├── syllabus
└── ECBS5211.pdf
└── README.md
/syllabus/ECBS5211.pdf:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/daroczig/CEU-R-prod/HEAD/syllabus/ECBS5211.pdf
--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
1 | Here you can find the materials for the "[Data Engineering 3: Batch Jobs and APIs](https://ceu.studyguide.timeedit.net/modules/ECBS5211?type=CORE)" course, part of the [MSc in Business Analytics](https://courses.ceu.edu/programs/ms/master-science-business-analytics) at CEU. For the previous editions, see [2017/2018](https://github.com/daroczig/CEU-R-prod/tree/2017-2018), [2018/2019](https://github.com/daroczig/CEU-R-prod/tree/2018-2019), [2019/2020](https://github.com/daroczig/CEU-R-prod/tree/2019-2020), [2020/2021](https://github.com/daroczig/CEU-R-prod/tree/2020-2021), [2021/2022](https://github.com/daroczig/CEU-R-prod/tree/2021-2022), [2022/2023](https://github.com/daroczig/CEU-R-prod/tree/2022-2023), and [2023/2024](https://github.com/daroczig/CEU-R-prod/tree/2023-2024).
2 |
3 | ## Table of Contents
4 |
5 | * [Table of Contents](#table-of-contents)
6 | * [Schedule](#schedule)
7 | * [Location](#location)
8 | * [Syllabus](#syllabus)
9 | * [Technical Prerequisites](#technical-prerequisites)
10 | * [Class Schedule](#class-schedule)
11 |
12 | * [Home assignment](#home-assignment)
13 | * [Getting help](#getting-help)
14 |
15 | ## Schedule
16 |
17 | 3 x 2 x 100 mins on April 28, May 5 and 12:
18 |
19 | * 13:30 - 15:10 session 1
20 | * 15:10 - 15:40 break
21 | * 15:40 - 17:20 session 2
22 |
23 | ## Location
24 |
25 | In-person at the Vienna campus (QS B-421).
26 |
27 | ## Syllabus
28 |
29 | Please find in the `syllabus` folder of this repository.
30 |
31 | ## Technical Prerequisites
32 |
33 | 1. You need a laptop with any operating system and stable Internet connection.
34 | 2. Please make sure that Internet/network firewall rules are not limiting your access to unusual ports (e.g. 22, 8787, 8080, 8000), as we will heavily use these in the class (can be a problem on a company network). CEU WiFi should have the related firewall rules applied for the class.
35 | 3. Join the Teams channel dedicated to the class at `ba-de4-2024` with the `symvo7f` team code
36 | 4. When joining remotely, it's highly suggested to get a second monitor where you can follow the online stream, and keep your main monitor for your own work. The second monitor could be an external screen attached to your laptop, e.g. a TV, monitor, projector, but if you don't have access to one, you may also use a tablet or phone to dial-in to the Zoom call.
37 |
38 | ## Class Schedule
39 |
40 | ## Week 1
41 |
42 | **Goal**: learn how to run and schedule R jobs in the cloud.
43 |
44 | ### Background: Example use-cases and why to use R in the cloud?
45 |
46 | Excerpts from https://daroczig.github.io/talks
47 |
48 | * "A Decade of Using R in Production" (Real Data Science USA - R meetup)
49 | * "Getting Things Logged" (RStudio::conf 2020)
50 | * "Analytics databases in a startup environment: beyond MySQL and Spark" (Budapest Data Forum 2018)
51 |
52 | ### Welcome to AWS!
53 |
54 | 1. Use the following sign-in URL to access the class AWS account: https://657609838022.signin.aws.amazon.com/console
55 | 2. Secure your access key(s), other credentials and any login information ...
56 |
57 | ... because a truly wise person learns from the mistakes of others!
58 |
59 | > "When I woke up the next morning, I had four emails and a missed phone call from Amazon AWS - something about 140 servers running on my AWS account, mining Bitcoin"
60 | -- [Hoffman said](https://www.theregister.co.uk/2015/01/06/dev_blunder_shows_github_crawling_with_keyslurping_bots)
61 |
62 | > "Nevertheless, now I know that Bitcoin can be mined with SQL, which is priceless ;-)"
63 | -- [Uri Shaked](https://medium.com/@urish/thank-you-google-how-to-mine-bitcoin-on-googles-bigquery-1c8e17b04e62)
64 |
65 | So set up 2FA (go to IAM / Users / username / Security credentials / Assigned MFA device): https://console.aws.amazon.com/iam
66 |
67 | PS probably you do not really need to store any access keys, but you may rely on roles (and the Key Management Service, and the Secrets Manager and so on)
68 |
69 |
70 | 3. Let's use the `eu-west-1` Ireland region
71 |
72 | ### Getting access to EC2 boxes
73 |
74 | **Note**: we follow the instructions on Windows in the Computer Lab, but please find below how to access the boxes from Mac or Linux as well when working with the instances remotely.
75 |
76 | 1. Create (or import) an SSH key in AWS (EC2 / Key Pairs): https://eu-west-1.console.aws.amazon.com/ec2/v2/home?region=eu-west-1#KeyPairs:sort=keyName including using the Owner tag!
77 | 2. Get an SSH client:
78 |
79 | * Windows -- Download and install PuTTY: https://www.putty.org
80 | * Mac -- Install PuTTY for Mac using homebrew or macports
81 |
82 | ```sh
83 | sudo brew install putty
84 | sudo port install putty
85 | ```
86 |
87 | * Linux -- probably the OpenSSH client is already installed, but to use the same tools on all operating systems, please install and use PuTTY on Linux too, eg on Ubuntu:
88 |
89 | ```sh
90 | sudo apt install putty
91 | ```
92 |
93 | 3. ~~Convert the generated pem key to PuTTY format~~No need to do this anymore, AWS can provide the key as PPK now.
94 |
95 | * GUI: https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/putty.html#putty-private-key
96 | * CLI:
97 |
98 | ```sh
99 | puttygen key.pem -O private -o key.ppk
100 | ```
101 |
102 | 4. Make sure the key is readable only by your Windows/Linux/Mac user, eg
103 |
104 | ```sh
105 | chmod 0400 key.ppk
106 | ```
107 |
108 | ### Create and connect to an EC2 box
109 |
110 | 1. Create an EC2 instance
111 |
112 | 0. Optional: create an Elastic IP for your box
113 | 1. Go the the Instances overview at https://eu-west-1.console.aws.amazon.com/ec2/v2/home?region=eu-west-1#Instances:sort=instanceId
114 | 2. Click "Launch Instance"
115 | 3. Provide a name for your server (e.g. `daroczig-de3-week1`) and some additional tags for resource tracking, including tagging downstream services, such as Instance and Volumes:
116 | * Class: `DE3`
117 | * Owner: `daroczig`
118 | 4. Pick the `Ubuntu Server 24.04 LTS (HVM), SSD Volume Type` AMI
119 | 5. Pick `t3a.small` (2 GiB of RAM should be enough for most tasks) instance type (see more [instance types](https://aws.amazon.com/ec2/instance-types))
120 | 6. Select your AWS key created above and launch
121 | 7. Pick a unique name for the security group after clicking "Edit" on the "Network settings"
122 | 8. Click "Launch instance"
123 | 9. Note and click on the instance id
124 |
125 | 2. Connect to the box
126 |
127 | 1. Specify the hostname or IP address
128 |
129 | 
130 |
131 | 2. Specify the "Private key file for authentication" in the Connection category's SSH/Auth/Credentials pane
132 | 3. Set the username to `ubuntu` on the Connection/Data tab
133 | 4. Save the Session profile
134 | 5. Click the "Open" button
135 | 6. Accept & cache server's host key
136 |
137 | Alternatively, you can connect via a standard SSH client on a Mac or Linux, something like:
138 |
139 | ```sh
140 | chmod 0400 /path/to/your/pem
141 | ssh -i /path/to/your/pem ubuntu@ip-address-of-your-machine
142 | ```
143 |
144 | As a last resort, use "EC2 Instance Connect" from the EC2 dashboard by clicking "Connect" in the context menu of the instance (triggered by right click in the table).
145 |
146 | ### Install RStudio Server on EC2
147 |
148 | 1. Look at the docs: https://www.rstudio.com/products/rstudio/download-server
149 | 2. First, we will upgrade the system to the most recent version of the already installed packages. Note, check on the concept of a package manager!
150 |
151 | Download Ubuntu `apt` package list:
152 |
153 | ```sh
154 | sudo apt update
155 | ```
156 |
157 | Optionally upgrade the system:
158 |
159 | ```sh
160 | sudo apt upgrade
161 | ```
162 |
163 | And optionally also reboot so that kernel upgrades can take effect.
164 |
165 | 3. Install R
166 |
167 | ```sh
168 | sudo apt install r-base
169 | ```
170 |
171 | To avoid manually answering "Yes" to the question to confirm installation, you can specify the `-y` flag:
172 |
173 | ```sh
174 | sudo apt install -y r-base
175 | ```
176 |
177 |
178 | 4. Try R
179 |
180 | ```sh
181 | R
182 | ```
183 |
184 | For example:
185 |
186 | ```r
187 | 1 + 4
188 | hist(mtcars$hp)
189 | # duh, where is the plot?!
190 | ```
191 |
192 | Exit:
193 |
194 | ```r
195 | q()
196 | ```
197 |
198 | Look at the files:
199 |
200 | ```sh
201 | ls
202 | ls -latr
203 | ```
204 |
205 |
206 | 5. Install RStudio Server
207 |
208 | ```sh
209 | wget https://download2.rstudio.org/server/jammy/amd64/rstudio-server-2024.12.1-563-amd64.deb
210 | sudo apt install -y gdebi-core
211 | sudo gdebi rstudio-server-2024.12.1-563-amd64.deb
212 | ```
213 |
214 | 6. Check process and open ports
215 |
216 | ```sh
217 | rstudio-server status
218 | sudo rstudio-server status
219 | sudo systemctl status rstudio-server
220 | sudo ps aux | grep rstudio
221 |
222 | sudo apt -y install net-tools
223 | sudo netstat -tapen | grep LIST
224 | sudo netstat -tapen
225 | ```
226 |
227 | 7. Look at the docs: http://docs.rstudio.com/ide/server-pro/
228 |
229 | ### Connect to the RStudio Server
230 |
231 | 1. Confirm that the service is up and running and the port is open
232 |
233 | ```console
234 | ubuntu@ip-172-31-12-150:~$ sudo netstat -tapen | grep LIST
235 | tcp 0 0 0.0.0.0:8787 0.0.0.0:* LISTEN 0 49065 23587/rserver
236 | tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN 0 15671 1305/sshd
237 | tcp6 0 0 :::22 :::* LISTEN 0 15673 1305/sshd
238 | ```
239 |
240 | 2. Try to connect to the host from a browser on port 8787, eg http://foobar.eu-west-1.compute.amazonaws.com:8787
241 | 3. Realize it's not working
242 | 4. Open up port 8787 in the security group, by selecting your security group and click "Edit inbound rules":
243 |
244 | 
245 |
246 | 5. Authentication: http://docs.rstudio.com/ide/server-pro/authenticating-users.html
247 | 6. Create a new user:
248 |
249 | sudo adduser ceu
250 |
251 | 7. Login & quick demo:
252 |
253 | ```r
254 | 1+2
255 | plot(mtcars)
256 | install.packages('fortunes')
257 | library(fortunes)
258 | fortune()
259 | fortune(200)
260 | system('whoami')
261 | ```
262 |
263 | 8. Reload webpage (F5), realize we continue where we left the browser :)
264 | 9. Demo the terminal:
265 |
266 | ```console
267 | $ whoami
268 | ceu
269 | $ sudo whoami
270 | ceu is not in the sudoers file. This incident will be reported.
