feat: per-annotator document assignment

- Add annotator_assignments.json config (docs list or docs_range per user)
- Documents API filters by ?user= query param
- Client passes username from OAuth cookie to documents fetch
- rafmacalaba gets docs 1-10, rafamacalaba gets docs 10-100
- Falls back to MAX_DOCS_TO_SCAN if no assignment

annotator_assignments.json

@@ -0,0 +1,22 @@
+{
+    "rafmacalaba": {
+        "docs": [
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10
+        ]
+    },
+    "rafamacalaba": {
+        "docs_range": [
+            10,
+            100
+        ]
+    }
+}

app/api/documents/route.js

@@ -1,7 +1,71 @@
 import { HF_DATASET_BASE_URL, MAX_DOCS_TO_SCAN } from '../../../utils/config.js';
 
-export async function GET() {
+const isHFSpace = () => process.env.HF_TOKEN && process.env.NODE_ENV !== 'development';
+
+/**
+ * Load annotator assignments config.
+ * Supports both local file and HF fetch.
+ */
+async function loadAssignments() {
     try {
+        if (isHFSpace()) {
+            const url = `${HF_DATASET_BASE_URL}/raw/main/annotation_data/annotator_assignments.json`;
+            const res = await fetch(url, {
+                headers: { 'Authorization': `Bearer ${process.env.HF_TOKEN}` },
+            });
+            if (res.ok) return await res.json();
+        } else {
+            const fs = await import('fs');
+            const path = await import('path');
+            const filePath = path.default.join(process.cwd(), 'annotator_assignments.json');
+            if (fs.default.existsSync(filePath)) {
+                return JSON.parse(fs.default.readFileSync(filePath, 'utf-8'));
+            }
+        }
+    } catch (e) {
+        console.warn('Could not load annotator assignments:', e.message);
+    }
+    return null;
+}
+
+/**
+ * Get the set of allowed doc indices for a user.
+ * Returns null if no assignments (= show all).
+ */
+function getAllowedDocs(assignments, username) {
+    if (!assignments || !username) return null;
+
+    const userConfig = assignments[username] || assignments[username.toLowerCase()];
+    if (!userConfig) return null;
+
+    const allowed = new Set();
+
+    // Explicit list: "docs": [1, 2, 3]
+    if (userConfig.docs) {
+        userConfig.docs.forEach(d => allowed.add(d));
+    }
+
+    // Range: "docs_range": [10, 100]  (inclusive)
+    if (userConfig.docs_range) {
+        const [start, end] = userConfig.docs_range;
+        for (let i = start; i <= end; i++) {
+            allowed.add(i);
+        }
+    }
+
+    return allowed.size > 0 ? allowed : null;
+}
+
+export async function GET(request) {
+    try {
+        // Get username from query param
+        const { searchParams } = new URL(request.url);
+        const username = searchParams.get('user');
+
+        // Load assignments
+        const assignments = await loadAssignments();
+        const allowedDocs = getAllowedDocs(assignments, username);
+
         // Fetch the index file from HF Datasets
         const linksUrl = `${HF_DATASET_BASE_URL}/raw/main/annotation_data/wbg_data/wbg_pdf_links.json`;
         const linksRes = await fetch(linksUrl, {
@@ -21,10 +85,18 @@ export async function GET() {
 
         const links = await linksRes.json();
 
-        // Filter to successful links and take the first N
-        const successLinks = links.filter(l => l.status === 'success').slice(0, MAX_DOCS_TO_SCAN);
+        // Filter to successful links
+        let successLinks = links.filter(l => l.status === 'success');
+
+        // If user has assignments, filter to allowed docs only
+        if (allowedDocs) {
+            successLinks = successLinks.filter(l => allowedDocs.has(l.index));
+        } else {
+            // No assignments — take first N
+            successLinks = successLinks.slice(0, MAX_DOCS_TO_SCAN);
+        }
 
-        // Parallel fetch — much faster than sequential scanning
+        // Parallel fetch
         const results = await Promise.allSettled(
             successLinks.map(async (link) => {
                 const docUrl = `${HF_DATASET_BASE_URL}/raw/main/annotation_data/wbg_extractions/doc_${link.index}/raw/doc_${link.index}_direct_judged.jsonl`;
@@ -58,7 +130,7 @@ export async function GET() {
             status: 200,
             headers: {
                 'Content-Type': 'application/json',
-                'Cache-Control': 'public, s-maxage=3600, stale-while-revalidate=59'
+                'Cache-Control': 'no-store'
             }
         });
     } catch (error) {

app/page.js

@@ -43,26 +43,10 @@ export default function Home() {
     const annotatablePages = currentDoc?.annotatable_pages ?? [];
     const currentPageNumber = annotatablePages[pageIdx] ?? null;
 
-    // Load documents on mount
-    useEffect(() => {
-        fetch('/api/documents')
-            .then(res => res.json())
-            .then(data => {
-                setDocuments(data);
-                if (data.length > 0) {
-                    setSelectedDocIndex(data[0].index);
-                    setPageIdx(0);
-                }
-                setLoading(false);
-            })
-            .catch(err => {
-                console.error("Failed to load documents", err);
-                setLoading(false);
-            });
-    }, []);
-
-    // Read HF OAuth cookie for annotator identity
+    // Read HF OAuth cookie and load assigned documents
     useEffect(() => {
+        // 1. Read username from cookie
+        let username = '';
         try {
             const cookie = document.cookie
                 .split('; ')
@@ -70,12 +54,33 @@ export default function Home() {
             if (cookie) {
                 const user = JSON.parse(decodeURIComponent(cookie.split('=').slice(1).join('=')));
                 if (user.username) {
+                    username = user.username;
                     setAnnotatorName(user.username);
                 }
             }
         } catch (e) {
             console.warn('Could not read hf_user cookie', e);
         }
+
+        // 2. Fetch documents (filtered by user if logged in)
+        const url = username
+            ? `/api/documents?user=${encodeURIComponent(username)}`
+            : '/api/documents';
+
+        fetch(url)
+            .then(res => res.json())
+            .then(data => {
+                setDocuments(data);
+                if (data.length > 0) {
+                    setSelectedDocIndex(data[0].index);
+                    setPageIdx(0);
+                }
+                setLoading(false);
+            })
+            .catch(err => {
+                console.error("Failed to load documents", err);
+                setLoading(false);
+            });
     }, []);
 
     // Update currentDoc when selection changes