Philip Newton's journal

It would also be good to have a searchable corpus, where you can search for individual morphemes, not just words.

Well, the Nunavut Hansard is a bilingual corpus (English, Inuktitut), though it might not be the most exciting of material.

InuktitutComputing.com also made available an aligned version, where the Inuktitut and English sentences were matched one to the other, as far as possible - and you can search their corpus.

That only goes by strings, though, I think, so searching for (say) -tuq- would not find the -juq- allomorph, and vice versa.

Here's an example search (for "*&*", i.e. all words that have an ampersand in them, the common typewriter representation of the letter ł, pronounced [ɬ]).