271 | ```
272 |
273 | 8. Grant sudo access to the new user by going back to SSH with `root` access:
274 |
275 | ```sh
276 | sudo apt install -y mc
277 | sudo mc
278 | sudo mcedit /etc/sudoers
279 | sudo adduser ceu admin
280 | man adduser
281 | man deluser
282 | ```
283 |
284 | Note 1: might need to relogin / restart RStudio / reload R / reload page .. to force a new shell login so that the updated group setting is applied
285 | Note 2: you might want to add `NOPASSWD` to the `sudoers` file:
286 |
287 | ```sh
288 | ceu ALL=(ALL) NOPASSWD:ALL
289 | ```
290 |
291 | Although also note (3) the related security risks.
292 |
293 | 9. Custom login page: http://docs.rstudio.com/ide/server-pro/authenticating-users.html#customizing-the-sign-in-page
294 | 10. Custom port (e.g. 80): http://docs.rstudio.com/ide/server-pro/access-and-security.html#network-port-and-address
295 |
296 | ```sh
297 | echo "www-port=80" | sudo tee -a /etc/rstudio/rserver.conf
298 | sudo rstudio-server restart
299 |
300 | ### Update R
301 |
302 | Note the pretty outdated R version ... so let's update R by using the apt repo managed by the CRAN team:
303 |
304 | ```sh
305 | wget -qO- https://cloud.r-project.org/bin/linux/ubuntu/marutter_pubkey.asc | sudo tee -a /etc/apt/trusted.gpg.d/cran_ubuntu_key.asc
306 | sudo add-apt-repository "deb https://cloud.r-project.org/bin/linux/ubuntu $(lsb_release -cs)-cran40/"
307 |
308 | ## fetch most recent list of packages and version, then auto-upgrade
309 | sudo apt-get update && sudo apt-get -y upgrade
310 | ```
311 |
312 | ## Week 2
313 |
314 | Quiz: https://forms.office.com/e/SgME7mHPM9 (5 mins)
315 |
316 | Recap on what we covered last week:
317 |
318 | 1. 2FA/MFA in AWS
319 | 2. Creating EC2 nodes
320 | 3. Connecting to EC2 nodes via SSH/Putty (note the difference between the PPK and PEM key formats)
321 | 4. Updating security groups
322 | 5. Installing RStudio Server
323 | 6. The difference between R console and Shell
324 | 7. The use of `sudo` and how to grant `root` (system administrator) privileges
325 | 8. Adding new Linux users, setting password, adding to group
326 | 9. Messing up package updates and making RStudio Server unusable
327 |
328 | Note that you do NOT need to do the instructions below marked with the 💪 emoji -- those have been already done for you, and the related steps are only included below for documenting what has been done and demonstrated in the class.
329 |
330 | ### Amazon Machine Images
331 |
332 | 💪 Instead of starting from scratch, let's create an Amazon Machine Image (AMI) from the EC2 node we used last week, so that we can use that as the basis of all the next steps:
333 |
334 | * Find the EC2 node in the EC2 console
335 | * Right click, then "Image and templates" / "Create image"
336 | * Name the AMI and click "Create image"
337 | * It might take a few minutes to finish
338 |
339 | Then you can use the newly created `de3-week2` AMI to spin up a new instance for you:
340 |
341 | 1. Go the the Instances overview at https://eu-west-1.console.aws.amazon.com/ec2/v2/home?region=eu-west-1#Instances:sort=instanceId
342 | 2. Click "Launch Instance"
343 | 3. Provide a name for your server (e.g. `daroczig-de3-week2`) and some additional tags for resource tracking, including tagging downstream services, such as Instance and Volumes:
344 | * Class: `DE3`
345 | * Owner: `daroczig`
346 | 4. Pick the `de3-week2` AMI
347 | 5. Pick `t3a.small` (2 GiB of RAM should be enough for most tasks) instance type (see more [instance types](https://aws.amazon.com/ec2/instance-types))
348 | 6. Select your AWS key created above and launch
349 | 7. Select the `de3` security group (granting access to ports 22, 8000, 8080, and 8787)
350 | 8. Click "Advanced details" and select `ceudataserver` IAM instance profile
351 | 9. Click "Launch instance"
352 | 10. Note and click on the instance id
353 |
354 | ### 💪 Create a user for every member of the team
355 |
356 | We'll export the list of IAM users from AWS and create a system user for everyone.
357 |
358 | 1. Attach a newly created IAM EC2 Role (let's call it `ceudataserver`) to the EC2 box and assign 'Read-only IAM access' (`IAMReadOnlyAccess`):
359 |
360 | 
361 |
362 | 
363 |
364 | 
365 |
366 | 2. Install AWS CLI tool (note that using the snap package manager as it was removed from the apt repos):
367 |
368 | ```
369 | sudo snap install aws-cli --classic
370 | ```
371 |
372 | 3. List all the IAM users: https://docs.aws.amazon.com/cli/latest/reference/iam/list-users.html
373 |
374 | ```
375 | aws iam list-users
376 | ```
377 |
378 | 4. Install R packages from JSON parsing and logging (in the next steps) from the apt repo instead of CRAN sources as per https://github.com/eddelbuettel/r2u
379 |
380 | ```sh
381 | wget -q -O- https://eddelbuettel.github.io/r2u/assets/dirk_eddelbuettel_key.asc | sudo tee -a /etc/apt/trusted.gpg.d/cranapt_key.asc
382 | sudo add-apt-repository "deb [arch=amd64] https://r2u.stat.illinois.edu/ubuntu noble main"
383 | sudo apt update
384 |
385 | sudo apt-get install -y r-cran-jsonlite r-cran-logger r-cran-glue
386 | ```
387 |
388 | Note that all dependencies (let it be an R package or system/Ubuntu package) have been automatically resolved and installed.
389 |
390 | Don't forget to click on the brush icon to clean up your terminal output if needed.
391 |
392 | Optionally [enable `bspm`](https://github.com/eddelbuettel/r2u#step-5-use-bspm-optional) to enable binary package installations via the traditional `install.packages` R function.
393 |
394 | 5. Export the list of users from R:
395 |
396 | ```
397 | library(jsonlite)
398 | users <- fromJSON(system('aws iam list-users', intern = TRUE))
399 | str(users)
400 | users[[1]]$UserName
401 | ```
402 |
403 | 6. Create a new system user on the box (for RStudio Server access) for every IAM user, set password and add to group:
404 |
405 | ```
406 | library(logger)
407 | library(glue)
408 | for (user in users[[1]]$UserName) {
409 |
410 | ## remove invalid character
411 | user <- sub('@.*', '', user)
412 | user <- sub('.', '_', user, fixed = TRUE)
413 |
414 | log_info('Creating {user}')
415 | system(glue("sudo adduser --disabled-password --quiet --gecos '' {user}"))
416 |
417 | log_info('Setting password for {user}')
418 | system(glue("echo '{user}:secretpass' | sudo chpasswd")) # note the single quotes + placement of sudo
419 |
420 | log_info('Adding {user} to sudo group')
421 | system(glue('sudo adduser {user} sudo'))
422 |
423 | }
424 | ```
425 |
426 | Note, you may have to temporarily enable passwordless `sudo` for this user (if have not done already) :/
427 |
428 | ```
429 | ceu ALL=(ALL) NOPASSWD:ALL
430 | ```
431 |
432 | Check users:
433 |
434 | ```
435 | readLines('/etc/passwd')
436 | ```
437 |
438 | ### 💪 Install Jenkins to schedule R commands
439 |
440 | 
441 |
442 | 1. Install Jenkins from the RStudio/Terminal: https://www.jenkins.io/doc/book/installing/linux/#debianubuntu
443 |
444 | ```sh
445 | sudo apt install -y fontconfig openjdk-17-jre
446 |
447 | sudo wget -O /usr/share/keyrings/jenkins-keyring.asc \
448 | https://pkg.jenkins.io/debian-stable/jenkins.io-2023.key
449 | echo deb [signed-by=/usr/share/keyrings/jenkins-keyring.asc] \
450 | https://pkg.jenkins.io/debian-stable binary/ | sudo tee \
451 | /etc/apt/sources.list.d/jenkins.list > /dev/null
452 | sudo apt-get update
453 | sudo apt-get install -y jenkins
454 |
455 | # check which port is open by java (jenkins)
456 | sudo ss -tapen | grep java
457 | ```
458 |
459 | 2. Open up port 8080 in the related security group
460 | 3. Access Jenkins from your browser and finish installation
461 |
462 | 1. Read the initial admin password from RStudio/Terminal via
463 |
464 | ```sh
465 | sudo cat /var/lib/jenkins/secrets/initialAdminPassword
466 | ```
467 |
468 | 2. Proceed with installing the suggested plugins
469 | 3. Create your first user (eg `ceu`)
470 |
471 | Note that if loading Jenkins after getting a new IP takes a lot of time, it might be due to
472 | not be able to load the `theme.css` as trying to search for that on the previous IP (as per
473 | Jenkins URL setting). To overcome this, wait 2 mins for the `theme.css` timeout, login, disable
474 | the dark theme: https://github.com/jenkinsci/dark-theme-plugin/issues/458
475 |
476 | ### 💪 Update Jenkins for shared usage
477 |
478 | Update the security backend to use real Unix users for shared access (if users already created):
479 |
480 | 
481 |
482 | And allow `jenkins` to authenticate UNIX users and restart:
483 |
484 | ```sh
485 | sudo adduser jenkins shadow
486 | sudo systemctl restart jenkins
487 | ```
488 |
489 | Then make sure to test new user access in an incognito window to avoid closing yourself out :)
490 |
491 | ### 💪 Set up an easy to remember IP address
492 |
493 | Optionally you can associate a fixed IP address to your box:
494 |
495 | 1. Allocate a new Elastic IP address at https://eu-west-1.console.aws.amazon.com/ec2/v2/home?region=eu-west-1#Addresses:
496 | 2. Name this resource by assigning a "Name" tag
497 | 3. Associate this Elastic IP with your stopped box, then start it
498 |
499 | ### 💪 Set up an easy to remember domain name
500 |
501 | Optionally you can associate a subdomain with your node, using the above created Elastic IP address:
502 |
503 | 1. Go to Route 53: https://console.aws.amazon.com/route53/home
504 | 2. Go to Hosted Zones and click on `ceudata.net`
505 | 3. Create a new Record, where
506 |
507 | - fill in the desired `Name` (subdomain), eg `de3.ceudata.net`
508 | - paste the public IP address or hostname of your server in the `Value` field
509 | - click `Create`
510 |
511 | 4. Now you will be able to access your box using this custon (sub)domain, no need to remember IP addresses.
