YT YouTube2Text

YouTube2Text → Transcripts → How to set up a remote dev environment for algo-trading with Python in VSCode

How to set up a remote dev environment for algo-trading with Python in VSCode — Transcript

by Algo-trading with Saleh · 2,975 words · 482 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00let's say you have a great idea for a
0:02trading algorithm you're excited about
0:04it and you want to try it out but
0:06unfortunately your development
0:08environment is just not ready maybe it's
0:10because you haven't installed all the
0:11dependencies yet or maybe you are out of
0:14your office and cannot access your work
0:16computer or maybe you just want to play
0:18around with a couple of packages without
0:20messing up your existing environment if
0:22any of these describe your current
0:24situation you can use a remote dev
0:26environment
0:27it's really easy to set up at this after
0:29watching this tutorial and it's super
0:31affordable in fact you can get started
0:34for free
0:35[Music]
0:38when it comes to choosing a cloud server
0:40provider what we're looking for is a
0:42good balance of the quality and the
0:44price if you are living in the us i
0:46highly recommend using digitalocean
0:48there is starting server which is more
0:49than enough for our use case it starts
0:52from 5 bucks a month but if you use my
0:54link you can get it started for free the
0:56other cloud provider that i personally
0:58recommend and use is hestner their
1:01starting plan offers more resources for
1:03the same price and if you use my link
1:05you will get 20 euros free credits
1:08[Music]
1:11i already have a history account so i'm
1:13gonna log in
1:19i will choose the tutorials project
1:22since i'll be using this server for
1:24development is actually quite important
1:26if the location of the server is closest
1:28to me so that the ping would be lower
1:30for that reason i choose germany but you
1:32do whichever is closest to you
1:35ubuntu 2004 is exactly the image that i
1:38want so i'm not gonna change that
1:40i will choose the second server plan
1:42because for only 60 cents more per month
1:44i will get two cpu cores which i think
1:47totally worth it
1:50and now i have to choose the ssh key of
1:52my local machine
1:53in your case if your account is new you
1:55have to generate your ssh keys if you
1:58haven't already and add it by choosing
2:00the add ssh key button
2:02and at the end i will change the server
2:04name it's not really important what you
2:06set here but i prefer something that i
2:08can remember
2:10and now my server should be ready in
2:12just a few seconds
2:14for logging into my server i will be
2:16using the terminal app so i'm gonna copy
2:18the ip address from here
2:21and in my terminal i will type ssh root
2:24add sign and then paste my ip address it
2:26is asking me whether or not i trust this
2:28server and i do so i will type yes
2:32and we are in
2:34[Music]
2:37i will head over to jesse's
2:38documentation environment setup ubuntu
2:42and copy the command for running the
2:44installation script for ubuntu 2004
2:48i will place it here and press enter the
2:50installation process is going to take
2:52approximately seven minutes so i will
2:55skip this video and come back when it's
2:56done all right at the end it will ask
2:59you whether or not you want to change
3:00your default shell to z shell and i will
3:03type y to say yes
3:06at this point you should have everything
3:07installed just to check we will type
3:09python3
3:11python
3:12uppercase v and yes the python version
3:14is correct then i'll do the same for pip
3:17and just to show you a list of the
3:19packages that this script installed for
3:21us i will type pip list
3:24a couple of useful packages are y
3:27finance talib stats models
3:31quant stats
3:33numpy
3:34pandas and more
3:36let's create a new jesse project by
3:38copying the command from the
3:39documentation website and paste it into
3:42the terminal
3:44then i will cd into my project
3:46and create a new environment file by
3:48typing cp.mv.example.mv
3:53by typing ls i will see all the existing
3:56files inside my project
4:00now remember this is a remote server so
4:03to edit the files that are in it
4:05typically i have to use an editor that
4:07is already installed on that server by
4:10default the nano editor is what is
4:12installed on these types of server but
4:14the problem with it is that it kind of
4:16sucks and is very hard to use what i'd
4:18like to do instead is to open these
