In its defence, it's not so much odd as it is (arguably) over-enthusiastic. Pretty much every pollster is getting the same results, give or take, before accounting for don't knows, and then the differences are coming from how they're weighting them. Their pre-weighting result is 43-26, right in line with everyone else.
Traditionally those don't knows start to break one way or the other (historically more Tory) by the halfway point of a campaign, so if the polls do start to converge a bit in 2-3 weeks' time then we can say it was predictive and not necessarily an outlier.
I'm also usually sceptical of claims that pollsters are trying to rig elections through deceptive polling, but the "J" in JL Partners was Theresa May's private pollster when she was PM, and his business is selling polling services to Tories and US Republicans. They do have something of an interest in trying to make the race appear closer than it is.