512 |
513 | ### 💪 Configuring for standard ports
514 |
515 | To avoid using ports like `8787` and `8080` (and get blocked by the firewall installed on the CEU WiFi), let's configure our services to listen on the standard 80 (HTTP) and potentially on the 443 (HTTPS) port as well, and serve RStudio on the `/rstudio`, and Jenkins on the `/jenkins` path.
516 |
517 | For this end, we will use Nginx as a reverse-proxy, so let's install it first:
518 |
519 | ```shell
520 | sudo apt install -y nginx
521 | ```
522 |
523 | First, we need to edit the Nginx config to enable websockets for Shiny apps etc in `/etc/nginx/nginx.conf` under the `http` section:
524 |
525 | ```
526 | map $http_upgrade $connection_upgrade {
527 | default upgrade;
528 | '' close;
529 | }
530 | ```
531 |
532 | Then we need to edit the main site's configuration at `/etc/nginx/sites-enabled/default` to act as a proxy, which also do some transformations, eg rewriting the URL (removing the `/rstudio` path) before hitting RStudio Server:
533 |
534 | ```
535 | server {
536 | listen 80;
537 | rewrite ^/rstudio$ $scheme://$http_host/rstudio/ permanent;
538 | location /rstudio/ {
539 | rewrite ^/rstudio/(.*)$ /$1 break;
540 | proxy_pass http://localhost:8787;
541 | proxy_redirect http://localhost:8787/ $scheme://$http_host/rstudio/;
542 | proxy_http_version 1.1;
543 | proxy_set_header Upgrade $http_upgrade;
544 | proxy_set_header Connection $connection_upgrade;
545 | proxy_read_timeout 20d;
546 | }
547 | }
548 | ```
549 |
550 | And restart Nginx:
551 |
552 | ```shell
553 | sudo systemctl restart nginx
554 | ```
555 |
556 | Find more information at https://support.rstudio.com/hc/en-us/articles/200552326-Running-RStudio-Server-with-a-Proxy.
557 |
558 | Let's see if the port is open on the machine:
559 |
560 | ```shell
561 | sudo ss -tapen|grep LIST
562 | ```
563 |
564 | Let's see if we can access RStudio Server on the new path:
565 |
566 | ```shell
567 | curl localhost/rstudio
568 | ```
569 |
570 | Now let's see from the outside world ... and realize that we need to open up port 80!
571 |
572 | Now we need to tweak the config to support Jenkins as well, but the above Nginx rewrite hack will not work (see https://www.jenkins.io/doc/book/system-administration/reverse-proxy-configuration-troubleshooting/ for more details), so we will just make it a standard reverse-proxy, eg:
573 |
574 | ```
575 | server {
576 | listen 80;
577 | rewrite ^/rstudio$ $scheme://$http_host/rstudio/ permanent;
578 | location / {
579 |
580 | }
581 | location /rstudio/ {
582 | rewrite ^/rstudio/(.*)$ /$1 break;
583 | proxy_pass http://localhost:8787;
584 | proxy_redirect http://localhost:8787/ $scheme://$http_host/rstudio/;
585 | proxy_http_version 1.1;
586 | proxy_set_header Upgrade $http_upgrade;
587 | proxy_set_header Connection $connection_upgrade;
588 | proxy_read_timeout 20d;
589 | }
590 | location ^~ /jenkins/ {
591 | proxy_pass http://127.0.0.1:8080/jenkins/;
592 | proxy_set_header X-Real-IP $remote_addr;
593 | proxy_set_header X-Forwarded-For $remote_addr;
594 | proxy_set_header Host $host;
595 | }
596 | }
597 | ```
598 |
599 | And we also need to let Jenkins also know about the custom path, so uncomment `Environment="JENKINS_PREFIX=/jenkins"` in `/lib/systemd/system/jenkins.service`, then reload the Systemd configs and restart Jenkins:
600 |
601 | ```shell
602 | sudo systemctl daemon-reload
603 | sudo systemctl restart jenkins
604 | ```
605 |
606 | See more details at the [Jenkins reverse proxy guide](https://www.jenkins.io/doc/book/system-administration/reverse-proxy-configuration-with-jenkins/reverse-proxy-configuration-nginx/).
607 |
608 | Optionally, replace the default, system-wide `index.html` for folks visiting the root domain without either the `rstudio` or `jenkins` path (note that instead of the editing the file, which might be overwritten with package updates, it would be better to create a new HTML file and refer that from the Nginx configuration, but we will keep it simple and dirty for now):
609 |
610 | ```shell
611 | echo "Welcome to DE3! Are you looking for /rstudio or /jenkins?" | sudo tee /usr/share/nginx/html/index.html
612 | ```
613 |
614 | Then restart Jenkins, and good to go!
615 |
616 | It might be useful to also proxy port 8000 for future use via updating the Nginx config to:
617 |
618 | ```
619 | server {
620 | listen 80;
621 | rewrite ^/rstudio$ $scheme://$http_host/rstudio/ permanent;
622 | location / {
623 |
624 | }
625 | location /rstudio/ {
626 | rewrite ^/rstudio/(.*)$ /$1 break;
627 | proxy_pass http://localhost:8787;
628 | proxy_redirect http://localhost:8787/ $scheme://$http_host/rstudio/;
629 | proxy_http_version 1.1;
630 | proxy_set_header Upgrade $http_upgrade;
631 | proxy_set_header Connection $connection_upgrade;
632 | proxy_read_timeout 20d;
633 | }
634 | location ^~ /jenkins/ {
635 | proxy_pass http://127.0.0.1:8080/jenkins/;
636 | proxy_set_header X-Real-IP $remote_addr;
637 | proxy_set_header X-Forwarded-For $remote_addr;
638 | proxy_set_header Host $host;
639 | }
640 | location ^~ /8000/ {
641 | rewrite ^/8000/(.*)$ /$1 break;
642 | proxy_pass http://127.0.0.1:8000;
643 | proxy_set_header X-Real-IP $remote_addr;
644 | proxy_set_header X-Forwarded-For $remote_addr;
645 | proxy_set_header Host $host;
646 | }
647 | }
648 | ```
649 |
650 | This way you can access the above services via the below URLs:
651 |
652 | RStudio Server:
653 |
654 | * http://your.ip.address:8787
655 | * http://your.ip.address/rstudio
656 |
657 | Jenkins:
658 |
659 | * http://your.ip.address:8080/jenkins
660 | * http://your.ip.address/jenkins
661 |
662 | Port 8000:
663 |
664 | * http://your.ip.address:8000
665 | * http://your.ip.address/8000
666 |
667 | If you cannot access RStudio Server on port 80, you might need to restart `nginx` as per above.
668 |
669 | Next, set up SSL either with Nginx or placing an AWS Load Balancer in front of the EC2 node.
670 |
671 | ### Warmup exercises
672 |
673 | Install `ggplot2` or your preferred Python packages to replicate the below steps that we will automate later:
674 |
675 | 1. Install the `devtools` R package and a few others (binary distribution) in the RStudio/Terminal:
676 |
677 | ```sh
678 | sudo apt-get install -y r-cran-devtools r-cran-data.table r-cran-httr r-cran-jsonlite r-cran-data.table r-cran-stringi r-cran-stringr r-cran-glue r-cran-logger r-cran-snakecase
679 | ```
680 |
681 | 2. Switch back to the R console and install the `binancer` R package from GitHub to interact with crypto exchanges (note the extra dependency to be installed from CRAN, no need to update any already installed package):
682 |
683 | ```r
684 | devtools::install_github('daroczig/binancer', upgrade = FALSE)
685 | ```
686 |
687 | 3. First steps with live data: load the `binancer` package and then use the `binance_klines` function to get the last 3 hours of Bitcoin price changes (in USD) with 1-minute granularity -- resulting in an object like:
688 |
689 | ```r
690 | > str(klines)
691 | Classes ‘data.table’ and 'data.frame': 180 obs. of 12 variables:
692 | $ open_time : POSIXct, format: "2020-03-08 20:09:00" "2020-03-08 20:10:00" "2020-03-08 20:11:00" "2020-03-08 20:12:00" ...
693 | $ open : num 8292 8298 8298 8299 8298 ...
694 | $ high : num 8299 8299 8299 8299 8299 ...
695 | $ low : num 8292 8297 8297 8298 8296 ...
696 | $ close : num 8298 8298 8299 8298 8299 ...
697 | $ volume : num 25.65 9.57 20.21 9.65 24.69 ...
698 | $ close_time : POSIXct, format: "2020-03-08 20:09:59" "2020-03-08 20:10:59" "2020-03-08 20:11:59" "2020-03-08 20:12:59" ...
699 | $ quote_asset_volume : num 212759 79431 167677 80099 204883 ...
700 | $ trades : int 371 202 274 186 352 271 374 202 143 306 ...
701 | $ taker_buy_base_asset_volume : num 13.43 5.84 11.74 7.12 15.24 ...
702 | $ taker_buy_quote_asset_volume: num 111430 48448 97416 59071 126493 ...
703 | $ symbol : chr "BTCUSDT" "BTCUSDT" "BTCUSDT" "BTCUSDT" ...
704 | - attr(*, ".internal.selfref")=
705 | ```
706 |
707 | Click here for the code generating the above ...
708 |
709 | ```r
710 | library(binancer)
711 | klines <- binance_klines('BTCUSDT', interval = '1m', limit = 60*3)
712 | str(klines)
713 | summary(klines$close)
714 | ```
715 |
716 |
717 | 4. Visualize the data, eg on a simple line chart:
718 |
719 | 
720 |
721 | Click here for the code generating the above ...
722 |
723 | ```r
724 | library(ggplot2)
725 | ggplot(klines, aes(close_time, close)) + geom_line()
726 | ```
727 |
728 |
729 | 5. Now create a candle chart, something like:
730 |
731 | 
732 |
733 | Click here for the code generating the above ...
734 |
735 | ```r
736 | library(scales)
737 | ggplot(klines, aes(open_time)) +
738 | geom_linerange(aes(ymin = open, ymax = close, color = close < open), size = 2) +
739 | geom_errorbar(aes(ymin = low, ymax = high), size = 0.25) +
740 | theme_bw() + theme('legend.position' = 'none') + xlab('') +
741 | ggtitle(paste('Last Updated:', Sys.time())) +
742 | scale_y_continuous(labels = dollar) +
743 | scale_color_manual(values = c('#1a9850', '#d73027')) # RdYlGn
744 | ```
745 |
746 |
747 | 6. Compare prices of 4 currencies (eg BTC, ETH, BNB and XRP) in the past 24 hours on 15 mins intervals:
748 |
749 | 
750 |
751 | Click here for the code generating the above ...
