For the needs of machine-learning aided novel materials discovery, large databases containing various compounds and properties need to be constructed. One possible approach is to extract a large number of chemical structures from an existing database. Another approach relies on the generation of diverse positional isomers of a given parent structure by introducing different functional groups and varying their positions. The present work is focused on building a software for automatic generation of all non-redundant positional isomers of a parent structure to build or expand different datasets. An illustrative example is provided. The software package could be found at the repository https://github.com/carim2020/der-gen.