rm frontmatter.yml after consolidating with citation.cff

pnpm i -D js-yaml
fix link to landing page overlapping with nav
add image to Wrenformer model card

citation.cff

@@ -1,33 +1,38 @@
-cff-version: 1.2.0
-title: Matbench Discovery
+# cff-version: 1.2.0
 message: If you use this software, please cite it as below.
+title: Matbench Discovery
+subtitle: Can machine learning identify stable crystals?
 authors:
-  - family-names: Riebesell
-    given-names: Janosh
+  - given-names: Janosh
+    family-names: Riebesell
     affiliation: University of Cambridge, Lawrence Berkeley National Laboratory
     email: [email protected]
     orcid: https://orcid.org/0000-0001-5233-3462
     corresponding: true
-  - family-names: Goodall
-    given-names: Rhys
+    affil_key: 1, 2
+  - given-names: Rhys
+    family-names: Goodall
     affiliation: University of Cambridge
     orcid: https://orcid.org/0000-0002-6589-1700
-  - family-names: Jain
-    given-name: Anubhav
+    affil_key: 1
+  - given-names: Anubhav
+    family-names: Jain
     orcid: 0000-0001-5893-9967
     affiliation: Lawrence Berkeley National Laboratory
-  - family-names: Persson
-    given-name: Kristin
+    affil_key: 2
+  - given-names: Kristin
+    family-names: Persson
     orcid: 0000-0003-2495-5509
     affiliation: Lawrence Berkeley National Laboratory
-  - family-names: King-Smith
-    given-name: Emma
-    orcid: 0000-0002-2999-0955
-    affiliation: University of Cambridge
-  - family-names: Lee
-    given-name: Alpha
+    affil_key: 2
+  - given-names: Alpha
+    family-names: Lee
     orcid: 0000-0002-9616-3108
     affiliation: University of Cambridge
+    affil_key: 1
+affiliations:
+  - Cavendish Laboratory, University of Cambridge, UK
+  - Lawrence Berkeley National Laboratory, Berkeley, USA
 license: MIT
 license-url: https://github.com/janosh/matbench-discovery/blob/main/license"
 repository-code: https://github.com/janosh/matbench-discovery

models/cgcnn/metadata.yml

@@ -25,8 +25,10 @@
     Ensemble Size: 10
 
   notes:
-    description: Published in 2017, CGCNN was the first crystal graph convolutional neural network to directly learn 8 different DFT-computed material properties from a graph representing the atoms and bonds in a crystal. ![Illustration of the crystal graph convolutional neural networks](https://researchgate.net/profile/Tian-Xie-11/publication/320726915/figure/fig1/AS:635258345119746@1528468800829/Illustration-of-the-crystal-graph-convolutional-neural-networks-a-Construction-of-the.png)
-    long: It showed that just like in other areas of ML, given large training sets, embeddings that outperform human-engineered features could be learned directly from the data.
+    description: |
+      Published in 2018, CGCNN was the first crystal graph convolutional neural network to directly learn 8 different DFT-computed material properties from a graph representing the atoms and bonds in a crystal.
+      ![Illustration of the crystal graph convolutional neural networks](https://researchgate.net/profile/Tian-Xie-11/publication/320726915/figure/fig1/AS:635258345119746@1528468800829/Illustration-of-the-crystal-graph-convolutional-neural-networks-a-Construction-of-the.png)
+    long: CGCNN was among the first to show that just like in other areas of ML, given large enough training sets, neural networks can learn embeddings that reliably outperform all human-engineered structure features directly from the data.
 
 - model_name: CGCNN+P
   model_version: 0.1.0 # the aviary version
@@ -60,5 +62,7 @@
     Perturbations: 5
 
   notes:
-    description: This work proposes simple structure perturbations to augment CGCNN's training data of relaxed structures with randomly perturbed ones resembling unrelaxed structures that are mapped to the same DFT final energy during training. ![Step function PES](https://media.springernature.com/full/springer-static/image/art%3A10.1038%2Fs41524-022-00891-8/MediaObjects/41524_2022_891_Fig1_HTML.png?as=webp)
+    description: |
+      This work proposes simple structure perturbations to augment CGCNN's training data of relaxed structures with randomly perturbed ones resembling unrelaxed structures that are mapped to the same DFT final energy during training.
+      ![Step function PES](https://media.springernature.com/full/springer-static/image/art%3A10.1038%2Fs41524-022-00891-8/MediaObjects/41524_2022_891_Fig1_HTML.png?as=webp)
     long: The model is essentially taught the potential energy surface (PES) is a step-function that maps each valley to its local minimum. The expectation is that during testing on unrelaxed structures, the model will predict the energy of the nearest basin in the PES. The authors confirm this by demonstrating a lowering of the energy error on unrelaxed structures.