752 |
753 | ```r
754 | library(data.table)
755 | klines <- rbindlist(lapply(
756 | c('BTCUSDT', 'ETHUSDT', 'BNBUSDT', 'XRPUSDT'),
757 | binance_klines,
758 | interval = '15m', limit = 4*24))
759 | ggplot(klines, aes(open_time)) +
760 | geom_linerange(aes(ymin = open, ymax = close, color = close < open), size = 2) +
761 | geom_errorbar(aes(ymin = low, ymax = high), size = 0.25) +
762 | theme_bw() + theme('legend.position' = 'none') + xlab('') +
763 | ggtitle(paste('Last Updated:', Sys.time())) +
764 | scale_color_manual(values = c('#1a9850', '#d73027')) +
765 | facet_wrap(~symbol, scales = 'free', nrow = 2)
766 | ```
767 |
768 |
769 |
770 |
771 | 7. Some further useful functions:
772 |
773 | - `binance_ticker_all_prices()`
774 | - `binance_coins_prices()`
775 | - `binance_credentials` and `binance_balances`
776 |
777 | 8. Create an R script that reports and/or plots on some cryptocurrencies, ideas:
778 |
779 | - compute the (relative) change in prices of cryptocurrencies in the past 24 / 168 hours
780 | - go back in time 1 / 12 / 24 months and "invest" $1K in BTC and see the value today
781 | - write a bot buying and selling crypto on a virtual exchange
782 |
783 | ### Schedule R commands
784 |
785 | Let's schedule a Jenkins job to check on the Bitcoin prices every hour!
786 |
787 | 1. Visit Jenkins using the `/jenkins` URL path of your instance's public IP address
788 | 2. Use your UNIX username and a password to log in
789 |
790 | If logging in takes a long time, it might be due to the Dark Theme plugin trying to load a CSS file from an outdated location absed on the Jenkins URL configured at `/jenkins/manage/configure`. To overcome this issue, wait it out at the above URL and update the IP address to your new dynamic IP address, or disable the plugin at `/jenkins/manage/pluginManager/installed` and then restart Jenkins at the bottom of the page.
791 |
792 | 3. Create a "New Item" (job):
793 |
794 | 1. Enter the name of the job: `get current Bitcoin price`
795 | 2. Pick "Freestyle project"
796 | 3. Click "OK"
797 | 4. Add a new "Execute shell" build step
798 | 5. Enter the below command to look up the most recent BTC price
799 |
800 | ```sh
801 | R -e "library(binancer);binance_coins_prices()[symbol == 'BTC', usd]"
802 | ```
803 |
804 | 6. Run the job
805 |
806 | 
807 |
808 | 4. Debug & figure out what's the problem ...
809 | 5. Install R packages system wide from RStudio/Terminal (more on this later):
810 |
811 | Either start R in the terminal as the root user (via `sudo R`) and run the previous `devtools::install_github` command there, or with a one-liner:
812 |
813 | ```sh
814 | sudo Rscript -e "library(devtools);withr::with_libpaths(new = '/usr/local/lib/R/site-library', install_github('daroczig/binancer', upgrade = FALSE))"
815 | ```
816 |
817 | 6. Rerun the job
818 |
819 | 
820 |
821 | ### Schedule R scripts
822 |
823 | 1. Create an R script with the below content and save on the server, eg as `/home/ceu/bitcoin-price.R`:
824 |
825 | ```r
826 | library(binancer)
827 | prices <- binance_coins_prices()
828 | paste('The current Bitcoin price is', prices[symbol == 'BTC', usd])
829 | ```
830 |
831 | 2. Follow the steps from the [Schedule R commands](#schedule-r-commands) section to create a new Jenkins job, but instead of calling `R -e "..."` in shell step, reference the above R script using `Rscript` instead:
832 |
833 | ```shell
834 | Rscript /home/ceu/de3.R
835 | ```
836 |
837 | Alternatively, you could also install little R for this purpose:
838 |
839 | ```shell
840 | sudo apt install -y r-cran-littler
841 | r /home/ceu/de3.R
842 | ```
843 |
844 | Note the permission error, so let's add the `jenkins` user to the `ceu` group:
845 |
846 | ```shell
847 | sudo adduser jenkins ceu
848 | ```
849 |
850 | Then restart Jenkins from the RStudio Server terminal:
851 |
852 | ```shell
853 | sudo systemctl restart jenkins
854 | ```
855 |
856 | A better solution will be later to commit our R script into a git repo, and make it part of the job to update from the repo.
857 |
858 | 3. Create an R script that generates a candlestick chart on the BTC prices from the past hour, saves as `btc.png` in the workspace, and update every 5 minutes!
859 |
860 | Example solution for the above ...
861 |
862 | ```r
863 | library(binancer)
864 | library(ggplot2)
865 | library(scales)
866 | klines <- binance_klines('BTCUSDT', interval = '1m', limit = 60)
867 | g <- ggplot(klines, aes(open_time)) +
868 | geom_linerange(aes(ymin = open, ymax = close, color = close < open), size = 2) +
869 | geom_errorbar(aes(ymin = low, ymax = high), size = 0.25) +
870 | theme_bw() + theme('legend.position' = 'none') + xlab('') +
871 | ggtitle(paste('Last Updated:', Sys.time())) +
872 | scale_y_continuous(labels = dollar) +
873 | scale_color_manual(values = c('#1a9850', '#d73027'))
874 | ggsave('btc.png', plot = g, width = 10, height = 5)
875 | ```
876 |
877 |
878 | 1. Enter the name of the job: `Update BTC candlestick chart`
879 | 2. Pick "Freestyle project"
880 | 3. Click "OK"
881 | 4. Add a new "Execute shell" build step
882 | 5. Enter the below command to look up the most recent BTC price
883 |
884 | ```sh
885 | Rscript /home/ceu/plot.R
886 | ```
887 | 6. Run the job
888 | 7. Look at the workspace that can be accessed from the sidebar menu of the job.
889 |
890 | ### ScheduleR improvements
891 |
892 | 1. Use a git repository to store the R scripts and fetch the most recent version on job start:
893 |
894 | 1. Configure the Jenkins job to use "Git" in the "Source Code Management" section, and use e.g. https://gist.github.com/daroczig/e5d3ee3664549932bb7f23ce8e93e472 as the repo URL, and specify the branch (`main`).
895 | 2. Update the Execute task section to refer to the `btcprice.R` file of the repo instead of the hardcoded local path.
896 | 3. Make edits to the repo, e.g. update lookback to 3 hours and check a future job output.
897 |
898 | 2. Set up e-mail notifications via eg https://app.mailjet.com/signin
899 |
900 | 1. 💪 Sign up, confirm your e-mail address and domain
901 | 2. 💪 Take a note on the SMTP settings, eg
902 |
903 | * SMTP server: in-v3.mailjet.com
904 | * Port: 465
905 | * SSL: Yes
906 | * Username: ***
907 | * Password: ***
908 |
909 | 3. 💪 Configure Jenkins at http://de3.ceudata.net/jenkins/configure
910 |
911 | 1. Set up the default FROM e-mail address at "System Admin e-mail address": jenkins@ceudata.net
912 | 2. Search for "Extended E-mail Notification" and configure
913 |
914 | * SMTP Server
915 | * Click "Advanced"
916 | * Check "Use SMTP Authentication"
917 | * Enter User Name from the above steps
918 | * Enter Password from the above steps
919 | * Check "Use SSL"
920 | * SMTP port: 465
921 |
922 | 5. Set up "Post-build Actions" in Jenkins: Editable Email Notification - read the manual and info popups, configure to get an e-mail on job failures and fixes
923 | 6. Configure the job to send the whole e-mail body as the deault body template for all outgoing emails
924 |
925 | ```shell
926 | ${BUILD_LOG, maxLines=1000}
927 | ```
928 |
929 | 3. Look at other Jenkins plugins, eg the Slack Notifier: https://plugins.jenkins.io/slack
930 |
931 | ### Intro to redis
932 |
933 | We need a persistent storage for our Jenkins jobs ... let's give a try to a key-value database:
934 |
935 | 1. 💪 Install server
936 |
937 | ```
938 | sudo apt install redis-server
939 | ss -tapen | grep LIST
940 | ```
941 |
942 | Test using the CLI tool:
943 |
944 | ```
945 | redis-cli get foo
946 | redis-cli set foo 42
947 | redis-cli get foo
948 | redis-cli del foo
949 | ```
950 |
951 | 2. 💪 Install an R client
952 |
953 | Although we could use the `RcppRedis` available on CRAN, so thus also in our already added apt repo, but `rredis` provides some convenient helpers that we plan to use, so we are going to install that as well from a custom R package repository to also demonstrate how `drat` works:
954 |
955 | ```
956 | sudo apt install -y r-cran-rcppredis
957 | sudo Rscript -e "withr::with_libpaths(new = '/usr/local/lib/R/site-library', install.packages('rredis', repos=c('https://ghrr.github.io/drat', 'https://cloud.r-project.org')))"
958 | ```
959 |
960 | 3. Interact from R
961 |
962 | ```r
963 | ## set up and initialize the connection to the local redis server
964 | library(rredis)
965 | redisConnect()
966 |
967 | ## set/get values
968 | redisSet('foo', 'bar')
969 | redisGet('foo')
970 |
971 | ## increment and decrease counters
972 | redisIncr('counter')
973 | redisIncr('counter')
974 | redisIncr('counter')
975 | redisGet('counter')
976 | redisDecr('counter')
977 | redisDecr('counter2')
978 |
979 | ## get multiple values at once
980 | redisMGet(c('counter', 'counter2'))
981 |
982 | ## list all keys
983 | redisKeys()
984 |
985 | ## drop all keys
986 | redisDelete(redisKeys())
987 | ```
988 |
989 | For more examples and ideas, see the [`rredis` package vignette](https://rdrr.io/cran/rredis/f/inst/doc/rredis.pdf) or try the interactive, genaral (not R-specific) [redis tutorial](https://try.redis.io).
990 |
991 | 4. Exercises
992 |
993 | - Create a Jenkins job running every minute to cache the most recent Bitcoin and Ethereum prices in Redis
994 | - Write an R script in RStudio that can read the Bitcoin and Ethereum prices from the Redis cache
995 |
996 | Example solution ...
997 |
998 | ```r
999 | library(binancer)
1000 | library(data.table)
1001 | prices <- binance_coins_prices()
1002 |
1003 | library(rredis)
1004 | redisConnect()
1005 |
1006 | redisSet('username:price:BTC', prices[symbol == 'BTC', usd])
1007 | redisSet('username:price:ETH', prices[symbol == 'ETH', usd])
1008 |
1009 | redisGet('username:price:BTC')
1010 | redisGet('username:price:ETH')
1011 |
1012 | redisMGet(c('username:price:BTC', 'username:price:ETH'))
1013 | ```
1014 |
1015 |
1016 | Example solution using a helper function doing some logging ...
1017 |
1018 | ```r
1019 | library(binancer)
1020 | library(logger)
1021 | library(rredis)
1022 | redisConnect()
1023 |
1024 | store <- function(s) {
1025 | ## TODO use the checkmate pkg to assert the type of symbol
1026 | log_info('Looking up and storing {s}')
1027 | value <- binance_coins_prices()[symbol == s, usd]
1028 | key <- paste('username', 'price', s, sep = ':')
1029 | redisSet(key, value)
1030 | log_info('The price of {s} is {value}')
1031 | }
1032 |
1033 | store('BTC')
1034 | store('ETH')
1035 |
1036 | ## list all keys with the "price" prefix and lookup the actual values
1037 | redisMGet(redisKeys('username:price:*'))
1038 | ```
1039 |
1040 |
1041 | More on databases at the "Mastering R" class in the Spring semester ;)
1042 |
1043 | ### Interacting with MS Teams
1044 |
1045 | 1. Join the #bots-bots-bots channel in the DE3 course's MS Teams
1046 | 2. Click on "Connectors" in the channel's config and add an incoming webhook with your username and optional logo, store the URL for later use
1047 | 3. 💪 Install the `teamr` package from CRAN
1048 |
1049 | ```shell
1050 | sudo apt install -y r-cran-teamr
1051 | ```
1052 |
1053 | 4. Use the webhook URL to send messages to the channel from R:
1054 |
1055 | ```r
1056 | library(teamr)
1057 | webhook_url <- "https://ceuedu.webhook.office.com/webhookb2/..."
