MariaDB 11.8 adds Vector Datatype

The lat­est Mari­aDB release adds a Vec­tor datatype, which is great for stor­ing and query­ing embed­dings. Embed­dings are basi­cal­ly a rep­re­sen­ta­tion of con­cepts for AI/GPT. I’ve used this when cre­at­ing a knowl­edge­base that can have ques­tions asked effi­cient­ly against it even at scale.

RustFS for Local Development S3-Compatible Storage

RustFS for Local Development S3-Compatible Storage

S3 and S3-com­pat­i­ble stor­age is an indis­pens­able com­po­nent of mod­ern web appli­ca­tions, espe­cial­ly when deploy­ing with con­tain­ers like Dock­er and Kubernetes.

When devel­op­ing an appli­ca­tion local­ly, although Lar­avel and oth­er frame­works have uni­fied stor­age helpers allow­ing switch­ing between file sys­tems (usu­al­ly local and S3), there are many quirks of object stor­age that will be over­looked if using a local filesystem.

Rather than use a host­ed S3 sys­tem dur­ing devel­op­ment, we can save costs, laten­cy (and time) and band­width using a local S3-com­pat­i­ble serv­er. MinIO was the go-to until recent­ly with con­tro­ver­sial license changes. RustFS is a good choice for local devel­op­ment. Here is how to set it up with Dock­er and Lar­avel for local development:

RustFS Dock­er:


1
2
3
4
5
6
7
8
9
10
11
docker run \
  -p 9000:9000 \
  -p 9001:9001 \
  -v rust-data:/data \
  -v rust-logs:/logs \
  -e RUSTFS_ACCESS_KEY=rustfsadmin \
  -e RUSTFS_SECRET_KEY=rustfsadmin \
  -e RUSTFS_CONSOLE_ENABLE=true \
  -e RUSTFS_SERVER_DOMAINS=storage.mydevdomain.test \
  -e RUSTFS_SIG_HEADER_WHITELIST="if-modified-since,range" \
  rustfs/rustfs:latest
Argu­mentMean­ing
1
dock­er run
Launch­es a new con­tain­er from the spec­i­fied image.
1
-p 9000:9000
Maps port 9000 inside the con­tain­er to port 9000 on your host. This is usu­al­ly where RustFS serves file stor­age, to be served publically.
1
-p 9001:9001
Maps port 9001 inside the con­tain­er to port 9001 on your host. RustFS web admin console.
1
-v rust-data:/data
Mounts a Dock­er vol­ume called 
1
rust-data
to 
1
/data
in the con­tain­er. This stores your per­sis­tent file data.
1
-v rust-logs:/logs
Mounts a Dock­er vol­ume called 
1
rust-logs
to 
1
/logs
in the con­tain­er. All logs gen­er­at­ed by RustFS go here.
1
-e RUSTFS_ACCESS_KEY=rustfsadmin
User­name for web admin con­sole access. Don’t use this in production!
1
-e RUSTFS_SECRET_KEY=rustfsadmin
Pass­word for web admin con­sole access. Don’t use this in production!
1
-e RUSTFS_CONSOLE_ENABLE=true
Enables the RustFS web con­sole, which allows you to mon­i­tor and man­age stor­age via a browser.
1
-e RUSTFS_SERVER_DOMAINS=storage.mydevdomain.test
Defines the domain(s) RustFS should respond to. 
1
-e RUSTFS_SIG_HEADER_WHITELIST=“if-modified-since,range”
Spec­i­fies which HTTP head­ers are allowed for signed requests. In prac­tice I’ve found this is required when serv­ing large files.
1
rustfs/rustfs:latest
The Dock­er image to run (lat­est ver­sion of RustFS).

To con­fig­ure this as a stor­age disk in Lar­avel, we can either cre­ate a new disk, or alter the S3 disk. You’ll have to decide which approach to take, which may depend on if you have mul­ti­ple disks and providers.

The admin inter­face of RustFS is very easy to use. If you are intend­ing files to be pub­lic and using Lar­avel, it is essen­tial to set your file sys­tem as pub­lic, and also set your RustFS buck­et as pub­lic. From expe­ri­ence you will just get silent errors with no error logs to help you.

config/filesystems.php


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
        's3' => [
            'driver' => 's3',
            'key' => env('AWS_ACCESS_KEY_ID'),
            'secret' => env('AWS_SECRET_ACCESS_KEY'),
            'region' => env('AWS_DEFAULT_REGION'),
            'bucket' => env('AWS_BUCKET'),
            'url' => env('AWS_URL'),
            'endpoint' => env('AWS_ENDPOINT'),
            'visibility' => 'public',
            'use_path_style_endpoint' => env('AWS_USE_PATH_STYLE_ENDPOINT', false),
            'throw' => false,
            'report' => false,
            'http' => [
                'verify' => env('AWS_SSL_VERIFY', true),
            ],
        ],

From expe­ri­ence I’ve found that RustFS requires use_path_style_endpoint, as the URL struc­ture has the buck­et name as the first path in the URL, con­trary to S3. This is the same in many com­pat­i­ble servers.

Dur­ing devel­op­ment, the option to ignore self-signed cer­tifi­cate errors is important.

