DEV Community: Sabrina Pereira

Plot FiftyOne visualisations with Seaborn

Sabrina Pereira — Mon, 01 May 2023 09:58:48 +0000

Why

I wanted more customisation options to draw figures for my thesis paper. It took me a while to figure this out, so I thought I'd share.

How

I'm assuming you already have a FiftyOne dataset with computed embeddings and visualization. If not, you'll need to create a dataset and compute the embeddings and visualization before proceeding.

I already have everything saved, so I load my dataset and the compute_visualization results before plotting:

import fiftyone as fo

# load dataset
dataset = fo.load_dataset("dataset_name")

# load computed visualisation
results = dataset.load_brain_results("vis_name")

I have a sample field called "vehicle_type" that I want to use as the hue in my seaborn plot. To obtain this information for each sample, I wrote a simple function:

def get_vehicle_type(sample_id):
    return dataset[sample_id]["vehicle_type"]

Next, I convert results.points into a pandas DataFrame and fetch the "vehicle_type" information from the FiftyOne dataset.

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt

# turn results.points into dataframe
df_viz = pd.DataFrame(results.points, columns=["x", "y"])
# get sample ids for each sample
df_viz["sample_id"] = results.sample_ids

# use sample id to get the sample field info I need
df_viz["vehicle_type"] = df_viz["sample_id"].apply(get_vehicle_type)

Finally, I plot the results using seaborn:

sns.scatterplot(data=df_viz, x='x', y='y', 
hue='vehicle_type', palette='mako_r', 
alpha=.9, s=1, edgecolor='none')
plt.title('Image Uniqueness')
plt.axis('off')
plt.show()

Seaborn allows for greater control over the appearance of the plot. Since I don't need the plot to be interactive, this is the perfect solution for creating uniform plots for my paper.

Final result

Extra: compute embeddings and visualisation

import fiftyone.zoo as foz

# compute embeddings
model = foz.load_zoo_model("mobilenet-v2-imagenet-torch")
embeddings = dataset.compute_embeddings(model)

# pickle embeddings for later use, this the computation takes a while
with open('embeddings.pkl', 'wb') as file:
    pickle.dump(embeddings, file)

# Compute visualization
results = fob.compute_visualization(
    dataset, embeddings=embeddings, seed=42, brain_key="vis_name"
)

Install and use Postgres in WSL

Sabrina Pereira — Fri, 24 Feb 2023 23:42:52 +0000

To keep this short and sweet, I'll assume you know your way around bash, Postgres and already have WSL installed.

Install Postgres

To install Postgres and run it in WSL, all you have to do is the following:

Open your WSL terminal
Update your Ubuntu packages: sudo apt update
Once the packages have updated, install PostgreSQL (and the -contrib package which has some helpful utilities) with: sudo apt install postgresql postgresql-contrib
Confirm installation and get the version number: psql --version

Set a password

The default admin user, postgres, needs a password assigned in order to connect to a database. To set a password:

Enter the command: sudo passwd postgres
You will get a prompt to enter your new password.
Close and reopen your terminal.

You can access psql directly using sudo -u postgres psql. You should see your prompt change to:

postgres=#

To change databases just use \c mydb.

You can also use su - postgres to go into the postgres user. Here you use the password you set up above. Your prompt should change to:

postgres@mycomputername:~$

From here, you can use psql to login into any database.

Creating a database

To create a database, just use the following command:

createdb mydb

You can change mydb to whatever name you want to give your database. To access it, just enter psql mydb in the command line. Now your prompt should look like this:

mydb=#

To create tables in a database from a file, use the following command:

psql -U postgres -q mydb < <file-path/file.sql>

Useful commands

\l lists all databases. Works from any database.
\dt lists all tables in the current database.
\c <db name> switch to a different database.

Use `psql` without sudo

Create a Postgres user with the same name as your Ubuntu username using the following command. And when it asks, make the new role a superuser.

