Putting an oral corpus into a book changes it, and the scholarship on Ifa is unusually frank about this.
A recited verse has a performer, an occasion, an audience and a purpose. Transcribed, it loses all four. It also loses tone, since Yoruba is a tone language and most publication in English drops the marks; it loses the variation between one telling and the next, because a transcript freezes one telling; and it acquires a fixity the original never had. Bascom's 1969 volume prints Yoruba text with facing English translation precisely so that a reader can see some of what the translation is doing. Abimbola's work treats the ese as poetry with its own formal conventions, which recovers something a plain prose rendering loses.
None of that makes the record untrustworthy. It makes it a particular kind of evidence: good for showing that material of a certain shape existed, held by identifiable people, at a known date, and poor for settling what any figure means in general.
Ika Meji is documented as one of the sixteen principal figures and the atlas states that with confidence. How much verse stands under it, and which verse is authoritative, are questions with answers only inside a lineage. Those answers are conferred through initiation and are not reproduced here.