4:21files inside an editor that i am already
4:23used to such as pycharm or vs code
4:26fortunately there is an extension for vs
4:28code called remote ssh that allows us to
4:31do exactly that so go ahead download and
4:34install vs code for your platform if you
4:36haven't already and then head over and
4:38open the extensions section
4:41and inside the search box type remote
4:44via ssh
4:46choose the first result and click on the
4:48install button
4:53all right now that we have the extension
4:55installed it's time to connect to our
4:57server to connect to your server open
4:59the command palette by pressing command
5:02shift and p on mac or if you are on
5:04windows press shift control np
5:08and enter ssh
5:12and choose add new ssh host
5:15here i will type ssh root add sign and
5:18then paste in the ip address of my
5:20server and choose the first address for
5:21saving the configurations everything
5:24looks good i'm gonna open the command
5:25palette one more time and this time
5:27choose connect to host and choose the ip
5:29address of my server
5:33everything looks good if i run the ls
5:35command you will see my project files
5:37i'm going to sit into my project
5:39all right so now what i want to see is
5:41the files of the project inside vs codes
5:44built in explorer to open the project
5:47directory inside vs code i will click on
5:49the open folder button and choose my bot
5:52directory and click on the ok button
5:56and i will say yes i do trust the others
5:58which is me
6:00and now i can actually edit the project
6:02files that are hosted on the remote
6:04server as if they were hosted on my
6:06local machine it feels the same i can
6:08just edit something press save and
6:11that's it
6:13[Music]
6:17this looks really good but one thing
6:19that we're still missing is auto
6:21completion for instance let's say that i
6:24want to use the
6:25size to quantity utility function of
6:27jesse for that i can say quantity equals
6:31utils dot and now if you had auto
6:34completion i would have seen the methods
6:36that exist inside the utils module so to
6:39enable auto completion i need to install
6:42an extension so i will open the
6:44extension panel and look up python and
6:47intellisense is the extension that i
6:49want
6:51so i'll go ahead and click on the
6:53install in ssh green button it should
6:56take a few seconds
6:59so now if i head back to my strategy
7:02file remove the dot and press it again i
7:05will get auto completion so i'm gonna
7:08choose size to quantity
7:10and give it 500 bucks and the entry
7:12price would be my current price which is
7:14self.price
7:17and change the quantity from one to the
7:20quantity variable
7:22so now the editor can understand my
7:24python code and is able to help me write
7:27my strategies faster than before
7:32to start jesse i will open the built-in
7:35terminal inside vs code and execute
7:37jesse run
7:40the reason i'm getting this error is
7:42because by default the environment file
7:44that we created by copying from the
7:46example file is set to work with docker
7:49so for a native environment setup like
7:51ours we need to set postgres and redis
7:54hosts into localhost those values
7:56already exist so all i have to do now is
7:59to comment this line and uncomment this
8:01one and i'll do the same for redis
8:04all right let's give this another try
8:07and this time we get the database is not
8:09ready and after 10 seconds it fails with
8:12an authentication error that's because
8:15even though the installer script that we
8:17used earlier installed postgres for us
8:20it didn't create a database user and
8:22password so let's head over to the
8:24documentation again and navigate into
8:27the environment setup and under ubuntu
8:31you will find the postgresql section and
8:34i can just copy and paste these commands
8:36inside my terminal to create a username
8:39database and password
8:48so if i execute the command one more
8:51time it should work
8:55here's something really interesting
8:57usually when you're using jesse on a
8:58remote server you must use your server's
9:01external ip address in order to access
9:03it but vs code can do something magical
9:05for you if you run the same command
9:07inside vs code's built-in terminal it
9:09will automatically bind your local ip
9:11address into the remote server's
9:13external ip
9:15so if inside my local browser i enter
9:18localhost with 9000 port it will open
9:21the dashboard of my remote server and
9:23it's really cool because it feels like
9:25you're running things on your local
