Mass spectrometry news and literature from UniMZ. A compact bilingual static publication for GitHub Pages.
Run python scripts/build.py with Python 3.10 or newer, then python -m unittest discover -s tests. No packages are required. Preview with python -m http.server 8000. Generated HTML, CSS, JavaScript and RSS are committed, so GitHub Pages can serve the repository root directly. The check workflow verifies that generated pages match their source.
The existing GitHub Pages settings determine deployment. This project does not change those settings. The confirmed public URL is https://mznews.unimz.org, recorded in site.json; it controls canonical, RSS and sitemap links. CNAME preserves this custom domain. Relative navigation supports both a custom domain and the GitHub project path. Preserve any existing CNAME file.
Create one UTF-8 JSON file in content/posts/, using the launch post as a complete example. Every post has exactly these fields:
slug: stable lowercase URL slug; do not rename after publication.date:YYYY-MM-DD, using the Beijing publication date.category:News,Latest,Spotlight, orClassics.titleandsummary: objects withenandzhtext.body:enandzharrays of plain-text paragraphs.links: a list of{ "label": "...", "url": "https://..." }; local site paths may start with/.papers: an ordered array, empty for an announcement. Each paper hastitle(original title),authors(complete ordered array of names),journal,year(string),url(original HTTPS paper link), andcommentary(enandzheditorial text).
Optional paper fields are first_online (verified YYYY-MM-DD date), publication_status (en and zh text, for example identifying a preprint or journal article), read_scope (en and zh text identifying the evidence actually read, such as abstract only), and doi (a DOI identifier beginning with 10., without a resolver URL). Provide these when verified; omit unknown metadata rather than guessing. A publication year does not establish its first-online date. Metadata appears beside the bibliographic information in each language; field labels stay English.
Each Latest entry is one complete daily digest with zero, one or many papers, in the selected reading order. There is no fixed count, cap or filler requirement. A zero-paper digest must honestly explain the coverage and that no new qualifying papers were found in both language bodies; never invent a paper to fill an entry. Create at most one digest for a date. Spotlight and Classics each introduce one paper. Preserve original paper titles and author names in both languages; translate only editorial prose. Interface and navigation stay English. Use bibliographic information, original commentary and paper links. Do not reproduce papers or add evaluation metrics or private material.
For new Latest daily digests, derive titles from the Beijing publication date: English Literature Brief (Month D, YYYY) (full English month, unpadded day; preserve parentheses, comma and spaces) and Chinese YYYY年M月D日文献简报 (unpadded month/day). Do not add topic subtitles, paper counts or version labels. Preserve original scientific paper titles, stable slugs and historical article titles; this rule does not apply to News, Spotlight or Classics.
The build validates the explicit schema and safely escapes all text. Unknown fields fail validation. English and Chinese article pages have stable URLs under posts/YYYY-MM-DD-slug/ and posts/YYYY-MM-DD-slug/zh/. Home combines all categories in reverse date order. Empty sections remain empty until actual articles are added. RSS contains one entry per post. Search works in either language; without JavaScript all articles remain available.
After editing, rebuild, run the checks, and commit the source together with generated files. A pull request runs the same checks before merge. Deleting or changing a published slug requires separately removing its old generated directory and considering existing external links.