# omarabedelkader.github.io

This repository builds and publishes a personal academic website and CVs.

## What is happening here?

The project uses **Markdown sources** plus a small **Python + Pandoc build pipeline**:

- `sources/` contains the content (`site.md`, `cv.md`, `cv-fr.md`, and blog posts under `sources/blog/`).
- `build/` contains Python scripts that convert Markdown into:
  - `docs/index.html` and `docs/fr/index.html` (website pages for GitHub Pages)
  - `docs/blog/`, `docs/fr/blog/`, and RSS feeds generated from Markdown blog posts
  - `cv/cv-en.pdf` and `cv/cv-industry.pdf` (PDF CV outputs)
- `templates/` contains the website static assets (`style.css`, `main.js`, `header.html`, favicon).
- `resources/` contains citation and LaTeX support files (`publications.bib`, `apa.csl`, `header.tex`).
- `runner/` contains shell wrappers that set up a virtual environment, install dependencies, and run each build script.

## Build flow

`run.sh` creates/activates `.venv`, installs `requirements.txt`, then runs six deterministic tasks:

1. Translate `sources/cv.md` to `sources/cv-fr.md`
2. Build `cv/cv-en.pdf` and `cv/cv-fr.pdf`
3. Translate `sources/site.md` to `sources/site-fr.md`
4. Translate published blog posts from `sources/blog/*.md` to `sources/blog-fr/*.md`
5. Generate website news snippets
6. Build `docs/index.html`, `docs/fr/index.html`, blog pages, RSS feeds, and `docs/sitemap.xml`

The translator protects Markdown links, URLs, inline code, fenced code blocks, frontmatter, raw LaTeX commands, and the `{{PUBLICATIONS_FROM_BIB}}` token. The website build copies CSS/JS/favicon into `docs/`, converts the site Markdown to HTML, copies CV PDFs into `docs/cv/`, injects a footer with a "Last updated" label, and includes the client-side script.

## How to run

```bash
# run all build steps
bash run.sh
```

`run.sh` is quiet by default: it prints each task name and the final result, while successful command output is hidden. For troubleshooting, run `VERBOSE=1 bash run.sh`.

## Blog posts

Write English posts as Markdown files in `sources/blog/`. Files whose name starts with `_` are ignored, so `sources/blog/_template.md` is safe to keep as a template. Published posts are listed on `/blog/` and `/fr/blog/` as a full archive with grid/list views, search, year filters, and topic filters.

A post can be as small as:

```markdown
---
title: "My Post Title"
date: "2026-08-20"
description: "A short summary for indexes and SEO."
tags:
  - AI
draft: false
---

Post content in Markdown.
```

The opening `---` must start at the first column of the file. This is intentional: malformed front matter should fail the build instead of publishing broken instructions as a public article.

Then run `bash run.sh`. The build translates the post to French, writes the generated translation under `sources/blog-fr/`, publishes English and French HTML pages under `docs/blog/` and `docs/fr/blog/`, updates RSS feeds, and adds sitemap entries. Each post ends with an issue link for reporting problems.

Or run individual build targets:

```bash
bash runner/build-site.sh
bash runner/build-cv.sh
bash runner/build-industry.sh
```

## Deployment model

`docs/` is the publishable site directory for **GitHub Pages**.

So the normal edit cycle is:
1. Update content in `sources/` (and optionally templates/resources)
2. Rebuild
3. Commit generated output (`docs/`, `cv/`) and source changes
4. Push to GitHub

---

If you just want to update text/content, start with `sources/site.md` and rebuild.