Sync people data across website, CV, and lab-manual

.github/workflows/build-content.yml

@@ -26,6 +26,8 @@ jobs:
     steps:
       - name: Checkout repository
         uses: actions/checkout@v4
+        with:
+          submodules: true
 
       - name: Set up Python
         uses: actions/setup-python@v5

.github/workflows/build-cv.yml

@@ -28,6 +28,8 @@ jobs:
     steps:
       - name: Checkout repository
         uses: actions/checkout@v4
+        with:
+          submodules: true
 
       - name: Install TeX Live
         run: |

.gitmodules

@@ -0,0 +1,3 @@
+[submodule "lab-manual"]
+	path = lab-manual
+	url = https://github.com/ContextLab/lab-manual.git

AGENTS.md

@@ -20,6 +20,7 @@ contextlab.github.io/
 ├── scripts/            # Python build system (see scripts/AGENTS.md)
 ├── images/             # Assets: people/, publications/, software/, research/, news/
 ├── documents/          # CV files (JRM_CV.tex → .pdf, .html)
+├── lab-manual/         # Git submodule (ContextLab/lab-manual)
 └── tests/              # pytest suite for build system
 ```
 
@@ -29,7 +30,8 @@ contextlab.github.io/
 |------|----------|-------|
 | Add publication | `data/publications.xlsx` | Auto-builds via GitHub Actions |
 | Add team member | `scripts/onboard_member.py` | Processes photo, generates bio, updates CV |
-| Offboard member | `scripts/offboard_member.py` | Moves to alumni, updates CV |
+| Offboard member | `scripts/offboard_member.py` | Moves to alumni, updates CV + lab-manual |
+| Reconcile people | `scripts/reconcile_people.py` | Three-way sync: people.xlsx ↔ CV ↔ lab-manual |
 | Add software | `data/software.xlsx` | |
 | Add news | `data/news.xlsx` | Thumbnail in `images/news/` |
 | Update CV | `documents/JRM_CV.tex` | Auto-compiles to PDF+HTML |

CLAUDE.md

@@ -0,0 +1,104 @@
+# CLAUDE.md
+
+This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
+
+## Project Overview
+
+Static website for the Contextual Dynamics Lab at Dartmouth College. Hosted on GitHub Pages at [context-lab.com](https://context-lab.com). Content pages (publications, people, software, news) are **auto-generated** from Excel spreadsheets via a Python build system — never edit the root HTML for those pages directly.
+
+## Commands
+
+```bash
+# Install dependencies
+pip install -r requirements-build.txt
+
+# Validate spreadsheet data
+cd scripts && python validate_data.py
+
+# Build all content pages (publications, people, software, news)
+cd scripts && python build.py
+
+# Build CV (requires XeLaTeX + Dartmouth Ruzicka fonts)
+cd scripts && python build_cv.py
+
+# Run full test suite
+python -m pytest tests/ -v
+
+# Run a single test file
+python -m pytest tests/test_build_publications.py -v
+
+# Pre-push validation (validation + build + tests)
+cd scripts && python pre_push_check.py
+
+# Local dev server
+python3 -m http.server 8000
+
+# Add hand-drawn borders to images
+python scripts/add_borders.py image.png images/publications/
+python scripts/add_borders.py photo.jpg images/people/ --face
+
+# Onboard new member (from scripts/ dir)
+python onboard_member.py "First Last" --rank "grad student" --photo headshot --skip-llm
+
+# Offboard member to alumni (from scripts/ dir)
+python offboard_member.py "member name" --end-year 2025
+
+# Reconcile people across website, CV, and lab-manual
+cd scripts && python reconcile_people.py --dry-run  # report only
+cd scripts && python reconcile_people.py             # apply auto-fixes
+
+# Initialize lab-manual submodule (required for reconciliation)
+git submodule update --init
+```
+
+## Architecture
+
+### Content Pipeline
+
+```
+data/*.xlsx  →  scripts/build_*.py  →  root *.html (auto-generated)
+                    ↑
+            templates/*.html (markers like <!-- PUBLICATIONS_PAPERS -->)
+```
+
+Each `build_*.py` follows the same pattern:
+1. Load spreadsheet with `utils.load_spreadsheet_all_sheets()`
+2. Generate HTML fragments
+3. Inject into template via `utils.inject_content(template, output, {"MARKER": html})`
+
+### Key Directories