1058 | # create new connector card
1059 | cc <- connector_card$new(hookurl = webhook_url)
1060 | # set the title and text of the message
1061 | cc$title("Hi from R!")
1062 | cc$text("This is a test message sent using the `teamr` package.")
1063 | # send the message
1064 | cc$send()
1065 | ```
1066 |
1067 | ## Week 3
1068 |
1069 | Quiz: https://forms.office.com/e/wRAxGqirdV (5 mins)
1070 |
1071 | Recap on what we covered last week:
1072 |
1073 | - Amazon Machine Image
1074 | - Shared RStudio Server
1075 | - Settup up Jenkins
1076 | - Reverse proxy for more conenvient access to services
1077 | - Scheduling R commands in Jenkins
1078 | - Scheduling R script in Jenkins
1079 | - Further Jenkins features, such as e-mail notifications and git integration
1080 | - Introduction to Redis
1081 | - Interacting with MS Teams
1082 |
1083 | New server for this week:
1084 |
1085 | 1. Go the the Instances overview at https://eu-west-1.console.aws.amazon.com/ec2/v2/home?region=eu-west-1#Instances:sort=instanceId
1086 | 2. Click "Launch Instance"
1087 | 3. Provide a name for your server (e.g. `daroczig-de3-week3`) and some additional tags for resource tracking, including tagging downstream services, such as Instance and Volumes:
1088 | * Class: `DE3`
1089 | * Owner: `daroczig`
1090 | 4. Pick the `de3` AMI
1091 | 5. Pick `t3a.small` (2 GiB of RAM should be enough for most tasks) instance type (see more [instance types](https://aws.amazon.com/ec2/instance-types))
1092 | 6. Select your AWS key created above and launch
1093 | 7. Select the `de3` security group (granting access to ports 22, 8000, 8080, and 8787)
1094 | 8. Click "Advanced details" and select `ceudataserver` IAM instance profile
1095 | 9. Click "Launch instance"
1096 | 10. Note and click on the instance id
1097 |
1098 | #### Storing the secret webhook URL
1099 |
1100 | 1. Do NOT store the webhook URL in plain-text (e.g. in your R script)!
1101 | 2. Let's use Amazon's Key Management Service: https://github.com/daroczig/CEU-R-prod/raw/2017-2018/AWR.Kinesis/AWR.Kinesis-talk.pdf (slides 73-75)
1102 | 3. 💪 Instead of using the Java SDK referenced in the above talk, let's install `boto3` Python module and use via `reticulate`:
1103 |
1104 | ```shell
1105 | sudo apt install -y python3-boto3
1106 | sudo apt install -y r-cran-reticulate r-cran-botor
1107 | ```
1108 |
1109 | Let's also let R know that we want to use the globally installed Python interpreter and its packages instead of setting up local virtual environments by adding the following to your `/etc/R/Renviron.site` file:
1110 |
1111 | ```shell
1112 | RETICULATE_PYTHON=/usr/bin/python3
1113 | ```
1114 |
1115 | 4. 💪 Create a key in the Key Management Service (KMS): `alias/de3`
1116 | 5. 💪 Grant access to that KMS key by creating an EC2 IAM role at https://console.aws.amazon.com/iam/home?region=eu-west-1#/roles with the `AWSKeyManagementServicePowerUser` policy and explicit grant access to the key in the KMS console
1117 | 6. 💪 Attach the newly created IAM role if not yet done
1118 | 7. Use this KMS key to encrypt the Slack token:
1119 |
1120 | ```r
1121 | library(botor)
1122 | botor(region = 'eu-west-1')
1123 | kms_encrypt(webhook_url, key = 'alias/de3')
1124 | ```
1125 |
1126 | Note, if R asks you to install Miniconda, say NO, as Python3 and the required packages have already been installed system-wide.
1127 |
1128 | 8. Store the ciphertext and use `kms_decrypt` to decrypt later, see eg
1129 |
1130 | ```r
1131 | kms_decrypt("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")
1132 | ```
1133 |
1134 | 9. 💪 Alternatively, use the AWS Parameter Store or Secrets Manager, see eg https://eu-west-1.console.aws.amazon.com/systems-manager/parameters/?region=eu-west-1&tab=Table and granting the `AmazonSSMReadOnlyAccess` policy to your IAM role or user.
1135 |
1136 | ```r
1137 | ssm_get_parameter('/teams/daroczig')
1138 | ```
1139 |
1140 | 10. Store your own webhook in the Parameter Store and use it in your R script.
1141 |
1142 | ### Job Scheduler exercises
1143 |
1144 | * Create a Jenkins job to alert if Bitcoin price is below $80K or higher than $100K.
1145 | * Limit this job to alert only once per hour, but alert should kick off as soon as possible (so don't schedule to run hourly, instead use a state to track when the last alert was sent).
1146 | * Create a Jenkins job to alert if Bitcoin price changed more than $200 in the past hour.
1147 | * Create a Jenkins job to alert if Bitcoin price changed more than 5% in the past day.
1148 | * Create a Jenkins job running hourly to generate a candlestick chart on the price of BTC and ETH.
1149 |
1150 | Example solution for the first exercise ...
1151 |
1152 | ```r
1153 | ## get data right from the Binance API
1154 | library(binancer)
1155 | btc <- binance_klines('BTCUSDT', interval = '1m', limit = 1)$close
1156 |
1157 | ## or from the local cache (updated every minute from Jenkins as per above)
1158 | library(rredis)
1159 | redisConnect()
1160 | btc <- redisGet('username:price:BTC')
1161 |
1162 | ## log whatever was retreived
1163 | library(logger)
1164 | log_info('The current price of a Bitcoin is ${btc}')
1165 |
1166 | ## get the last alert time
1167 | last_alert <- redisGet('username:alert:last')
1168 | if (is.null(last_alert)) {
1169 | last_alert <- 0
1170 | }
1171 | since_last_alert <- as.numeric(
1172 | difftime(Sys.time(), last_alert,
1173 | units = "secs"))
1174 |
1175 | ## send alert
1176 | if (since_last_alert >= 60 && (btc < 80000 | btc > 100000)) {
1177 | library(botor)
1178 | botor(region = 'eu-west-1')
1179 | webhook_url <- ssm_get_parameter('/teams/username')
1180 | library(teamr)
1181 | cc <- connector_card$new(hookurl = webhook_url)
1182 | cc$title('Bitcoin price alert!')
1183 | cc$text(paste('The current price of a Bitcoin is:', btc))
1184 | cc$send()
1185 | }
1186 | ```
1187 |
1188 |
1189 |
1190 | ### Make API endpoints
1191 |
1192 | 1. 💪 Install plumber: [rplumber.io](https://www.rplumber.io)
1193 |
1194 | ```sh
1195 | sudo apt install -y r-cran-plumber
1196 | ```
1197 |
1198 | 2. Create an API endpoint to show the min, max and mean price of a BTC in the past hour!
1199 |
1200 | Create `~/plumber.R` with the below content:
1201 |
1202 | ```r
1203 | library(binancer)
1204 |
1205 | #* BTC stats
1206 | #* @get /btc
1207 | function() {
1208 | klines <- binance_klines('BTCUSDT', interval = '1m', limit = 60L)
1209 | klines[, .(min = min(close), mean = mean(close), max = max(close))]
1210 | }
1211 | ```
1212 |
1213 | Start the plumber application wither via clicking on the "Run API" button or the below commands:
1214 |
1215 | ```r
1216 | library(plumber)
1217 | pr("plumber.R") %>% pr_run(host='0.0.0.0', port=8000)
1218 | ```
1219 |
1220 | 3. Add a new API endpoint to generate the candlestick chart with dynamic symbol (default to BTC), interval and limit! Note that you might need a new `@serializer`, function arguments, and type conversions as well.
1221 |
1222 | Example solution for the above ...
1223 |
1224 | ```r
1225 | library(binancer)
1226 | library(ggplot2)
1227 | library(scales)
1228 |
1229 | #* Generate plot
1230 | #* @param symbol coin pair
1231 | #* @param interval:str enum
1232 | #* @param limit integer
1233 | #* @get /klines
1234 | #* @serializer png
1235 | function(symbol = 'BTCUSDT', interval = '1m', limit = 60L) {
1236 | klines <- binance_klines(symbol, interval = interval, limit = as.integer(limit)) # NOTE int conversion
1237 | library(scales)
1238 | p <- ggplot(klines, aes(open_time)) +
1239 | geom_linerange(aes(ymin = open, ymax = close, color = close < open), size = 2) +
1240 | geom_errorbar(aes(ymin = low, ymax = high), size = 0.25) +
1241 | theme_bw() + theme('legend.position' = 'none') + xlab('') +
1242 | ggtitle(paste('Last Updated:', Sys.time())) +
1243 | scale_y_continuous(labels = dollar) +
1244 | scale_color_manual(values = c('#1a9850', '#d73027')) # RdYlGn
1245 | print(p)
1246 | }
1247 | ```
1248 |
1249 |
1250 | 4. Add a new API endpoint to generate a HTML report including both the above!
1251 |
1252 | Example solution for the above ...
