git clone https://github.com/ledgerW/gf-blogger.git

1. Install Anaconda

2. Create a virtual conda environment for this project:
   From project root: conda env create -f environment.yml
   Activate env: activate gf-blogger

Get API keys for the below services and put them in the .env (below)
OpenAI
Serper

Populate goldfinch_blogger/.env.template and rename to .env (keep in goldfinch_blogger/)

There are two examples:

• goldfinch_blogger/outline.txt
• goldfinch_blogger/section.txt

The SECTION: string is used for parsing sections, so it must be used. Just keep the format in the examples, but otherwise anything can go in the section description.

These are the input for the "blog writer". The file can be called anything. A relative path to it will be passed to the "blog writer" CLI from the terminal

Output will be found in goldfinch_blogger/new_post/ (it will be created). It will produce the final blog post, which will be called blog.txt, as well as files containing individual sections and the final prompts containing all the stuffed context for each section, which will located in goldfinch_blogger/new_post/sections/.

from ./goldfinch_blogger: python GoldfinchBlogger.py --outline_path outline.txt

Notes

• Here's an example of how all this can be deployed fully serverless with backend and frontend CI/CD using Seed and Vercel...
  Oxpecker
  (This is a private repo - I'll send you an invite)