+
+| Directory | Purpose |
+|-|-|
+| `data/` | Source spreadsheets (publications.xlsx, people.xlsx, software.xlsx, news.xlsx) + Dartmouth fonts |
+| `templates/` | HTML templates with `<!-- MARKER -->` injection points |
+| `scripts/` | Python build system, validation, onboarding/offboarding tools |
+| `tests/` | pytest suite — conftest.py adds `scripts/` to sys.path |
+| `documents/` | CV source (JRM_CV.tex) and generated outputs (PDF, HTML) |
+| `css/style.css` | Single stylesheet with CSS variables at top |
+| `js/main.js` | All JS components (9 init functions) |
+| `lab-manual/` | Git submodule — [ContextLab/lab-manual](https://github.com/ContextLab/lab-manual) (init with `git submodule update --init`) |
+
+### CV Pipeline
+
+`documents/JRM_CV.tex` → `scripts/build_cv.py` + `scripts/extract_cv.py` (custom LaTeX→HTML parser) → `documents/JRM_CV.pdf` + `documents/JRM_CV.html`
+
+### GitHub Actions
+
+- **build-content.yml**: Triggers on changes to `data/`, `templates/`, `scripts/`. Validates, builds, runs tests, auto-commits regenerated HTML.
+- **build-cv.yml**: Triggers on changes to `documents/JRM_CV.tex`, CV scripts, or `css/cv.css`. Compiles LaTeX, runs tests, auto-commits PDF+HTML.
+
+## Critical Rules
+
+- **Never edit auto-generated root HTML** (`publications.html`, `people.html`, `software.html`, `news.html`) — edit `data/*.xlsx` or `templates/*.html` instead
+- **Never use `!important`** in CSS without explicit justification
+- **Never add inline styles** to templates — use CSS classes
+- **Always run tests before pushing**: `python -m pytest tests/ -v`
+- **All headings are lowercase** via `text-transform: lowercase` in CSS
+
+## Design System
+
+- **Colors**: `--primary-green: rgb(0, 112, 60)`, `--bg-green: rgba(0, 112, 60, 0.2)`, `--dark-text: rgba(0, 0, 0, 0.7)`
+- **Font**: Nunito Sans, 300 weight body, 14px base, 1.7 line-height
+- **Images**: 500x500px with hand-drawn green borders (use `scripts/add_borders.py`), face-detect cropping for people photos (`--face`)
+- **Forms**: Formspree backend — endpoint in form `action` attribute

README.md

@@ -223,6 +223,7 @@ The script will:
 - Generate or edit the bio using a local LLM (gpt-oss-20b)
 - Add the member to `people.xlsx`
 - Add the member to `JRM_CV.tex`
+- Update `lab_manual.tex` in the lab-manual submodule (if initialized)
 - Invite to GitHub organization and teams (if `--github` provided)
 - Share Google Calendars with appropriate permissions (if `--gmail` provided)
 - Rebuild `people.html`
@@ -268,10 +269,31 @@ python offboard_member.py --list-no-photo
 The script will:
 - Move the member from `members` sheet to `alumni_undergrads` in `people.xlsx`
 - Update `JRM_CV.tex` to add the end date
+- Update `lab_manual.tex` in the lab-manual submodule (if initialized)
 - Prompt to rebuild `people.html`
 
 **Idempotent**: Running twice with the same name detects the member is already offboarded.
 
+#### Reconciling People Data
+
+The reconciliation tool compares member/alumni data across `people.xlsx` (source of truth), `JRM_CV.tex`, and the lab-manual's `lab_manual.tex`:
+
+```bash
+cd scripts
+
+# Report discrepancies without making changes
+python reconcile_people.py --dry-run
+
+# Apply auto-fixes (people.xlsx entries missing from other sources)
+python reconcile_people.py
+```
+
+The tool uses fuzzy name matching to catch spelling variations and nicknames. Discrepancies are categorized as:
+- **Auto-resolved**: People in people.xlsx missing from CV or lab-manual (auto-added)
+- **Flagged for review**: People in other sources missing from people.xlsx (requires manual review)
+
+**Requires**: Lab-manual submodule initialized (`git submodule update --init`)
+
 #### Adding Alumni (Manual)
 
 1. Open `data/people.xlsx`
@@ -301,6 +323,9 @@ The script will:
 # Install dependencies
 pip install -r requirements-build.txt
 
+# Initialize lab-manual submodule (required for reconciliation)
+git submodule update --init
+
 # Validate data files
 python scripts/validate_data.py