Remove hardcoded override in Publications; rely on pipeline + data

- Publications.jsx: drop paper-specific URL override
- PUBLICATION_PROCEDURE: point to central workflow, run sync script

Author: tserre

PUBLICATION_PROCEDURE.md

@@ -1,64 +1,48 @@
 # PROCEDURE FOR ADDING PUBLICATIONS
 
+**Lab website publications** come from the central repo. Do not edit `src/data/publications_by_year.json` by hand—it is generated by the sync script.
+
+**Full pipeline:** See the central repo doc  
+`/Users/tserre/Projects/prj_cv/CENTRAL_TO_LAB_AND_CV_WORKFLOW.md`
+
 ## CENTRAL SOURCE OF TRUTH
 **Location:** `/Users/tserre/Projects/prj_cv/data/publications_structured.json`
 
-This is the SINGLE file that controls all publications for both CV and lab website.
+Single source for both CV and lab website. Edit that file, then run the pipeline scripts.
 
-## STEP-BY-STEP PROCEDURE
+## PIPELINE (run after editing central file)
 
-### 1. Edit the Central File
-Edit `/Users/tserre/Projects/prj_cv/data/publications_structured.json`
+### 1. Edit Central File
+Edit `/Users/tserre/Projects/prj_cv/data/publications_structured.json`.  
+Use the **official published URL** (Cell, OpenReview, etc.) once a paper is published; arXiv only for preprints.
 
-**Structure:**
-```json
-{
-  "under_review": [
-    {
-      "title": "Paper Title",
-      "authors": "Author1, Author2 & T. Serre",
-      "journal": "Journal Name",
-      "url": "https://arxiv.org/abs/XXXX",
-      "year": "2025"
-    }
-  ],
-  "peer_reviewed": [...],
-  "unpublished_preprints": [...]
-}
+### 2. Sync to This Repo (Lab Website)
+From this repo:
+```bash
+cd /Users/tserre/Projects/Lab-Website
+python3 scripts/sync_from_central_publications.py
 ```
+This overwrites `src/data/publications_by_year.json` from the central file (and keeps existing `pdfPath` values).
 
-**Conventions:**
-- Use † for first equal authors, ‡ for last equal authors
-- Use sentence case capitalization (first word + after colons)
-- Include arXiv URLs when available
-- Place new papers at bottom of appropriate year section
-
-### 2. Regenerate CV
+### 3. Regenerate CV (in central repo)
 ```bash
 cd /Users/tserre/Projects/prj_cv/scripts
-python generate_cv.py
+python3 generate_cv.py
 ```
 
-### 3. Update Lab Website
-**Option A:** Edit HTML source in `/Users/tserre/Projects/Lab-Website/public/misc/publications`
-**Option B:** Run lab website update scripts (if available)
-
-### 4. Commit Changes
+### 4. Rebuild and Deploy
 ```bash
-cd /Users/tserre/Projects/prj_cv
-git add -A
-git commit -m "Add new publication: [Title]"
-git push origin main
+cd /Users/tserre/Projects/Lab-Website
+npm run build
+# then your normal deploy (e.g. push to trigger GitHub Actions)
 ```
 
 ## CRITICAL RULES
-- ✅ ALWAYS edit the central JSON file first
-- ✅ NEVER edit generated LaTeX files directly
-- ✅ Use consistent formatting conventions
-- ❌ Never edit `/Users/tserre/Projects/prj_cv/latex/serre_cv.tex` directly
-- ❌ Never edit `/Users/tserre/Projects/prj_cv/latex/serre_cv.pdf` directly
+- ✅ Update the central file first, then run `sync_from_central_publications.py`
+- ❌ Do not edit `src/data/publications_by_year.json` directly—it is produced by the sync script
+- The file `src/data/officialPublicationUrls.js` is a display override map; keep it in sync with central when you add official links so it mirrors the single source of truth
 
 ## VERIFICATION
-1. Check generated CV PDF looks correct
-2. Check lab website displays correctly
-3. Ensure both sources are identical
+1. Central file has the correct URL for published papers
+2. After sync, lab website publications page shows the correct link
+3. CV PDF matches central data

src/pages/Publications/Publications.jsx

@@ -12,21 +12,12 @@ export function Publications() {
     const [searchQuery, setSearchQuery] = useState("");
     const [selectedYear, setSelectedYear] = useState("All");
 
-    // Official journal URL overrides (used when mapping or publication.url could be stale/wrong)
-    const BETTER_AI_JOURNAL_URL = "https://www.cell.com/trends/cognitive-sciences/fulltext/S1364-6613(25)00349-3";
-
     // Function to get the official publication URL (journal/conference/OpenReview)
     const getOfficialUrl = (publication) => {
-        // Explicit overrides for papers that must always link to the journal (not arXiv)
-        if (publication.title && publication.title.includes("Better artificial intelligence does not mean better models of biology")) {
-            return BETTER_AI_JOURNAL_URL;
-        }
-        // Then try the mapping
         const mappedUrl = getOfficialPublicationUrl(publication.title);
         if (mappedUrl) {
             return mappedUrl;
         }
-        // If URL is not a PDF and not our internal publications page, use it as official URL
         if (publication.url && 
             !publication.url.endsWith('.pdf') && 
             !publication.url.includes('/papers/') &&