https://github.com/Data4DM/BayesSD/discussions/234 and https://github.com/Data4DM/BayesSD/discussions/224
1.1 document ingestion
1.2 data preprocessing
2.1 document embedding
2.2 metadata embedding
3.1 query processing
3.2 embedding matching
4.1 state retriever
4.2 path evaluation
4.3 bottleneck detector
5.1 continuous learning
5.2 model update