Personal Branding Poster Generator

한국어 README

A Python/FastAPI project and Codex skill bundle that generates a personal branding poster from one portrait photo and one PDF resume.

Features

• Validates portrait image uploads (jpg, jpeg, png)
• Extracts and structures text from PDF resumes
• Generates rule-based summaries and hashtags
• Removes image backgrounds with rembg and falls back to OpenCV GrabCut
• Composes a monochrome portrait with neon silhouette layers
• Saves square poster output as PNG
• Saves generated metadata as JSON
• Supports FastAPI, CLI, and Codex skill execution

Project Structure

app/
  main.py
  models/
  routes/
  services/
agents/
scripts/
sample_inputs/
SKILL.md
tests/
requirements.txt
README.md
README.ko.md

The output/ directory is created automatically during execution.

Installation

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

CLI Usage

python -m app.main --image /path/to/photo.jpg --resume /path/to/resume.pdf --output-dir output

You can also use the repository-level runner script, which prepares the virtual environment and dependencies automatically.

python scripts/run_branding_poster.py --image /path/to/photo.jpg --resume /path/to/resume.pdf --output-dir output

For the built-in sample demo:

python scripts/run_branding_poster.py --sample --output-dir output

Generated files:

• output/poster.png
• output/metadata.json

API Usage

uvicorn app.main:app --reload

Example request:

curl -X POST "http://127.0.0.1:8000/poster/generate" \
  -F "photo=@/path/to/photo.jpg" \
  -F "resume_pdf=@/path/to/resume.pdf"

Tests

pytest

Using It As a Codex Skill

This repository doubles as a Codex skill bundle. After publishing it to GitHub, Codex can install or open the repository and use SKILL.md plus agents/openai.yaml to run it as an execution-focused skill.

Example flow:

$personal-branding-poster Create a personal branding poster from this portrait photo and resume PDF.

Sample flow:

$personal-branding-poster Show me the sample result using the repository demo inputs.

Codex can run scripts/run_branding_poster.py to:

• prepare the local environment
• generate the poster
• return poster.png and metadata.json

Example

Sample input files:

• Image: sample_inputs/lenna_test_image.png
• Resume: sample_inputs/lenna_resume_sample.pdf
• Output: sample_output/lenna_poster_sample.png

input (+ resume.pdf)	output

Notes

• Summaries and hashtags are rule-based so the pipeline can be upgraded to an LLM-backed version later.
• If rembg fails or is unavailable, the code falls back to OpenCV GrabCut.
• The text overlay is auto-sized and wrapped to stay within the poster bounds.