9:27machine and to prove it to you that this
9:29dashboard is indeed running on the
9:31remote server i will open the about
9:33section and as you can see the operating
9:35system is linux while my local machine
9:38is a mac so that's how i know this for
9:40sure
9:41[Music]
9:44to install jupyter lab inside the
9:46server's terminal i will execute pip
9:48install jewfighter lab it's going to
9:51take several seconds and now we can
9:53start using it to insert jupyter lab i
9:56will execute jupyter
9:58hyphen
9:59lab
10:00it throws an error saying that running
10:02as root is not recommended but i don't
10:05really mind it so i'm going to skip it
10:06by using the allow root flag at the end
10:09of it and just like before vs code is
10:12going to do its magic thing and bind my
10:15local address into the external ip
10:17address of the server so if i copy this
10:19url and paste it into my browser i
10:22should be able to see my jupyter lab
10:24workflow
10:26let's create a notebook to execute some
10:28python code in case you're not already
10:30familiar with due python notebooks they
10:32are great for researching using python
10:34they allow you to write your code inside
10:37separate blocks and see the result right
10:39there to show you a good example of this
10:42inside this notebook i'm gonna import
10:44some candles using jesse's research
10:46module calculate the 50 period sma of
10:50those candles and display both of them
10:52on a chart first i'm going to import a
10:55few libraries that i need
10:57from date time import date time
11:01import
11:02matplotlib
11:04dot
11:05pi plot
11:07as plt
11:09this next line is not actually required
11:11but i'm gonna do it just to silent an
11:13annoying warning of pandas
11:17next i will import the research module
11:19from jessie and also jesus technical
11:21indicator module
11:23ptc candles echo
11:27research dot
11:29get candles
11:31the first parameter is exchange name
11:34which i will pass binance
11:36the second one is the symbol which i
11:39will pass btc
11:41usdt
11:43i will pass 30 minutes as a time frame
11:47and then i will pass a starting date
11:49i want 10 days of candles
11:52from 10th of november 2021 until 10 days
11:55later
11:57if i execute this
11:59i will get an error saying that candle
12:02not found in database the reason i'm
12:04getting this error is because i didn't
12:06import any candles with jesse
12:09so i'm going to create a new terminal
12:11window on my vs code
12:13and execute the jsu-run command to
12:16run the dashboard
12:19after entering my password and logging
12:22in
12:23i'm gonna navigate into the import
12:25candles page
12:27and choose binance as my exchange
12:30btc usct as my symbol
12:34and change the starting date to 9th of
12:36november which is one day before the
12:37actual date that i need just in case the
12:40importing process is going to take
12:42approximately five minutes so i'm going
12:44to skip the video and come back when
12:46it's done
12:47all right now that we have our candles
12:49let's head back to the jupiter lab and
12:51try executing that block of code one
12:54more time
12:56this time i didn't get any errors which
12:58means it worked
12:59next i want the moving average of these
13:01candles with the period of 50. i can get
13:04that by saying ta.sma
13:08and pass in my candles
13:10and the next parameter is the period and
13:13because i need all the moving average
13:14values and not just the last one i'm
13:17going to pass sequential equals true
13:20just to be sure i'm gonna print the
13:22values
13:24i can minimize the output if i don't
13:26need it anymore by clicking on the blue
13:29line
13:31inside the chart i'm going to need only
13:33the closing prices and not the whole
13:34candles so i need a numpy array of the
13:37closing prices only and since jesse's
13:39candles have the type of numpy array
13:42that's easy to do
13:47let's print it just to be sure the
13:49format is correct
13:51and minimize it
13:54all right now i have everything that i
13:56need for my charts except for timestamps
13:58we already have timestamps in our
14:00candles but they are in the integer
14:02format which isn't what the matplotlib
14:04library expects
14:06so i need to convert them into datetime
14:08objects
14:10to do that i'm going to create a python
14:12list named times and then loop through
14:15every single candle
14:17create new time date out of those
14:19milliseconds timestamps and store it
14:21inside the times list
14:28let's print it just to make sure the
14:29format is correct
