Allow uploading of samples >1 year old

[martinmchugh]

We are looking out for possible re-infections and may look to sequence old samples stored in the freezer that we didn't have capacity for at the time of collection. These could well be older than 1y but currently these cannot be uploaded to Majora.

Can this be revisited to remove the block, or an additional metadata flag added to indicate these are definately old samples and not a data entry error?

[SamStudio8]

Hi Martin! We did have a proposal to remove the date check entirely which was rejected by the metadata working group, it's deemed too useful. I think a flag would be the best way forward and I've previously discussed this with Anthony too. We're currently working on scaling solutions and will likely not be able to do this particularly soon. A rather unpleasant hack to get these samples in would be to push them with a valid date (ie. not 365 days old), then push the correct date with Ocarina put biosample --partial, or the uploader - as we recently added a patch to skip the past date check on samples that already exist. I wouldn't want to be seen as recommending this solution but if the samples are urgent then that would be the way to do it without waiting for us...!

…

On Wed, 9 Jun 2021, 13:46 martinmchugh, ***@***.***> wrote: We are looking out for possible re-infections and may look to sequence old samples stored in the freezer that we didn't have capacity for at the time of collection. These could well be older than 1y but currently these cannot be uploaded to Majora. Can this be revisited to remove the block, or an additional metadata flag added to indicate these are definately old samples and not a data entry error? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#84>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAIN6OU2MDSKXP6Z4YNU5VLTR5PC7ANCNFSM46L7BOOA> .

[rambaut]

Would it work to put an initial received_date in to create the entry and then add the collection_date and fix the received_date? Might make data integrity a little bit more robust.

[SamStudio8]

That's much nicer. I think that should work fine :)

…

On Wed, 9 Jun 2021, 14:20 Andrew Rambaut, ***@***.***> wrote: Would it work to put an initial received_date in to create the entry and then add the collection_date and fix the received_date? Might make data integrity a little bit more robust. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#84 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAIN6ORHTAVCI4UTTNAGILDTR5S77ANCNFSM46L7BOOA> .

[SamStudio8]

Just to be safe, I'd make sure the dates are correct before uploading any sequences in case they are gobbled up by a pipeline before you get a chance to fix them!

…

On Wed, 9 Jun 2021, 14:23 Sam Nicholls, ***@***.***> wrote: That's much nicer. I think that should work fine :) On Wed, 9 Jun 2021, 14:20 Andrew Rambaut, ***@***.***> wrote: > Would it work to put an initial received_date in to create the entry and > then add the collection_date and fix the received_date? Might make data > integrity a little bit more robust. > > — > You are receiving this because you commented. > Reply to this email directly, view it on GitHub > <#84 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAIN6ORHTAVCI4UTTNAGILDTR5S77ANCNFSM46L7BOOA> > . >

[martinmchugh]

That sounds good, I agree that losing the data check entirely would probably end up letting through many errors in response to reduced effort in a handful of genuine cases. Happy to do the workaround Andrew suggests.

[SamStudio8]

Moving this to our backlog (#62) as there is a viable workaround for the moment