2 should be Arhopala sublustris.
Key line 1, 7, 8, 95,100, 120, 126, 127. The alea subgroup of the cleander group is particularly difficult.
However, of the 3 spp. with the hindwing postdiscal spot in space 6 shaped thus( broadest above; narrow below ; convex on the outer margin.) and overlaping the cellend bar, A. aroa can be discounted because of its forewing rounded termen. A. milleri is a deep indigo blue with the border a thread.
This one have the upper a bright blue (flash-wise) and the dark wing border wider than a thread.
4 is Allotinus unicolor. Readily IDed by the hindwing postdiscal spot in space 6 being well inside of the spot in space 7 above.
14 should be Jamides celeno. There are examples in which the bands are very broad and the striae collide.
21 is probably Potanthus confucius. The poor degree of overlap is not an indicator of its upperside feature.
22 is Notocrypta clavata. Band extends to costa; absence of spot in space 4; inner notch not right-angled.
26 should be Jamides malaccanus. There are 3 complete bands on the hindwing.
30 is Ypthima huebneri. Y. huebneri is separated from the Sri Lankan ceylonica which have a large white patch.
36 is Curetis sperthis. Black specklings & a forewing band which is almost parallel to the termen.