1253 |
1254 | 💪 Update the `markdown` package:
1255 |
1256 | ```shell
1257 | sudo apt install -y r-cran-markdown
1258 | ```
1259 |
1260 | Create an R markdown for the reporting:
1261 |
1262 | `````md
1263 | ---
1264 | title: "report"
1265 | output: html_document
1266 | date: "`r Sys.Date()`"
1267 | ---
1268 |
1269 | ```{r setup, include=FALSE}
1270 | knitr::opts_chunk$set(echo = FALSE, warning=FALSE)
1271 | library(binancer)
1272 | library(ggplot2)
1273 | library(scales)
1274 | library(knitr)
1275 |
1276 | klines <- function() {
1277 | binance_klines('BTCUSDT', interval = '1m', limit = 60L)
1278 | }
1279 | ```
1280 |
1281 | Bitcoin stats:
1282 |
1283 | ```{r stats}
1284 | kable(klines()[, .(min = min(close), mean = mean(close), max = max(close))])
1285 | ```
1286 |
1287 | On a nice plot:
1288 |
1289 | ```{r plot}
1290 | ggplot(klines(), aes(open_time, )) +
1291 | geom_linerange(aes(ymin = open, ymax = close, color = close < open), size = 2) +
1292 | geom_errorbar(aes(ymin = low, ymax = high), size = 0.25) +
1293 | theme_bw() + theme('legend.position' = 'none') + xlab('') +
1294 | ggtitle(paste('Last Updated:', Sys.time())) +
1295 | scale_y_continuous(labels = dollar) +
1296 | scale_color_manual(values = c('#1a9850', '#d73027'))
1297 | ```
1298 | `````
1299 |
1300 | And the plumber file:
1301 |
1302 | ```r
1303 | library(binancer)
1304 | library(ggplot2)
1305 | library(scales)
1306 | library(rmarkdown)
1307 | library(plumber)
1308 |
1309 | #' Gets BTC data from the past hour
1310 | #' @return data.table
1311 | klines <- function() {
1312 | binance_klines('BTCUSDT', interval = '1m', limit = 60L)
1313 | }
1314 |
1315 | #* BTC stats
1316 | #* @get /stats
1317 | function() {
1318 | klines()[, .(min = min(close), mean = mean(close), max = max(close))]
1319 | }
1320 |
1321 | #* Generate plot
1322 | #* @get /plot
1323 | #* @serializer png
1324 | function() {
1325 | p <- ggplot(klines(), aes(open_time, )) +
1326 | geom_linerange(aes(ymin = open, ymax = close, color = close < open), size = 2) +
1327 | geom_errorbar(aes(ymin = low, ymax = high), size = 0.25) +
1328 | theme_bw() + theme('legend.position' = 'none') + xlab('') +
1329 | ggtitle(paste('Last Updated:', Sys.time())) +
1330 | scale_y_continuous(labels = dollar) +
1331 | scale_color_manual(values = c('#1a9850', '#d73027')) # RdYlGn
1332 | print(p)
1333 | }
1334 |
1335 | #* Generate HTML
1336 | #* @get /report
1337 | #* @serializer html
1338 | function(res) {
1339 | filename <- tempfile(fileext = '.html')
1340 | on.exit(unlink(filename))
1341 | render('report.Rmd', output_file = filename)
1342 | include_file(filename, res)
1343 | }
1344 | ```
1345 |
1346 | Run via:
1347 |
1348 | ```r
1349 | library(plumber)
1350 | pr('plumber.R') %>% pr_run(port = 8000)
1351 | ```
1352 |
1353 |
1354 | Try to DRY (don't repeat yourself!) this up as much as possible.
1355 |
1356 | ### R API containers
1357 |
1358 | Why API? Why R-based API? Examples
1359 |
1360 | * adtech
1361 | * healthtech
1362 |
1363 | 1. Write an R script that provides 3 API endpoints (look up examples from past week!):
1364 |
1365 | * `/stats` reports on the min/mean/max BTC price from the past 3 hours
1366 | * `/plot` generates a candlestick chart on the price of BTC from past 3 hours
1367 | * `/report` generates a HTML report including both the above
1368 |
1369 | Example solution for the above ...
1370 |
1371 | 💪 Update the `markdown` package:
1372 |
1373 | ```shell
1374 | sudo apt install -y r-cran-markdown
1375 | ```
1376 |
1377 | Create an R markdown for the reporting:
1378 |
1379 | `````md
1380 | ---
1381 | title: "report"
1382 | output: html_document
1383 | date: "`r Sys.Date()`"
1384 | ---
1385 |
1386 | ```{r setup, include=FALSE}
1387 | knitr::opts_chunk$set(echo = FALSE, warning=FALSE)
1388 | library(binancer)
1389 | library(ggplot2)
1390 | library(scales)
1391 | library(knitr)
1392 |
1393 | klines <- function() {
1394 | binance_klines('BTCUSDT', interval = '1m', limit = 60L)
1395 | }
1396 | ```
1397 |
1398 | Bitcoin stats:
1399 |
1400 | ```{r stats}
1401 | kable(klines()[, .(min = min(close), mean = mean(close), max = max(close))])
1402 | ```
1403 |
1404 | On a nice plot:
1405 |
1406 | ```{r plot}
1407 | ggplot(klines(), aes(open_time, )) +
1408 | geom_linerange(aes(ymin = open, ymax = close, color = close < open), size = 2) +
1409 | geom_errorbar(aes(ymin = low, ymax = high), size = 0.25) +
1410 | theme_bw() + theme('legend.position' = 'none') + xlab('') +
1411 | ggtitle(paste('Last Updated:', Sys.time())) +
1412 | scale_y_continuous(labels = dollar) +
1413 | scale_color_manual(values = c('#1a9850', '#d73027'))
1414 | ```
1415 | `````
1416 |
1417 | And the plumber file:
1418 |
1419 | ```r
1420 | library(binancer)
1421 | library(ggplot2)
1422 | library(scales)
1423 | library(rmarkdown)
1424 | library(plumber)
1425 |
1426 | #' Gets BTC data from the past hour
1427 | #' @return data.table
1428 | klines <- function() {
1429 | binance_klines('BTCUSDT', interval = '1m', limit = 60L)
1430 | }
1431 |
1432 | #* BTC stats
1433 | #* @get /stats
1434 | function() {
1435 | klines()[, .(min = min(close), mean = mean(close), max = max(close))]
1436 | }
1437 |
1438 | #* Generate plot
1439 | #* @get /plot
1440 | #* @serializer png
1441 | function() {
1442 | p <- ggplot(klines(), aes(open_time, )) +
1443 | geom_linerange(aes(ymin = open, ymax = close, color = close < open), size = 2) +
1444 | geom_errorbar(aes(ymin = low, ymax = high), size = 0.25) +
1445 | theme_bw() + theme('legend.position' = 'none') + xlab('') +
1446 | ggtitle(paste('Last Updated:', Sys.time())) +
1447 | scale_y_continuous(labels = dollar) +
1448 | scale_color_manual(values = c('#1a9850', '#d73027')) # RdYlGn
1449 | print(p)
1450 | }
1451 |
1452 | #* Generate HTML
1453 | #* @get /report
1454 | #* @serializer html
1455 | function(res) {
1456 | filename <- tempfile(fileext = '.html')
1457 | on.exit(unlink(filename))
1458 | render('report.Rmd', output_file = filename)
1459 | include_file(filename, res)
1460 | }
1461 | ```
1462 |
1463 | Run via:
1464 |
1465 | ```r
1466 | library(plumber)
1467 | pr('plumber.R') %>% pr_run(port = 8000)
1468 | ```
1469 |
1470 |
1471 | 2. Try to DRY (don't repeat yourself!) this up as much as possible.
1472 | 3. Bundle all the scripts into a single Docker image:
1473 |
1474 | a. 💪 Install Docker:
1475 |
1476 | ```shell
1477 | curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /etc/apt/keyrings/docker.gpg
1478 | echo \
1479 | "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.gpg] \
1480 | https://download.docker.com/linux/ubuntu \
1481 | $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
1482 | sudo apt-get update
1483 | sudo apt-get install -y docker-ce
1484 | ```
1485 |
1486 | b. Create a new file named `Dockerfile` (File/New file/Text file to avoid auto-adding the `R` file extension) with the below content to add the required files and set the default working directory to the same folder:
1487 |
1488 | ```
1489 | FROM rstudio/plumber
1490 |
1491 | RUN apt-get update && apt-get install -y pandoc && apt-get clean && rm -rf /var/lib/apt/lists/
1492 | RUN install2.r ggplot2
1493 | RUN installGithub.r daroczig/binancer
1494 | ADD report.Rmd /app/report.Rmd
1495 | ADD plumber.R /app/plumber.R
1496 | EXPOSE 8000
1497 | WORKDIR /app
1498 | ```
1499 |
1500 | Note the step installing the required R packages!
1501 |
1502 | c. Build the Docker image:
1503 |
1504 | ```sh
1505 | sudo docker build -t btc-report-api .
1506 | ```
1507 |
1508 | d. Run a container based on the above image:
1509 |
1510 | ```sh
1511 | sudo docker run -p 8000:8000 -ti btc-report-api plumber.R
1512 | ```
1513 |
1514 | e. Realize that it's not working, as we need to install `rmarkdown` package, so update the above `Dockerfile` on line 4, rebuild the image and run again.
1515 |
1516 | f. Test by visiting the `8000` port or the Nginx proxy at http://de3.ceudata.net/8000, e.g. Swagger docs at http://de3.ceudata.net/8000/__docs__/#/default/get_report or an endpoint directly at http://de3.ceudata.net/8000/report.
1517 |
1518 | ### Docker registry
1519 |
1520 | Now let's make the above created and tested Docker image available outside of the RStudio Server by uploading the Docker image to Elastic Container Registry (ECR):
1521 |
1522 | 1. 💪 Create a new private repository at https://eu-west-1.console.aws.amazon.com/ecr/home?region=eu-west-1, call it `de3-example-api`
1523 | 2. 💪 Assign the `EC2InstanceProfileForImageBuilderECRContainerBuilds` policy to the `ceudataserver` IAM role so that we get RW access to the ECR repositories. Tighten this role up in prod!
1524 | 3. 💪 Let's login to ECR on the RStudio Server so that we can upload the Docker image:
1525 |
1526 | ```sh
1527 | aws ecr get-login-password --region eu-west-1 | sudo docker login --username AWS --password-stdin 657609838022.dkr.ecr.eu-west-1.amazonaws.com
1528 | ```
1529 |
1530 | 4. 💪 Tag the already build Docker image for upload:
1531 |
1532 | ```sh
1533 | sudo docker tag btc-report-api:latest 657609838022.dkr.ecr.eu-west-1.amazonaws.com/de3-example-api:latest
1534 | ```
1535 |
1536 | 5. 💪 Push the Docker image:
1537 |
1538 | ```sh
1539 | sudo docker push 657609838022.dkr.ecr.eu-west-1.amazonaws.com/de3-example-api:latest
1540 | ```
1541 |
1542 | 6. Check the Docker repository at https://eu-west-1.console.aws.amazon.com/ecr/repositories/private/657609838022/de3-example-api?region=eu-west-1
1543 |
1544 | ### Docker service
1545 |
1546 | 1. Go to the Elastic Container Service (ECS) dashboard at https://eu-west-1.console.aws.amazon.com/ecs/home?region=eu-west-1#/
1547 | 2. Create a task definition for the Docker run:
1548 |
1549 | 1. Task name: `btc-api`
1550 | 2. Container name: `api`
1551 | 3. Image URI: `657609838022.dkr.ecr.eu-west-1.amazonaws.com/de3-example-api`
1552 | 4. Container port: 8000
1553 | 5. Command in the Docker configuration: `plumber.R`
1554 | 6. Review Task size, but default values should fine for this simple task
1555 |
1556 | 3. Create a new cluster, call it `BTC_API`, using Fargate. Don't forget to add the `Class` tag!
1557 | 4. Create a Service in the newly created Cluster at https://eu-west-1.console.aws.amazon.com/ecs/v2/clusters/btc-api/services?region=eu-west-1
1558 |
1559 | 1. Compute option can be "Launch type" for now
1560 | 2. Specify the Task Family as `btc-api`
1561 | 3. Provide the same as service name
1562 | 4. Use the `de3` security group
1563 | 5. Create a load balancer listening on port 80 (would need to create an SSL cert for HTTPS), and specify `/stats` as the healthcheck path, with a 10 seconds grace period
1564 | 6. Test the deployed service behind the load balancer, e.g. https://btc-api-1417435399.eu-west-1.elb.amazonaws.com/report
1565 |
1566 | ## Home assignment
1567 |
1568 | The goal of this assignment is to confirm that you have a general understanding on how to build data pipelines using Amazon Web Services and R, and can actually implement a stream processing application (either running in almost real-time or batched/scheduled way) or R-based API in practice.
