Build fails because of issue with data downloaded by `EpiMix`

[mtiberti]

Backtrace:

Quitting from Moonlight2R.Rmd:738-758 [unnamed-chunk-20]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
<error/rlang_error>
Error in `data.frame()`:
! arguments imply differing number of rows: 1758481, 0
---
Backtrace:
    ▆
 1. └─Moonlight2R::GMA(...)
 2.   └─EpiMix::EpiMix(...)
 3.     └─EpiMix:::filterProbes(...)
 4.       └─EpiMix:::getProbeAnnotation(...)
 5.         └─base::data.frame(probe = names(manifest), gene = manifest$gene)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Error: processing vignette 'Moonlight2R.Rmd' failed with diagnostics:
arguments imply differing number of rows: 1758481, 0
--- failed re-building ‘Moonlight2R.Rmd’

I've seen this happening intermittently on different builders, and looks like it might be connected with corrupted data from ExperimentHub. In fact, the problem is generated by these lines:

        manifest <- EpiMix_getInfiniumAnnotation(plat = met.platform, genome = genome)
        ProbeAnnotation <- data.frame(probe = names(manifest), gene = manifest$gene)

where EpiMix_getInfiniumAnnotation only downloads from ExperimentHub

[mtiberti]

BioC suspects its a builder problem - we should check if it's fixed at the next building cycle

[mtiberti]

looks like we are all green now - builders managed to do their downloads

[mtiberti]

I'm reopening this - similar fails keep happening and feels like we would need a better long-term solution