14:34and finally we can display our chart i
14:36will start by setting the size of my
14:38chart a little bigger than the default
14:44and then i will start by plotting btc
14:46close prices
14:49i can do that by saying plt dot plot
14:53which is a function
14:55and then i pass my timestamps and then
14:58btc close prices
15:01and i set the color to orange
15:05and i set the label to
15:07btc
15:10next i will plot
15:12the 50 moving average
15:14by typing plt dot plot
15:19and then again i pass timestamps
15:22moving average values
15:24and i set the color to
15:26black and the label to
15:29sma 50
15:32and finally to display the chart i will
15:34say plt dot legend
15:38oops look like i forgot
15:40an equal sign
15:42i'm sure you guys saw it
15:44let's give it another try and this time
15:46it works as you can see we have a chart
15:48with btc prizes and the 50 moving
15:51average this was a simple example but it
15:54shows you how powerful this thing could
15:55be
15:58[Music]
16:00this tutorial is almost done but before
16:02i leave you guys i want to show you a
16:03common problem with this setup and the
16:05solution for it when running python
16:07processes like we just did whether using
16:09the terminal app or vs code's built-in
16:11terminal one common problem is that
16:14sometimes you close the terminal session
16:16whether intentionally or accidentally
16:18but the process keeps running in the
16:20background so the next time that you try
16:22to run the same command it will fail
16:24saying that the port is already in use
16:26by another process there are a couple of
16:28ways to solve this problem my solution
16:30is to use a tool called a screen which
16:32gives you better control over your
16:34terminal sessions here let me show you
16:37you can install a screen on your server
16:39by running
16:40apt install python y screen
16:45in my case it's already installed
16:48and now instead of executing jsu-run
16:50directly
16:51first i will create an end screen
16:53i can do that by saying screen
16:55python uppercase s and then i pass in a
16:59name
16:59which makes reattaching easier
17:02for this process i will name it jesse
17:04and now i'm actually inside a new screen
17:07window
17:08so if i execute jsurun in here
17:12and close my terminal windows entirely
17:15i can later on open another terminal and
17:17reattach to the same window and to do
17:20that i will execute a screen python r
17:23and passing the name of that window
17:25which was jesse and i can just continue
17:28as if i never close the terminal at all
17:31now let's create another terminal window
17:33for jupiter screen
17:36hyphen uppercase s
17:40and i will name it dewpiter
17:42and to run jupiter i will
17:44execute jupiter
17:47lab
17:50then again i have to pass the allow root
17:53flag
17:56and then i can copy and paste this url
17:58inside my browser to use jupyter but
18:01sometimes when you are running the
18:02process inside a screen window the port
18:05binding of vs code doesn't work anymore
18:07so what i can do here is to replace
18:10localhost with the external ip address
18:12of my server the reason this isn't
18:14working now is because when i started
18:16jupyter i didn't specify a host address
18:19specifically so by default it uses
18:21localhost but to access the application
18:23using an external address i need to
18:26either specify the ip address or to
18:28allow it to serve the application to all
18:30addresses i can pass in four zeros so
18:34i'm gonna stop the session inter y
18:36so this time after writing the same
18:38command
18:39i will pass in the ip option
18:41with the value of
18:43zero zero zero zero
18:47now it doesn't matter which url i copy
18:49because i will be changing the hostname
18:51either way
18:52now i'm gonna paste in the ip address of
18:55my server
18:57and this time it should work
19:01let's do the same for jesse's dashboard
19:04and it works perfectly that's it for
19:07this tutorial this was just about the
19:08installation of the remote dev
19:10environment if you want to see examples
19:12of me using jc's research module make
19:15sure to subscribe because i will be
19:17making videos about that thanks for
19:19watching

About this transcript

This page contains the full transcript of How to set up a remote dev environment for algo-trading with Python in VSCode by Algo-trading with Saleh, generated from the public captions YouTube serves with the video. The transcript has 2,975 words across 482 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.