IDSgrep is a general tree-matching utility whose main application is to querying databases of Han-script character structures. For instance, a user might ask for all characters with 月 on the right, and get a result including the characters 明, 朝, and 期. IDSgrep supports a powerful language for specifying structural queries, inspired by traditional Unix regular expression syntax; the structural data itself is represented in an extended version of Unicode Ideographic Description Sequences. The package includes, or can generate, dictionary data from Tsukurimashou, KanjiVG, CHISE, and EDICT2.
Read the IDSgrep user's manual (PDF).