Python generation script for BitBirds

Overview

BitBirds generation script

Intro

This is published under MIT license, which means you can do whatever you want with it - entirely at your own risk.

Please don't be an asshole. This is, like, grassroots and stuff.

Specifically I'm asking you in good faith not to directly knock off the BitBirds project, or otherwise screw me over for sharing this. Do not use this for anything hateful or discriminatory.

There is a YouTube video walkthrough to complement this ReadMe ...Link....

Setting the expectations

If you're new to programming you may struggle to set up the dependencies. If you're persistent, you can do it! I believe in you.

Often in technology, setting up a pre-requisite like PIP (a python asset installation tool) isn't something the developer thinks about in a given project because it has been on their computer for months or years.

Even having set up a number of dependencies just a few weeks ago for this project, I don't remember exactly how I worked through the various error messages. When you try to run the script, if it fails there will often be some useful nugget of information buried in the cryptic response blob. As a rule for life - Google is your friend, and others have probably encountered your exact error message. When asking questions on discord, stackoverflow, or wherever, say very specifically (1) what you've tried (2) what you expect as the result and (3) what issue/error you're encountering. That'll get a lot more useful feedback than just shouting HALP.

Dependencies

The dependenices were all installed with the terminal/command line. There is documentation abound about terminal generally, and these tools specifically, but unfortunately I did not save copies of the web pages I used. From memory the things I needed to setup were:

  • Python 3 (default on my mac was python 2.7)

  • PIP - a command-line installation mechanism for python assets.

IIRC I needed to use some special command with python 3 to use pip as an installation mechanism for the items below- perhaps pip3 install ... rather than pip install ...?

  • Pillow - python library to generate images - installed via terminal/command-line with PIP

  • NumPy - python library to work with arrays - installed via terminal/command-line with PIP

  • I don't think I had to install the 'random' library (included in the script) to use the number-randomization feature, but might have.

If you encounter specific setup items I haven't mentioned here let me know, and I'll add them.

How this script works

The video I've put on YouTube complements this overview.

We are iterating through a 'loop' once for each bird. The loop starts with a 'seed number' that is used to deterministically generate pseudo-random numbers. I say 'deterministically' and 'pseudo-random' because from the same seed number the 'random' output will always be the same. It's not truly random in a security or mathematical sense. I used the most recent ETH block at the time as my seed number - 11981207.

There is then a 'chain' of additional random numbers generated that are used to define all of the various traits of the birds. Many of the attributes generate a random number between 1-1000 and use that for some sort of logical statement (e.g. to decide beak color).

Interestingly, the way I've used this random-number chain seems to have resulted in some specific behavior and combinations that I can't yet personally explain. For example the way the bird type selection random number and beak color selection random number chain from one another seems to have resulted in no red-beaked woodpeckers. I also noticed that all of the four cockatoos generated seemed to have the exact same blue crest. Because I wanted more variety than that, in the minted NFTs I ran another batch of cockatoos and replaced the second, third, and fourth cockatoos in the original sequence with others that provided more variety. I made no other changes to the randomly generated set of birds, and if you run the script yourself (without changing the seed) you should see identical matches for each number. If you spot a pattern in why the random numbers behave in this way, I would love to discuss on twitter or discord.

  • Head color and throat color are a random 1-255 number generated into each of the three RGB values in a color.
  • Eye color looks at a random number between 1-1000 and if it's 47 or less, will give the bird crazy eyes. Crazy eyes always have the same pupil color (154, 0, 0) and then generate a random color for the 'white of the eye,' in the same way the head and throat color are generated.
  • Beak color is determined by an evaluation of another 1-1000 'random' number. Grey is most common, gold is also common, red is rare, and black is very rare.

The bird images are 24x24 arrays of variables, representing every pixel in the final image. I've used variables with two letters for each type of pixel (e.g. outline ol, head color hd, beak color bk), so as to keep the 'matrix' of pixel variables easy to see and work with. If you zoom way out on the code you may even be able to see a rough picture of the birds in the code, just from the slight differences in the variables.

The script uses another 1-1000 'random' number to decide which of the bird type templates to use. Basic bird is most common, at about 75% probability. Jay has a 15% probability, woodpecker has 6%, eagle has 3.5%, and cockatoo has half a percent probability.

From there, you're just about home free. The final bit of the loop re-sizes the generated bird from 24x24 pixels up to 480x480 pixels. It generates the image file path (dynamically, wherever you have the folder using the os library), and saves the image into the included folder bird_images.

Then it goes right back to the top of the loop, and does it again for the next bird, until the number of defined loops is completed.

Wrap up

I sincerely hope this inspires someone to learn a new skill, take up coding, or generally expand their horizons! I won't profess to be a professional coder, but I am a technologist in my day job and have found it to be a fulfilling and rewarding life path. This BitBirds project has been a joy to be involved in. The community that has popped up around it already has been inspiring, and I'm excited to see it grow in the years to come.

If would like to show your thanks for this shared asset I'd encourage you to plant some trees! https://onetreeplanted.org/collections/all

If you feel absolutly compelled to send ETH or NFTs to me directly, please know that it is not necessary, but the BitBirds project hardware wallet address is: 0x1fd146a5e6152c5ACd3A013fBC42A243e4DfCe63

Thanks for everything!

Owner
BitBirds generation script!
Final Project for the Intel AI Readiness Boot Camp NLP (Jan)

NLP Boot Camp (Jan) Synopsis Full Name: Prameya Mohanty Name of your School: Delhi Public School, Rourkela Class: VIII Title of the Project: iTransect

TheCodingHub 1 Feb 01, 2022
Reading Wikipedia to Answer Open-Domain Questions

DrQA This is a PyTorch implementation of the DrQA system described in the ACL 2017 paper Reading Wikipedia to Answer Open-Domain Questions. Quick Link

Facebook Research 4.3k Jan 01, 2023
Différents programmes créant une interface graphique a l'aide de Tkinter pour simplifier la vie des étudiants.

