Yingda Cheng

Yingda Cheng

Professor at Virginia Tech

CLAUDE.md

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project

Personal academic homepage for Yingda Cheng (Professor at Virginia Tech), built with Jekyll on the academicpages template (a fork of Minimal Mistakes). This repo is yingdacheng.github.io — GitHub Pages builds and deploys it automatically on every push to master. There is no CI, no separate build/deploy step, no test suite, and no linter/formatter configured.

Local development

bundle install                  # install Ruby deps (delete Gemfile.lock first if it errors)
bundle exec jekyll liveserve    # serve at localhost:4000 with live reload, using _config.dev.yml overrides

Ruby-dev, bundler, and nodejs must be installed on the system first. Gemfile.lock is intentionally absent from version control (delete it if it reappears and causes vulnerability warnings).

The npm run build:js / uglify scripts in package.json minify assets/js/_main.js (plus vendor plugins) into assets/js/main.min.js. Run this after editing assets/js/_main.js — the minified file is what’s actually served.

Content structure

Standard Jekyll collections, each with its own directory: _pages, _posts, _talks, _teaching, _publications, _portfolio. Site-wide settings (including the author sidebar with social links) live in _config.yml; _config.dev.yml overrides a few values for local serving.

Some collection entries (_talks, _teaching, _publications, _portfolio) are still unmodified template placeholders (e.g. “Talk 1 on Relevant Topic in Your Field”). This is expected and intentional — don’t flag or try to fill these in unless the user specifically asks.

Generating publication/talk markdown

markdown_generator/ contains actively-used tooling that converts structured data into the markdown files these collections expect. Run scripts from inside markdown_generator/ — they write output to ../_publications/ or ../_talks/ using relative paths.

  • pubsFromBib.py — reads proceedings.bib and pubs.bib (BibTeX, via pybtex), writes one .md per entry to ../_publications/. Configure via the publist dict at the top of the file.
  • publications.py — reads publications.tsv (columns: pub_date, title, venue, excerpt, citation, site_url, paper_url; pub_date as YYYY-MM-DD), writes to ../_publications/.
  • talks.py — reads talks.tsv (columns: title, type, url_slug, venue, date, location, talk_url, description; date as YYYY-MM-DD), writes to ../_talks/.

Each .py file has a corresponding .ipynb with more documentation; they do the same thing.

talkmap.py (project root) scrapes the location field from every file in _talks/, geocodes it via geopy/Nominatim, and regenerates the cluster map in talkmap/. Run it from inside _talks/. Requires glob, getorg, geopy (not declared in any manifest — install manually if missing).

Git workflow

Commit directly to master — no branching or PR process for this repo.