Feat/knowledge graphs

knowledge-graphs/docker-compose.yml

@@ -6,7 +6,7 @@ services:
     environment:
       JAVA_OPTS: "-Darcadedb.server.rootPassword=arcadedb"
     healthcheck:
-      test: ["CMD-SHELL", "wget -qO- http://localhost:2480/api/v1/ready || exit 1"]
+      test: ["CMD", "curl", "-sf", "http://localhost:2480/api/v1/ready"]
       interval: 5s
       timeout: 3s
       retries: 20

knowledge-graphs/queries/queries.sh

@@ -76,6 +76,8 @@ LIMIT 10
 echo ""
 echo "=== Query 5: GraphRAG Hybrid (Vector Seed + Citation Expansion) ==="
 echo "Find topics reachable via citation graph from papers most similar to [0.8, 0.2, 0.1, 0.1]."
+echo "Note: paper IDs are hardcoded because ArcadeDB does not support a SELECT subquery"
+echo "inside a MATCH 'where' clause; these IDs are the top-3 vector-similarity results."
 echo ""
 query "sql" "
 SELECT topic.name AS topic, count(*) AS connections

knowledge-graphs/setup.sh

@@ -8,8 +8,10 @@ DB_NAME="KnowledgeGraph"
 
 # ── Wait for ArcadeDB ─────────────────────────────────────────────────────────
 echo "Waiting for ArcadeDB at ${ARCADEDB_URL}..."
+retries=0
 until curl -sf -u "${ARCADEDB_USER}:${ARCADEDB_PASS}" \
     "${ARCADEDB_URL}/api/v1/ready" > /dev/null 2>&1; do
+  (( retries++ >= 30 )) && { echo "ArcadeDB did not start after 60s"; exit 1; }
   sleep 2
 done
 echo "ArcadeDB is ready."

knowledge-graphs/sql/02-data.sql

@@ -73,13 +73,13 @@ CREATE EDGE COVERS FROM (SELECT FROM Paper WHERE id = 'p9') TO (SELECT FROM Topi
 CREATE EDGE COVERS FROM (SELECT FROM Paper WHERE id = 'p10') TO (SELECT FROM Topic WHERE id = 't6')
 CREATE EDGE COVERS FROM (SELECT FROM Paper WHERE id = 'p10') TO (SELECT FROM Topic WHERE id = 't2')
 -- PaperActivity documents for trending query (varied counts so results are ordered meaningfully)
-INSERT INTO PaperActivity SET paperId = 'p1', citationCount = 15, ts = date()
-INSERT INTO PaperActivity SET paperId = 'p2', citationCount = 22, ts = date()
-INSERT INTO PaperActivity SET paperId = 'p3', citationCount = 18, ts = date()
-INSERT INTO PaperActivity SET paperId = 'p8', citationCount = 30, ts = date()
-INSERT INTO PaperActivity SET paperId = 'p10', citationCount = 25, ts = date()
-INSERT INTO PaperActivity SET paperId = 'p1', citationCount = 8, ts = date()
-INSERT INTO PaperActivity SET paperId = 'p2', citationCount = 10, ts = date()
-INSERT INTO PaperActivity SET paperId = 'p9', citationCount = 12, ts = date()
-INSERT INTO PaperActivity SET paperId = 'p3', citationCount = 7, ts = date()
-INSERT INTO PaperActivity SET paperId = 'p5', citationCount = 5, ts = date()
+INSERT INTO PaperActivity SET paperId = 'p1', citationCount = 15, ts = '2024-01-15T10:00:00'
+INSERT INTO PaperActivity SET paperId = 'p2', citationCount = 22, ts = '2024-01-15T10:01:00'
+INSERT INTO PaperActivity SET paperId = 'p3', citationCount = 18, ts = '2024-01-15T10:02:00'
+INSERT INTO PaperActivity SET paperId = 'p8', citationCount = 30, ts = '2024-01-15T10:03:00'
+INSERT INTO PaperActivity SET paperId = 'p10', citationCount = 25, ts = '2024-01-15T10:04:00'
+INSERT INTO PaperActivity SET paperId = 'p1', citationCount = 8, ts = '2024-01-16T10:00:00'
+INSERT INTO PaperActivity SET paperId = 'p2', citationCount = 10, ts = '2024-01-16T10:01:00'
+INSERT INTO PaperActivity SET paperId = 'p9', citationCount = 12, ts = '2024-01-16T10:02:00'
+INSERT INTO PaperActivity SET paperId = 'p3', citationCount = 7, ts = '2024-01-16T10:03:00'
+INSERT INTO PaperActivity SET paperId = 'p5', citationCount = 5, ts = '2024-01-16T10:04:00'