I could use a day off to reflect on my practice with LLM, but it is just not happening right now so persevere and sort of phone it in. In phoning it in I am generating some interesting experiments.
Today I asked:
create a gitlab-ci pipeline that does something creative with code
I am obsessed with pipelines. It is a way to divide up code based on stage, development, testing, and deployment, and to apply certain guide rails and checks to each stage. I have a whole pipeline metaphysics that I will at some point discuss. The technological society — which is our current condition operates under pipeline metaphysics.
The Result
Creating a creative GitLab CI/CD pipeline largely depends on your specific project and goals, but I can provide an example of a creative pipeline that performs code analysis, generates ASCII art from your code, and then deploys it to a static website. This is a fun and unique way to visualize your code.
1. First, make sure your repository has the following structure:
“`
project-root/
├── .gitlab-ci.yml
├── src/
│ ├── your_code.py
├── scripts/
│ ├── generate_ascii_art.sh
├── public/
“`2. Create a `.gitlab-ci.yml` file in the project’s root directory:
“`yaml
stages:
— build
— deployvariables:
ARTIFACTS_DIR: artifacts
PUBLIC_DIR: publicgenerate_ascii_art:
stage: build
script:
— mkdir -p $ARTIFACTS_DIR
— ./scripts/generate_ascii_art.sh src/your_code.py > $ARTIFACTS_DIR/ascii_art.txt
artifacts:
paths:
— $ARTIFACTS_DIR/ascii_art.txtdeploy_to_website:
stage: deploy
only:
— master
script:
— cp $ARTIFACTS_DIR/ascii_art.txt $PUBLIC_DIR/index.html
artifacts:
paths:
— $PUBLIC_DIR
“`This pipeline has two stages: `build` and `deploy`.
3. Create the `generate_ascii_art.sh` script in the `scripts/` directory:
“`bash
#!/bin/bashcode_file=$1
python3 -c “
from art import *
art = text2art(open(‘$code_file’).read(), ‘block’)
print(art)
“ > ascii_art.txt
“`This script uses the `art` library in Python to generate ASCII art from your code and saves it in `ascii_art.txt`.
4. Make sure you have the `art` library installed in your Python environment. You can install it with:
“`bash
pip install art
“`5. Create a simple HTML template in `public/index.html` to display the generated ASCII art:
“`html
<!DOCTYPE html>
<html>
<head>
<title>Generated ASCII Art</title>
</head>
<body>
<pre>
<! — Include ASCII art here →
$(cat ascii_art.txt)
</pre>
</body>
</html>
“`Now, when you push code changes to the `master` branch, GitLab CI/CD will automatically generate ASCII art from your code and deploy it to a static website. You can access the generated ASCII art by opening the website hosted in the `public` directory of your repository.
This is just one example of a creative GitLab CI/CD pipeline. You can adapt and expand upon it based on your project’s needs and the creative ideas you want to implement.
OpenAI. (2023). ChatGPT (September 25 Version) [Large language model]. https://chat.openai.com
The Commentary
This did not work out of the box — as is the case with most LLM work. But it 95% worked. And I fixed it mostly by asking follow up questions not editing the code myself. What is most impressive is that the LLM created a series of pieces, the pipeline code, the directory structure, and code run by the pipeline.
The ascii art does not generate. I suspect it has something to do with my python program that generates a Fibonacci sequence.But I am not going to debug that. Instead I am really amazed what did happen and then maybe in the future I’ll go back and revisit these early experiments.
This project and pipeline exist in gitlab.
Leave a Reply