1569 |
1570 | ### Tech setup
1571 |
1572 | To minimize the system administration and some of the already-covered engineering tasks for the students, the below pre-configured tools are provided as free options, but students can decide to build their own environment (on the top of or independently from these) and feel free to use any other tools:
1573 |
1574 | * `de3` Amazon Machine Image that you can use to spin up an EC2 node with RStudio Server, Shiny Server, Jenkins, Redis and Docker installed & pre-configured (use your AWS username and the password shared on Slack previously) along with the most often used R packages (including the ones we used for stream processing, eg `botor`, `AWR.Kinesis` and the `binancer` package)
1575 | * `de3` EC2 IAM role with full access to Kinesis, Dynamodb, Cloudwatch and the `slack` token in the Parameter Store
1576 | * `de3` security group with open ports for RStudio Server and Jenkins
1577 | * lecture and seminar notes at https://github.com/daroczig/CEU-R-prod
1578 |
1579 | ### Required output
1580 |
1581 | Make sure to clean-up your EC2 nodes, security groups, keys etc created in the past weeks, as left-over AWS resources will contribute negative points to your final grade! E.g. the EC2 node you created on the second week should be terminated.
1582 |
1583 | * Minimal project (for grade up to "B"): schedule a Jenkins job that runs every hour getting the past hour's 1-minute interval klines data on ETH prices (in USD). The job should be configured to pull the R script at the start of the job either from a private or public git repo or gist. Then
1584 |
1585 | * Find the min and max price of ETH in the past hour, and post these stats in the `#bots-bots-bots` MS Teams channel. Make sure to set your username for the message, and use a custom emoji as the icon.
1586 | * Set up email notification for the job when it fails.
1587 |
1588 | * Recommended project (for grade up to "A"): Deploy an R-based API in ECS (like we did on the last week) for analyzing recent Binance (or any other real-time) data. The API should include at least 4 endpoints using different serializers, and these endpoints should be other than the ones we covered in the class. **At least one endpoint should have at least a few parameters.** Build a Docker image, push it to ECR, and deploy as service in ECS. Document the steps required to set up ECR/ECS with screenshots, then delete all services after confirming that everything works correctly.
1589 |
1590 | Regarding feedback: by default, I add a super short feedback on Moodle as a comment to your submission (e.g. "good job" or "excellent" for grade A, or short details on why it was not A). If you want to receive more detailed feedback, please send me an email to schedule a quick call. If you want early feedback (before grading), send me an email at least a week before the submission deadline!
1591 |
1592 | ### Delivery method
1593 |
1594 | * Create a PDF document that describes your solution and all the main steps involved with low level details: attach screenshots (including the URL nav bar and the date/time widget of your OS, so like full-screen and not area-picked screenshots) of your browser showing what you are doing in RStudio Server, Jenkins, in the AWS dashboards, or example messages posted in MS Teams, and make sure that the code you wrote is either visible on the screenshots, or included in the PDF.
1595 |
1596 | * STOP the EC2 Instance you worked on, but don’t terminate it, so I can start it and check how it works. Note that your instance will be terminated by me after the end of the class.
1597 | * Include the `instance_id` on the first page of the PDF, along with your name or student id.
1598 | * Upload the PDF to Moodle.
1599 |
1600 | ### Submission deadline
1601 |
1602 | Midnight (CET) on May 31, 2025.
1603 |
1604 | ## Extra: Stream processing using R and AWS
1605 |
1606 | An introduction to stream processing with AWS Kinesis and R: https://github.com/daroczig/CEU-R-prod/raw/2017-2018/AWR.Kinesis/AWR.Kinesis-talk.pdf (presented at the Big Data Day Los Angeles 2016, EARL 2016 London and useR! 2017 Brussels)
1607 |
1608 | This section describes how to set up a Kinesis stream with a few on-demand shards on the live Binance transactions read from its websocket -- running in a Docker container, then feeding the JSON lines to Kinesis via the Amazon Kinesis Agent.
1609 |
1610 | ### 💪 Setting up a demo stream
1611 |
1612 | 1. Start a `t3a.micro` instance running "Amazon Linux 2 AMI" (where it's easier to install the Kinesis Agent compared to using eg Ubuntu) with a known key. Make sure to **set a name** and enable termination protection (in the instance details)! Use SSH, Putty or eg the browser-based SSH connection. Note that the default username is `ec2-user` instead of `ubuntu`.
1613 |
1614 | 2. Install Docker (note that we are not on Ubuntu today, but using Red Hat's `yum` package manager):
1615 |
1616 | ```
1617 | sudo yum install docker
1618 | sudo service docker start
1619 | sudo service docker status
1620 | ```
1621 |
1622 | 3. Let's use a small Python app relying on the Binance API to fetch live transactions and store in a local file:
1623 |
1624 | * sources: https://github.com/daroczig/ceu-de3-docker-binance-streamer
1625 | * docker: https://cloud.docker.com/repository/registry-1.docker.io/daroczig/ceu-de3-docker-binance-streamer
1626 |
1627 | Usage:
1628 |
1629 | ```
1630 | screen -RRd streamer
1631 | sudo docker run -ti --rm --log-opt max-size=50m daroczig/ceu-de3-docker-binance-streamer >> /tmp/transactions.json
1632 | ## "C-a c" to create a new screen, then you can switch with C-a "
1633 | ls -latr /tmp
1634 | tail -f /tmp/transactions.json
1635 | ```
1636 |
1637 | 4. Install the Kinesis Agent:
1638 |
1639 | As per https://docs.aws.amazon.com/firehose/latest/dev/writing-with-agents.html#download-install:
1640 |
1641 | ```
1642 | sudo yum install -y aws-kinesis-agent
1643 | ```
1644 |
1645 | 5. Create a new Kinesis Stream (called `crypto`) at https://eu-west-1.console.aws.amazon.com/kinesis. Don't forget to tag it (Class, Owner)!
1646 |
1647 | 6. Configure the Kinesis Agent:
1648 |
1649 | ```
1650 | sudo yum install mc
1651 | sudo mcedit /etc/aws-kinesis/agent.json
1652 | ```
1653 |
1654 | Running the above commands, edit the config file to update the Kinesis endpoint, the name of the stream on the local file path:
1655 |
1656 | ```
1657 | {
1658 | "cloudwatch.emitMetrics": true,
1659 | "kinesis.endpoint": "https://kinesis.eu-west-1.amazonaws.com",
1660 | "firehose.endpoint": "",
1661 |
1662 | "flows": [
1663 | {
1664 | "filePattern": "/tmp/transactions.json*",
1665 | "kinesisStream": "crypto",
1666 | "partitionKeyOption": "RANDOM"
1667 | }
1668 | ]
1669 | }
1670 | ```
1671 |
1672 | Note that extra star at the end of the `filePattern` to handle potential issues when file is copy/truncated (logrotate).
1673 |
1674 | 7. Restart the Agent:
1675 |
1676 | ```
1677 | sudo service aws-kinesis-agent start
1678 | ```
1679 |
1680 | 8. Check the status and logs:
1681 |
1682 | ```
1683 | sudo service aws-kinesis-agent status
1684 | sudo journalctl -xe
1685 | ls -latr /var/log/aws-kinesis-agent/aws-kinesis-agent.log
1686 | tail -f /var/log/aws-kinesis-agent/aws-kinesis-agent.log
1687 | ```
1688 |
1689 | 9. Make sure that the IAM role (eg `kinesis-admin`) can write to Kinesis and Cloudwatch, eg by attaching the `AmazonKinesisFullAccess` policy, then restart the agent
1690 |
1691 | ```
1692 | sudo service aws-kinesis-agent restart
1693 | ```
1694 |
1695 | 10. Check the AWS console's monitor if all looks good there as well
1696 | 11. Note for the need of permissions to `cloudwatch:PutMetricData` (see example `cloudwatch-putmetrics` policy).
1697 | 12. Optionally set up a cronjob to truncate that the file from time to time:
1698 |
1699 | ```sh
1700 | 5 * * * * /usr/bin/truncate -s 0 /tmp/transactions.json
1701 | ```
1702 |
1703 | 13. Set up an alert in Cloudwatch if streaming stops
1704 |
1705 | ### A simple stream consumer app in R
1706 |
1707 | As the `botor` package was already installed, we can rely on the power of `boto3` to interact with the Kinesis stream. The IAM role attached to the node already has the `AmazonKinesisFullAccess` policy attached, so we have permissions to read from the stream.
1708 |
1709 | First we need to create a shard iterator, then using that, we can read the actual records from the shard:
1710 |
1711 | ```r
1712 | library(botor)
1713 | botor(region = 'eu-west-1')
1714 | shard_iterator <- kinesis_get_shard_iterator('crypto', '0')
1715 | records <- kinesis_get_records(shard_iterator$ShardIterator)
1716 | str(records)
1717 | ```
1718 |
1719 | Let's parse these records:
1720 |
1721 | ```r
1722 | records$Records[[1]]
1723 | records$Records[[1]]$Data
1724 |
1725 | library(jsonlite)
1726 | fromJSON(as.character(records$Records[[1]]$Data))
1727 | ```
1728 |
1729 | ### Parsing and structuring records read from the stream
1730 |
1731 | Exercises:
1732 |
1733 | * parse the loaded 25 records into a `data.table` object with proper column types. Get some help on the data format from the [Binance API docs](https://github.com/binance/binance-spot-api-docs/blob/master/web-socket-streams.md#trade-streams)!
1734 | * count the overall number of coins exchanged
1735 | * count the overall value of transactions in USD (hint: `binance_ticker_all_prices()` and `binance_coins_prices()`)
1736 | * visualize the distribution of symbol pairs
1737 |
1738 | A potential solution that you should not look at before thinking ...
