to log in to our user and to create an index that is going 01:00.480 --> 01:04.920 to store the embeddings of the documentation of blockchain. 01:04.920 --> 01:13.160 So now let's give our index a meaningful name, so I'll call it long chain dash doc dash index. 01:13.360 --> 01:19.320 And we're going to tell pinecone which embeddings model we're going to be using to index our documents. 01:19.480 --> 01:22.920 So we're going to choose text embeddings really small of OpenAI. 01:23.280 --> 01:24.200 The dimension. 01:24.200 --> 01:29.000 So the size of the vector is going to be 1536. 01:29.280 --> 01:34.600 And here we're going to be using cosine similarity to determine the distances between the vectors. 01:35.160 --> 01:35.520 All right. 01:35.560 --> 01:37.880 By the way this is the new pinecone UI. 01:38.040 --> 01:39.360 So it's pretty similar. 01:39.560 --> 01:47.240 Um you here can write your index name and and yeah here in order to choose the embedding model, this 01:47.240 --> 01:48.520 is how the UI looks. 01:48.760 --> 01:55.360 And make sure you select the one 536 um in the embeddings dimension. 01:55.760 --> 01:56.640 Um, yeah. 01:56.640 --> 01:58.840 So this is the new UI in pinecone. 01:58.840 --> 02:00.440 Just important to show you this. 02:02.360 --> 02:02.880 All right. 02:02.880 --> 02:04.800 Let's go and set configuration. 02:05.040 --> 02:07.360 We're going to go with the serverless option here. 02:07.360 --> 02:09.760 And we're simply now creating the index. 02:09.760 --> 02:13.280 And in a couple of seconds our index will be initialized. 02:13.280 --> 02:13.440 Relaxed. 02:13.560 --> 02:19.760 By the way, notice that our index is deployed on AWS cloud and this is the region. 02:19.760 --> 02:21.880 Now we have flexibility on this. 02:21.880 --> 02:24.520 Let's say we want to deploy it on Google Cloud. 02:24.560 --> 02:29.360 Then we simply needed to mark the Google Cloud logo and it will be deployed there. 02:29.560 --> 02:34.960 And this is important because AWS customers would want their index to be hosted on AWS. 02:35.280 --> 02:38.840 And GCP customers would want it to be hosted on Google Cloud. 02:38.840 --> 02:45.120 So this helps with latency and with commercial agreements of companies working with those cloud providers. 02:45.520 --> 02:49.480 And the region is important, for example, for GDPR compliance. 02:49.480 --> 02:55.560 So for that, we'll need our vector store to be deployed only in a Europe data center. 02:57.280 --> 02:57.840 All right. 02:57.840 --> 03:03.840 So now we can ingest the documents of the documentation into our index. 03:04.120 --> 03:07.800 So link chain is going to be making requests to pinecone on behalf of us. 03:07.800 --> 03:11.520 So we need to copy your API key and give it to link chain. 03:11.640 --> 03:15.160 So let's copy it and let's put it in the dot env file. 03:15.320 --> 03:24.000 And let me open now the project in PyCharm for the first time in this video and I created a dot env 03:24.040 --> 03:29.000 file, which of course I didn't commit to the repository because I don't want to commit any secrets 03:29.400 --> 03:29.920 anyway. 03:29.920 --> 03:34.320 So I created a dot env file and in the dot env file I'll create. 03:34.360 --> 03:38.000 I'll simply put here the pinecone API key which I copied. 03:38.000 --> 03:44.600 And also we need to plug in of course the OpenAI API key or any other LLM vendor you're using. 03:45.640 --> 03:46.080 All right. 03:46.080 --> 03:50.120 So I want to deactivate my virtual environment my automatic one. 03:50.120 --> 03:52.880 And I want now to take a look at the Pip file. 03:52.880 --> 03:57.960 And here we have all the packages we need for our project which I already pre-prepared. 03:58.680 --> 04:02.400 And you can see for example we have Lang Lang pinecone. 04:02.840 --> 04:08.440 And if I'm going to look at the log file with the exact versions we can see, we have for example the 04:08.440 --> 04:12.040 latest version currently of Linkchain 0.26. 04:12.640 --> 04:17.920 And I remind you, you might have a different version depending when you're taking the course. 04:17.960 --> 04:24.000 Of course, I'll keep updating this course and the repository to be updated with the latest version. 04:24.000 --> 04:28.360 So if there are breaking changes I will update the code of course, and the videos. 04:28.400 --> 04:28.800 All right. 04:28.800 --> 04:30.920 So I just want to address a couple of things. 04:31.200 --> 04:36.880 I added a new logger.py file, which you should be seeing on the left side as well. 04:36.920 --> 04:40.120 This is something which I added now as the boilerplate. 04:40.160 --> 04:42.480 We're going to be using it in future videos. 04:42.840 --> 04:46.320 And this backend directory you can ignore. 04:46.360 --> 04:48.520 We'll be creating it in future videos. 04:48.680 --> 04:56.240 And also this docs directory you can ignore will be downloading the link chain docs in the next couple 04:56.280 --> 04:56.920 of videos. 04:57.280 --> 05:00.040 All right so now I want to run Pipenv install. 05:00.040 --> 05:03.640 And this will take all the information in the pip file dot log file. 05:03.840 --> 05:07.840 And it's going to install all the dependencies with the correct versions. 05:08.040 --> 05:14.320 So if you run that you're going to have the exact same version as I'm having right now of all the packages. 05:15.520 --> 05:19.800 So we're going to create a new file I'll call it ingestion.py. 05:19.840 --> 05:28.400 And this file is going to hold all of the implementation of ingestion the link chain documentation embedding 05:28.400 --> 05:33.160 it into vectors then storing those vectors into pinecone the vector store.