The Uirapuru project addresses the challenge of querying incomplete or inconsistent graph data by developing a hybrid, constraint-aware querying system that performs prediction and validation only when needed during query evaluation. By selectively adding plausible edges based on likelihood thresholds, it enables queries to combine existing and predicted connections while maintaining constraint consistency. These principles will extend the Knowledge Mapping Layer of FLAMI, enhancing FAIR-compliant data repositories with richer, uncertainty-aware graph exploration and reasoning.