1739 |
1740 | ```shell
1741 | library(data.table)
1742 | dt <- rbindlist(lapply(records$Records, function(record) {
1743 | fromJSON(as.character(record$Data))
1744 | }))
1745 |
1746 | str(dt)
1747 |
1748 | setnames(dt, 'a', 'seller_id')
1749 | setnames(dt, 'b', 'buyer_id')
1750 | setnames(dt, 'E', 'event_timestamp')
1751 | ## Unix timestamp / Epoch (number of seconds since Jan 1, 1970): https://www.epochconverter.com
1752 | dt[, event_timestamp := as.POSIXct(event_timestamp / 1000, origin = '1970-01-01')]
1753 | setnames(dt, 'q', 'quantity')
1754 | setnames(dt, 'p', 'price')
1755 | setnames(dt, 's', 'symbol')
1756 | setnames(dt, 't', 'trade_id')
1757 | setnames(dt, 'T', 'trade_timestamp')
1758 | dt[, trade_timestamp := as.POSIXct(trade_timestamp / 1000, origin = '1970-01-01')]
1759 | str(dt)
1760 |
1761 | for (id in grep('_id', names(dt), value = TRUE)) {
1762 | dt[, (id) := as.character(get(id))]
1763 | }
1764 | str(dt)
1765 |
1766 | for (v in c('quantity', 'price')) {
1767 | dt[, (v) := as.numeric(get(v))]
1768 | }
1769 |
1770 | library(binancer)
1771 | binance_coins_prices()
1772 |
1773 | dt[, .N, by = symbol]
1774 | dt[symbol=='ETHUSDT']
1775 | dt[, from := substr(symbol, 1, 3)]
1776 | dt <- merge(dt, binance_coins_prices(), by.x = 'from', by.y = 'symbol', all.x = TRUE, all.y = FALSE)
1777 | dt[, value := as.numeric(quantity) * usd]
1778 | dt[, sum(value)]
1779 | ```
1780 |
1781 |
1782 | ### Actual stream processing instead of analyzing batch data
1783 |
1784 | Let's write an R function to increment counters on the number of transactions per symbols:
1785 |
1786 | 1. Get sample raw data as per above (you might need to get a new shard iterator if expired):
1787 |
1788 | ```r
1789 | records <- kinesis_get_records(shard_iterator$ShardIterator)$Record
1790 | ```
1791 |
1792 | 2. Function to parse and process it
1793 |
1794 | ```r
1795 | txprocessor <- function(record) {
1796 | symbol <- fromJSON(as.character(record$Data))$s
1797 | log_info(paste('Found 1 transaction on', symbol))
1798 | redisIncr(paste('USERNAME', 'tx', symbol, sep = ':'))
1799 | }
1800 | ```
1801 |
1802 | 3. Iterate on all records
1803 |
1804 | ```r
1805 | library(logger)
1806 | library(rredis)
1807 | redisConnect()
1808 | for (record in records) {
1809 | txprocessor(record)
1810 | }
1811 | ```
1812 |
1813 | 4. Check counters
1814 |
1815 | ```r
1816 | symbols <- redisMGet(redisKeys('^USERNAME:tx:*'))
1817 | symbols
1818 |
1819 | symbols <- data.frame(
1820 | symbol = sub('^USERNAME:tx:', '', names(symbols)),
1821 | N = as.numeric(symbols))
1822 | symbols
1823 | ```
1824 |
1825 | 5. Visualize
1826 |
1827 | ```r
1828 | library(ggplot2)
1829 | ggplot(symbols, aes(symbol, N)) + geom_bar(stat = 'identity')
1830 | ```
1831 |
1832 | 6. Rerun step (1) and (3) to do the data processing, then (4) and (5) for the updated data visualization.
1833 |
1834 | 7. 🤦
1835 |
1836 | 8. Let's make use of the next shard iterator:
1837 |
1838 | ```r
1839 | ## reset counters
1840 | redisDelete(redisKeys('USERNAME:tx:*'))
1841 |
1842 | ## get the first shard iterator
1843 | shard_iterator <- kinesis_get_shard_iterator('crypto', '0')$ShardIterator
1844 |
1845 | while (TRUE) {
1846 |
1847 | response <- kinesis_get_records(shard_iterator)
1848 |
1849 | ## get the next iterator
1850 | shard_iterator <- response$NextShardIterator
1851 |
1852 | ## extract records
1853 | records <- response$Record
1854 | for (record in records) {
1855 | txprocessor(record)
1856 | }
1857 |
1858 | ## summarize
1859 | symbols <- redisMGet(redisKeys('USERNAME:tx:*'))
1860 | symbols <- data.frame(
1861 | symbol = sub('^symbol:', '', names(symbols)),
1862 | N = as.numeric(symbols))
1863 |
1864 | ## visualize
1865 | print(ggplot(symbols, aes(symbol, N)) + geom_bar(stat = 'identity') + ggtitle(sum(symbols$N)))
1866 | }
1867 | ```
1868 |
1869 | ### Stream processor daemon
1870 |
1871 | 0. So far, we used the `boto3` Python module from R via `botor` to interact with AWS, but this time we will integrate Java -- by calling the AWS Java SDK to interact with our Kinesis stream, then later on to run a Java daemon to manage our stream processing application.
1872 |
1873 | 1. 💪 First, let's install Java and the `rJava` R package:
1874 |
1875 | ```shell
1876 | sudo apt install r-cran-rjava
1877 | ```
1878 |
1879 | 2. 💪 Then the R package wrapping the AWS Java SDK and the Kinesis client, then update to the most recent dev version right away:
1880 |
1881 | ```shell
1882 | sudo apt install r-cran-awr.kinesis
1883 | sudo R -e "withr::with_libpaths(new = '/usr/local/lib/R/site-library', install.packages('AWR', repos = 'https://daroczig.gitlab.io/AWR'))"
1884 | sudo R -e "withr::with_libpaths(new = '/usr/local/lib/R/site-library', devtools::install_github('daroczig/AWR.Kinesis', upgrade = FALSE))"
1885 | ```
1886 |
1887 | 3. 💪 Note, after installing Java, you might need to run `sudo R CMD javareconf` and/or restart R or the RStudio Server via `sudo rstudio-server restart` :/
1888 |
1889 | ```shell
1890 | Error : .onLoad failed in loadNamespace() for 'rJava', details:
1891 | call: dyn.load(file, DLLpath = DLLpath, ...)
1892 | error: unable to load shared object '/usr/lib/R/site-library/rJava/libs/rJava.so':
1893 | libjvm.so: cannot open shared object file: No such file or directory
1894 | ```
1895 |
1896 | 4. And after all, a couple lines of R code to get some data from the stream via the Java SDK (just like we did above with the Python backend):
1897 |
1898 | ```r
1899 | library(rJava)
1900 | library(AWR.Kinesis)
1901 | records <- kinesis_get_records('crypto', 'eu-west-1')
1902 | str(records)
1903 | records[1]
1904 |
1905 | library(jsonlite)
1906 | fromJSON(records[1])
1907 | ```
1908 |
1909 | 1. Create a new folder for the Kinesis consumer files: `streamer`
1910 |
1911 | 2. Create an `app.properties` file within that subfolder
1912 |
1913 | ```
1914 | executableName = ./app.R
1915 | regionName = eu-west-1
1916 | streamName = crypto
1917 | applicationName = my_demo_app_sadsadsa
1918 | AWSCredentialsProvider = DefaultAWSCredentialsProviderChain
1919 | ```
1920 |
1921 | 3. Create the `app.R` file:
1922 |
1923 | ```r
1924 | #!/usr/bin/Rscript
1925 | library(logger)
1926 | log_appender(appender_file('app.log'))
1927 | library(AWR.Kinesis)
1928 | library(methods)
1929 | library(jsonlite)
1930 |
1931 | kinesis_consumer(
1932 |
1933 | initialize = function() {
1934 | log_info('Hello')
1935 | library(rredis)
1936 | redisConnect(nodelay = FALSE)
1937 | log_info('Connected to Redis')
1938 | },
1939 |
1940 | processRecords = function(records) {
1941 | log_info(paste('Received', nrow(records), 'records from Kinesis'))
1942 | for (record in records$data) {
1943 | symbol <- fromJSON(record)$s
1944 | log_info(paste('Found 1 transaction on', symbol))
1945 | redisIncr(paste('symbol', symbol, sep = ':'))
1946 | }
1947 | },
1948 |
1949 | updater = list(
1950 | list(1/6, function() {
1951 | log_info('Checking overall counters')
1952 | symbols <- redisMGet(redisKeys('symbol:*'))
1953 | log_info(paste(sum(as.numeric(symbols)), 'records processed so far'))
1954 | })),
1955 |
1956 | shutdown = function()
1957 | log_info('Bye'),
1958 |
1959 | checkpointing = 1,
1960 | logfile = 'app.log')
1961 | ```
1962 |
1963 | 4. 💪 Allow writing checkpointing data to DynamoDB and CloudWatch in IAM
1964 |
1965 | 5. Convert the above R script into an executable using the Terminal:
1966 |
1967 | ```shell
1968 | cd streamer
1969 | chmod +x app.R
1970 | ```
1971 |
1972 | 6. Run the app in the Terminal:
1973 |
1974 | ```
1975 | /usr/bin/java -cp /usr/local/lib/R/site-library/AWR/java/*:/usr/local/lib/R/site-library/AWR.Kinesis/java/*:./ \
1976 | com.amazonaws.services.kinesis.multilang.MultiLangDaemon \
1977 | ./app.properties
1978 | ```
1979 |
1980 | 7. Check on `app.log`
1981 |
1982 | ### Shiny app showing the progress
1983 |
1984 | 1. Reset counters
1985 |
1986 | ```r
1987 | library(rredis)
1988 | redisConnect()
1989 | keys <- redisKeys('symbol*')
1990 | redisDelete(keys)
1991 | ```
1992 |
1993 | 2. 💪 Install the `treemap` package
1994 |
1995 | ```
1996 | sudo apt install r-cran-httpuv r-cran-shiny r-cran-xtable r-cran-htmltools r-cran-igraph r-cran-lubridate r-cran-tidyr r-cran-quantmod r-cran-broom r-cran-zoo r-cran-htmlwidgets r-cran-tidyselect r-cran-rlist r-cran-rlang r-cran-xml r-cran-treemap r-cran-highcharter
1997 | ```
1998 |
1999 | 3. Run the below Shiny app
2000 |
2001 | ```r
2002 | ## packages for plotting
2003 | library(treemap)
2004 | library(highcharter)
2005 |
2006 | ## connect to Redis
2007 | library(rredis)
2008 | redisConnect()
2009 |
2010 | library(shiny)
2011 | library(data.table)
2012 | ui <- shinyUI(highchartOutput('treemap', height = '800px'))
2013 | server <- shinyServer(function(input, output, session) {
2014 |
2015 | symbols <- reactive({
2016 |
2017 | ## auto-update every 2 seconds
2018 | reactiveTimer(2000)()
2019 |
2020 | ## get frequencies
2021 | symbols <- redisMGet(redisKeys('symbol:*'))
2022 | symbols <- data.table(
2023 | symbol = sub('^symbol:', '', names(symbols)),
2024 | N = as.numeric(symbols))
2025 |
2026 | ## color top 3
2027 | symbols[, color := 1]
2028 | symbols[symbol %in% symbols[order(-N)][1:3, symbol], color := 2]
2029 |
2030 | ## return
2031 | symbols
2032 |
2033 | })
2034 |
2035 | output$treemap <- renderHighchart({
2036 | tm <- treemap(symbols(), index = c('symbol'),
2037 | vSize = 'N', vColor = 'color',
2038 | type = 'value', draw = FALSE)
2039 | N <- sum(symbols()$N)
2040 | hc_title(hctreemap(tm, animation = FALSE),
2041 | text = sprintf('Transactions (N=%s)', N))
2042 | })
2043 |
2044 | })
2045 | shinyApp(ui = ui, server = server, options = list(port = 3838))
2046 | ```
2047 |
2048 | We will learn more about Shiny in the upcoming Data Visualization 4 class :)
2049 |
2050 |
2051 |
2052 | Will be updated from week to week.
2053 |
2054 | ## Getting help
2055 |
2056 | File a [GitHub ticket](https://github.com/daroczig/CEU-R-prod/issues).
2057 |
--------------------------------------------------------------------------------