The sorta-descriptive script takes a text file of image descriptions generated by MidJourney and sorts them into nine columns in a CSV file to prepare them for further data analysis.
The MidJourney description prompt generator helps users understand the way MidJourney perceives and describes images. This is particularly useful for those seeking to create consistent characters in MidJourney; as it allows them to analyze the themes and patterns that MidJourney identifies in similar images. By harnessing a deeper understanding of the AI's output, users can then refine their prompts to achieve their desired consistency.
This repository builds upon that knowledge to help users prepare their data for qualitative analysis of themes. It is assumed that users of this repo are copying and pasting the MidJourney descriptions into a .txt file for further analysis.
The sorta-descriptive repo contains a Python script designed to streamline the process of preparing data for analysis in cost-effective software alternatives to NVIVO and MAXQDA. These alternatives often require data to be submitted in a .csv or other spreadsheet format as opposed to a .txt file.
The sorta-descriptive script takes a text file of image descriptions generated by MidJourney and sorts it into nine columns in a CSV file:
- Subject
- Style
- Core
- Punk
- Wave
- Resolution
- Real
- Quality
- Additional
The resulting CSV file can then be used on its own or uploaded to various data analysis websites or software for further analytical insights.
This tutorial is designed with beginners in mind, offering a user-friendly solution.
- Sign up for a free account on Replit.com.
- Click on "+ Create Repl".
- Select Python as the programming language.
- Copy and paste the code from this repo into your
main.pyfile. - Name your MidJourney descriptions text file "descriptions.txt" and upload it to the repl.
- Click "Run" at the top of the page.
- After the script is running, you should see ten new files added to the "Files" section on the left-hand side of the Repl. These include nine separated files for subject, style, core, punk, wave, resolution, real, quality, and additional; and one combined file called
combined.csv. - Right-click on any of the output files to download them.
If you find this repository useful and would like to show your appreciation, consider buying me a coffee through Buy Me a Coffee.
Thank you for visiting my GitHub page!