The final step is to set up a reverse proxy on your favorite web serv­er to serve the files. Here’s an exam­ple con­fig in Nginx (don’t for­get to update your hosts file):


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
server {
    listen 80;
    listen 443 ssl;
    server_name storage.mydevdomain.test bucketname.storage.mydevdomain.test;

    location / {
        proxy_pass http://127.0.0.1:9900;

        # Preserve client + host info (important for S3-style services)
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;

        # RustFS / S3 compatibility tweaks
        proxy_http_version 1.1;
        proxy_set_header Connection "";

        # Large uploads support
        client_max_body_size 0;
        proxy_request_buffering off;
    }

    # SSL (Laragon)
    ssl_certificate "/etc/ssl/mycert.crt";
    ssl_certificate_key "/etc/ssl/mykey.key";
    ssl_session_timeout 5m;
    ssl_protocols TLSv1 TLSv1.1 TLSv1.2;
    ssl_ciphers ALL:!ADH:!EXPORT56:RC4+RSA:+HIGH:+MEDIUM:+LOW:+SSLv3:+EXP;
    ssl_prefer_server_ciphers on;

    charset utf-8;
}

That’s it! Now you have local S3-com­pat­i­ble stor­age for local devel­op­ment, that won’t cost you in stor­age and ingress/egress fees, has low laten­cy and won’t use up your bandwidth.

Drone 3D Survey

I always want­ed to do this, a 3D drone survey.

I took pho­tographs from my drone in var­i­ous grid pat­terns. These pho­tographs were then processed using Open­DroneMap. The process took a lot of com­put­ing pow­er, with some tak­ing 48 hours on a 24-core processor.

Mea­sure­ments can be tak­en for heights, widths, dis­tances and vol­ume accu­rate to around an inch.

You can view it online here:
https://www.jonhassall.com/drone_viewer_disused_vehicle_maintenance/viewers/disused_maintenance_area.html

Use your mouse to spin it around.

Using CSS Parameters to Stop Content Shifting with Dynamic Vue/React Content

Con­tent Shift­ing, also known as Con­tent Jump­ing, hap­pens when ele­ments of a web page change height while a page is load­ing. This can be dis­ori­en­tat­ing to users, and in extreme cas­es can cause prob­lems if a user clicks on an ele­ment as it moves, caus­ing unde­sired input. It’s also expect­ed in 2021 that search engines will begin penal­is­ing con­tent that exhibits con­tent shifting.

To avoid con­tent shift­ing, the ele­men­t’s height would be set regard­less of the dynam­ic con­tent it contains.

When using Vue or React JavaScript frame­works, I’ve found that CSS para­me­ters are use­ful. In my exam­ple, a Vue app loads dynam­ic con­tent and dis­plays it. Unfor­tu­nate­ly this caus­es con­tent shifting.

To avoid this, I pass the num­ber of rows in the HTML/view:


1
<div class="container" style="--preload-row-count: {{ $count }};">...</div>

In my CSS/SASS stylesheet, I then use this para­me­ter in a cal­cu­la­tion. There are two columns, so the num­ber of rows is divid­ed by 2. It is then mul­ti­plied by the height of a sin­gle row, and a min­i­mum height is set for the con­tain­er so that it will not shift:


1
2
3
4
.container{
    --calculated-board-rows: calc( var(--preload-row-count) / 2);
    min-height: calc( var(--calculated-board-rows) * 50px );
}

This fair­ly sim­ple method is an effec­tive way to stop con­tent shift­ing using CSS para­me­ters and calculations.

Twilio Studio — IVR and Chat Bots

Twilio Studio — IVR and Chat Bots

I’ve used Twilio for a while for pro­gram­mat­i­cal­ly send­ing and receiv­ing SMS mes­sages. There’s also a visu­al edi­tor called Stu­dio that can be used to make call and mes­sage flows:

It can be con­nect­ed to Twilio Autopi­lot to make AI-pow­ered bots. Tasks are trained with sam­ple phras­es. These sam­ple phras­es are vari­a­tions on what would be said to trig­ger an action e.g. ‘Call recep­tion,’ ‘Front desk,’ ‘Talk to a human.’

An exam­ple that comes to mind, is mak­ing a call han­dling sys­tem for an office. Rather than a voice menu that details each option fol­lowed by a num­ber, the caller could sim­ply say who they want­ed to talk to or what their request was about, and the sys­tem would han­dle it. This is far more respect­ful of the caller’s time com­pared to hav­ing them lis­ten to a long list of choices.

It works with SMS and voice calls, and seems a good way to build an IVR (Inter­ac­tive Voice Menu) sys­tem. TwiML can be used for more com­pli­cat­ed tasks, while still using Studio/Autopilot. The pric­ing is a lit­tle high­er than if you were to use a self-host­ed sys­tem, but there are so many com­pli­cat­ed func­tion­al­i­ties it seems well worth pay­ing the extra, as it would save time and reduce complexity.

I built a remote con­trolled car that used Twil­io’s cell­phone ser­vice. You can read about it here.

Mikrotik Script — Wake On LAN All Devices from DHCP

Mikrotik Script — Wake On LAN All Devices from DHCP

A lit­tle script to wake up all Mikrotik devices on your net­work that have a DHCP lease, using Wake On LAN.

It may be a good idea to set your DHCP lease expiry to some­thing long like 7 days. A short expiry may mean dynam­ic leas­es have expired if the device has been switched off for some time.

It should also wake up any devices with sta­t­ic leas­es, so you could also set all your devices as sta­t­ic in your DHCP.

1
2
3
4
5
/ip dhcp-serv­er lease
:fore­ach i in=[find] do={
:put ([get $i address].”,”.[get $i mac-address].”,”.[get $i server].”,”.[get $i host-name].”,”.[get $i expires-after])
:exe­cute { /tool wol interface=bridge [get $i mac-address] }
}