GP211-Grand-Projet Ce repertoire contient tout les programmes nécessaires au bon fonctionnement de notre projet-logiciel. Cette interface graphique es

1 Dec 21, 2021
Code for "Generating Disentangled Arguments with Prompts: a Simple Event Extraction Framework that Works"

GDAP The code of paper "Code for "Generating Disentangled Arguments with Prompts: a Simple Event Extraction Framework that Works"" Event Datasets Prep

45 Oct 29, 2022
An Explainable Leaderboard for NLP

ExplainaBoard: An Explainable Leaderboard for NLP Introduction | Website | Download | Backend | Paper | Video | Bib Introduction ExplainaBoard is an i

NeuLab 319 Dec 20, 2022
A Non-Autoregressive Transformer based TTS, supporting a family of SOTA transformers with supervised and unsupervised duration modelings. This project grows with the research community, aiming to achieve the ultimate TTS.

A Non-Autoregressive Transformer based TTS, supporting a family of SOTA transformers with supervised and unsupervised duration modelings. This project grows with the research community, aiming to ach

Keon Lee 237 Jan 02, 2023
Analyse japanese ebooks using MeCab to determine the difficulty level for japanese learners

japanese-ebook-analysis This aim of this project is to make analysing the contents of a japanese ebook easy and streamline the process for non-technic

Christoffer Aakre 14 Jul 23, 2022
An IVR Chatbot which can exponentially reduce the burden of companies as well as can improve the consumer/end user experience.

IVR-Chatbot Achievements 🏆 Team Uhtred won the Maverick 2.0 Bot-a-thon 2021 organized by AbInbev India. ❓ Problem Statement As we all know that, lot

ARYAMAAN PANDEY 9 Dec 08, 2022
Code for the paper "BERT Loses Patience: Fast and Robust Inference with Early Exit".

Patience-based Early Exit Code for the paper "BERT Loses Patience: Fast and Robust Inference with Early Exit". NEWS: We now have a better and tidier i

Kevin Canwen Xu 54 Jan 04, 2023
NL. The natural language programming language.

NL A Natural-Language programming language. Built using Codex. A few examples are inside the nl_projects directory. How it works Write any code in pur

2 Jan 17, 2022
मराठी भाषा वाचविण्याचा एक प्रयास. इंग्रजी ते मराठीचा शब्दकोश. An attempt to preserve the Marathi language. A lightweight and ad free English to Marathi thesaurus.

For English, scroll down मराठी शब्द मराठी भाषा वाचवण्यासाठी मी हा ओपन सोर्स प्रोजेक्ट सुरू केला आहे. माझ्या मते, आपली भाषा हळूहळू आणि कोणाचाही लक्षात

मुक्त स्त्रोत 20 Oct 11, 2022
This Project is based on NLTK It generates a RANDOM WORD from a predefined list of words, From that random word it read out the word, its meaning with parts of speech , its antonyms, its synonyms

This Project is based on NLTK(Natural Language Toolkit) It generates a RANDOM WORD from a predefined list of words, From that random word it read out the word, its meaning with parts of speech , its

SaiVenkatDhulipudi 2 Nov 17, 2021
Repository for fine-tuning Transformers 🤗 based seq2seq speech models in JAX/Flax.

Seq2Seq Speech in JAX A JAX/Flax repository for combining a pre-trained speech encoder model (e.g. Wav2Vec2, HuBERT, WavLM) with a pre-trained text de

Sanchit Gandhi 21 Dec 14, 2022
This repository contains the code, models and datasets discussed in our paper "Few-Shot Question Answering by Pretraining Span Selection"

Splinter This repository contains the code, models and datasets discussed in our paper "Few-Shot Question Answering by Pretraining Span Selection", to

Ori Ram 88 Dec 31, 2022
Official PyTorch implementation of "Dual Path Learning for Domain Adaptation of Semantic Segmentation".

Dual Path Learning for Domain Adaptation of Semantic Segmentation Official PyTorch implementation of "Dual Path Learning for Domain Adaptation of Sema

27 Dec 22, 2022
Takes a string and puts it through different languages in Google Translate a requested amount of times, returning nonsense.

PythonTextObfuscator Takes a string and puts it through different languages in Google Translate a requested amount of times, returning nonsense. Requi

2 Aug 29, 2022
Trained T5 and T5-large model for creating keywords from text

text to keywords Trained T5-base and T5-large model for creating keywords from text. Supported languages: ru Pretraining Large version | Pretraining B

Danil 61 Nov 24, 2022
A framework for evaluating Knowledge Graph Embedding Models in a fine-grained manner.

A framework for evaluating Knowledge Graph Embedding Models in a fine-grained manner.

NEC Laboratories Europe 13 Sep 08, 2022
Maix Speech AI lib, including ASR, chat, TTS etc.

Maix-Speech 中文 | English Brief Now only support Chinese, See 中文 Build Clone code by: git clone https://github.com/sipeed/Maix-Speech Compile x86x64 c

Sipeed 267 Dec 25, 2022
Unofficial Parallel WaveGAN (+ MelGAN & Multi-band MelGAN & HiFi-GAN & StyleMelGAN) with Pytorch

Parallel WaveGAN implementation with Pytorch This repository provides UNOFFICIAL pytorch implementations of the following models: Parallel WaveGAN Mel

Tomoki Hayashi 1.2k Dec 23, 2022