Without seeing the dome end cap of the element, I'm not 100% convinced that the bypass is installed upside down. If it used the Champ Labs, Fram poppet bypass design with compression spring combined, that might be true.
Otoh, it could still function if the top of the poppet opens inwards and seals back to the dome endcap when it closes. That would be similar to how another Purolator bypass (enclosed interior coil spring with a phenolic seal top) works but didn't use the composite/nylon poppet shown. Then it would still have a separate stamped compression spring. 'To me' seems a more reasonable explanation.
Again though, as I've never seen any Purolator filter previously with the poppet type bypass shown, can't rule anything out. Why I asked for more information on application number, date code and purchase location.