fix: reduce enrich batch size to 500 to avoid Neo4j memory pool limit

Adding prop_* properties to nodes (framework, http_method, auth_type, etc.)
increased per-node data 3-4x. With 140K nodes at batch size 2000, the
UNWIND transaction exceeded Neo4j's memory pool limit.

Reduced node and edge batch size from 2000 to 500. Matches the default
indexing batch size and avoids memory issues on large codebases.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: aksOps

src/main/java/io/github/randomcodespace/iq/cli/EnrichCommand.java

@@ -175,7 +175,8 @@ private int enrichFromCache(AnalysisCache cache, Path root, NumberFormat nf, Ins
             }
 
             // Bulk-load nodes in batches using UNWIND
-            int nodeBatchSize = 2000;
+            // Smaller batches to avoid Neo4j memory pool limit (nodes carry prop_* properties)
+            int nodeBatchSize = 500;
             int nodesLoaded = 0;
             int totalNodes = enrichedNodes.size();
             for (int i = 0; i < totalNodes; i += nodeBatchSize) {
@@ -263,7 +264,7 @@ private int enrichFromCache(AnalysisCache cache, Path root, NumberFormat nf, Ins
                 validEdgeMaps.add(props);
             }
 
-            int edgeBatchSize = 2000;
+            int edgeBatchSize = 500;
             int edgesLoaded = 0;
             int totalEdges = validEdgeMaps.size();
             CliOutput.info("  Loading " + nf.format(totalEdges) + " edges...");