models/wrenformer/metadata.yml

@@ -30,5 +30,9 @@ hyperparams:
   Ensemble Size: 10
 
 notes:
-  description: Wrenformer is a standard PyTorch Transformer Encoder trained to learn material embeddings from composition, space group, Wyckoff positions in a structure.
+  description: |
+    Wrenformer is a standard PyTorch Transformer Encoder trained to learn material embeddings from composition, space group, Wyckoff positions in a structure.
+    ![Model workings](https://science.org/cms/10.1126/sciadv.abn4117/asset/a29e0899-77d1-47c8-82e3-00ab87c3b8d5/assets/images/large/sciadv.abn4117-f1.jpg)
+    A ML–powered materials discovery workflow using Wrenformer's Wyckoff string inputs to predict formation energies for candidate materials in an enumerated library of Wyckoff representations (shapes are used to denote different Wyckoff positions and colors to denote different element types). Predicted formation energies are then compared against the known convex hull of stability. Structures satisfying the required symmetries are relaxed for materials predicted to be stable.
+
   long: It builds on [Roost](https://doi.org/10.1038/s41467-020-19964-7) and [Wren](https://doi.org/10.1126/sciadv.abn4117), by being a fast structure-free model that is still able to distinguish polymorphs through symmetry.

site/package.json

@@ -28,6 +28,7 @@
     "eslint": "^8.34.0",
     "eslint-plugin-svelte3": "^4.0.0",
     "hastscript": "^7.2.0",
+    "js-yaml": "^4.1.0",
     "katex": "^0.16.4",
     "mdsvex": "^0.10.6",
     "prettier": "^2.8.4",

site/src/app.d.ts

@@ -15,9 +15,9 @@ declare module '*metadata.yml' {
 }
 
 // paper metadata
-declare module '*frontmatter.yml' {
-  const frontmatter: import('$lib').Frontmatter
-  export = frontmatter
+declare module '*citation.cff' {
+  const data: import('$lib').Citation
+  export = data
 }
 
 declare module '*references.yaml' {

site/src/lib/index.ts

@@ -72,15 +72,20 @@ export type Reference = {
   ISSN?: string
 }
 
-export type Frontmatter = {
+export type Citation = {
   title: string
   subtitle?: string
-  tags: string[]
   authors: {
     name: string
     affiliation: string
+    affil_key: string
     orcid: string
   }[]
   affiliations: string[]
-  date: string
+  'date-released': string
+  license: string
+  'license-url': string
+  'repository-code': string
+  url: string
+  version: string
 }

site/src/routes/+layout.svelte

@@ -49,14 +49,21 @@
     max-width: 50em;
   }
   a[href='/'] {
-    font-size: 15pt;
-    position: absolute;
-    top: 2em;
-    left: 2em;
+    font-size: 14pt;
     background-color: rgba(255, 255, 255, 0.1);
     padding: 1pt 5pt;
     border-radius: 3pt;
     transition: 0.2s;
+    width: max-content;
+    box-sizing: border-box;
+    margin: 1em 0 0 1em;
+  }
+  @media (min-width: 900px) {
+    a[href='/'] {
+      position: absolute;
+      top: 1em;
+      left: 1em;
+    }
   }
   a[href='/']:hover {
     background-color: rgba(255, 255, 255, 0.2);

site/src/routes/paper/+layout.server.ts

@@ -1,11 +1,13 @@
 import fs from 'fs'
+import yml from 'js-yaml'
 import type { LayoutServerLoad } from './$types'
 
 export const load: LayoutServerLoad = async ({ route }) => {
   const data = fs.readFileSync(`src/routes/${route.id}/+page.md`, `utf8`)
+  const cff = fs.readFileSync(`../citation.cff`, `utf8`)
 
   // Count the number of words using a regular expression
   const word_count = data.match(/\b\w+\b/g)?.length ?? null
 
-  return { word_count }
+  return { word_count, ...yml.load(cff) }
 }

site/src/routes/paper/+layout.svelte

@@ -2,24 +2,25 @@
   import { References } from '$lib'
   import { pretty_num } from 'elementari/labels'
   import type { LayoutServerData } from './$types'
-  import { affiliations, authors, date, subtitle, title } from './frontmatter.yml'
   import { references } from './references.yaml'
 
   export let data: LayoutServerData
+
+  const authors = data.authors.map(
+    (auth) => `${auth[`given-names`]} ${auth[`family-names`]}<sup>${auth.affil_key}</sup>`
+  )
 </script>
 
-<h1>{title}<br /><small>{subtitle}</small></h1>
+<h1>{data.title}<br /><small>{data.subtitle}</small></h1>
 
 <address>
   <span>
-    {@html authors
-      .map((author) => `${author.name}<sup>${author.affiliation}</sup>`)
-      .join(`, `)}
+    {@html authors.join(`, `)}
   </span>
   <span>
-    {@html affiliations.map((affil, idx) => `${idx + 1}. ${affil}`).join(`<br/>`)}
+    {@html data.affiliations.map((affil, idx) => `${idx + 1}. ${affil}`).join(`<br/>`)}
   </span>
-  <span style="font-weight: lighter;">{date}</span>
+  <span style="font-weight: lighter;">{data[`date-released`]}</span>
 </address>
 
 <div>