$ sudo -u postgres createuser --interactive
Enter name of role to add: sabrina
Shall the new role be a superuser? (y/n) y

Then you have to change the pg_hba.conf file. It will be under /etc/postgresql/<postgres-version>/main. You will need sudo to edit this file.

sudo vi pg_hba.conf

Scroll to the bottom of the files, now change where it says peer, to trust, like so:

# Database administrative login by Unix domain socket
local   all             postgres                                trust

# TYPE  DATABASE        USER            ADDRESS                 METHOD

# "local" is for Unix domain socket connections only
local   all             all                                     trust

Accessing your database from Windows

You have to change the file postgresql.conf. Just uncomment the line for listen_address and change it to listen_address = '*'.
Set up a password for postgres admin user sudo -u postgres psql -c "ALTER USER postgres PASSWORD 'postgres';". This will change the password to postgres and this is what you use when connecting to a database. The password you set during install is for the postgres Ubuntu user.

Now you should be able to connect to your database from Windows using software such as TablePlus.

Need more?

This is just a quick overview of what to do to get up and running, for a more in-depth tutorial see this and the sources listed below.

Sources

WSL Documentation: Install PostgreSQL
Postgres documentation: 1.3. Creating a Database
StackOverflow: PostgreSQL: Why psql can't connect to server?
StackExchange: How do I list all databases and tables using psql?
StackOverflow: fatal role "root" does not exist
POSTGRESQL ON WSL2 FOR WINDOWS: INSTALL AND SETUP
StackOverflow: password authentication failed for user "postgres"

Vagrant and VirtualBox on Windows 11 and WSL

Sabrina Pereira — Mon, 13 Feb 2023 15:01:06 +0000

Download VirtualBox's latest version for Windows and its extension pack from here.

After installing VirtualBox, install the pack by going into tools > extensions then clicking install:

Install vagrant according to their instructions for Linux. All you have to do is to runt he following command in WSL:

wget -O- https://apt.releases.hashicorp.com/gpg | gpg --dearmor | sudo tee /usr/share/keyrings/hashicorp-archive-keyring.gpg
echo "deb [signed-by=/usr/share/keyrings/hashicorp-archive-keyring.gpg] https://apt.releases.hashicorp.com $(lsb_release -cs) main" | sudo tee /etc/apt/sources.list.d/hashicorp.list
sudo apt update && sudo apt install vagrant

Export some global variables to be able to run vagrant (these need to be reset each time you open a new terminal):

export VAGRANT_WSL_ENABLE_WINDOWS_ACCESS="1"
export PATH="$PATH:/mnt/c/Program Files/Oracle/VirtualBox"

You also have to add the folder where your project is located:

export VAGRANT_WSL_WINDOWS_ACCESS_USER_HOME_PATH="/mnt/c/<path-to-folder-where-Vagrantfile-is>"

And you have to ssh from PowerShell 🫠

ssh vagrant@127.0.0.1 -p 2222

Add this line to your Vagrantfile to be able to have a project in your ubuntu file system:

config.vm.synced_folder '.', '/vagrant', disabled: true

For more info on WSL installs, click here.

Miniconda in WSL

Sabrina Pereira — Tue, 27 Sep 2022 16:16:10 +0000

Install

Download the latest version of miniconda:

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh

Install it:

bash Miniconda3-latest-Linux-x86_64.sh

After Installation

To install conda's shell functions, first restart your shell, then:

conda init

If you'd prefer that conda's base environment is not activated on startup, set the auto_activate_base parameter to false:

conda config --set auto_activate_base false

Remove the installation file from your directory:

rm Miniconda3-latest-Linux-x86_64.sh

Using Miniconda

To get into your conda environment

conda activate

To create a new environment:

conda create --name <new-env-name>

To view all your environments:

conda env list

Conda cheat sheet PDF here.

Other useful things to know

Open current directory in file explorer

explorer.exe .