I've read it on the internet. It had interesting observations.
Here are a few notes for discussion.

It is said that 3 pieces (per side) was removed from earlier version of standart Shogi (their movement was not described). One of them is named elephant, and it's definitely the Drunk Elephant, all we know.
Two other pieces are leopards. I did not hear about them before. What are they? Could it be the Ferocious Leopard (猛豹, mōhyō) from larger Shogi variants?

Second, in the preface, added in later edition, it's said that Semedo is not name of the game, but the name of Portuguese missionary, who has described it. No further information about it. It's only said that Kisljuk was not able to find information about in any Chinese sources. Probably, Thomas Hyde's "Mandragorias, seu Historia Shahiludii" is the only available source, and exact rules and origins of this game will never be known. It could be played in a very small region of China.
There is some hope if somebody will find out, wich regions was visited by this missionary, visit these regions, and find this game if it's still played or described in some archieves (big chances that it's completely forgot). But these are probably too big efforts for finding game, wich is, most likely, a merely miniaturized version of Xiang-Qi, designed for casual playing.
Still, it have cultural significance, being, as far as i know, the only miniature chess variant, invented earlier than XIX century.
If there is description of normal Xiang-Qi in "Mandragorias", one can compare it with description of this misterious game, and it (unlikely) could give some clarity.
There may be another key - that it's played on squares rather than intersections. Like Ban-Qi - none-chess game, played with Xiang-Qi equipment and Jungle game with same hierarchy system (though Semedo's game recembles Xiang-Qi much more).

There is no description of the large Korean chess variant, expect mentoiding it's size and similarities with Dai-Shogi. But certainly there is some more information somewhere in sources, used by Kisljuk.
I hope, somebody knows something about it.

Mentoided recently patented large Xinag-Qi variant is, most probably, Wuji chess. It have different number of columns, but it could be mistake. This book have several mistakes (some are fixed preface of later edition). Some mistakes may be copied from sources (Chu Shogi Lion is described with same mistakes as in Murray's History of Chess).