Check metadata columns from environmental packages

[josenavas]

Extend the metadata templates to enforce/suggest the columns outlined in the MIMARKS standard for each environmental package.

This also means that when the environmental package of a study changes, we need to perform all the required changes on the DB storage regarding metadata plus do not allow to completely change it until the user provides the new columns.

[rob-knight]

Note that environmental package should not be at the study level because many studies now contain samples from multiple types of environment (eg soil plus water or human plus dog plus household surface).

On Nov 19, 2014, at 1:34 AM, josenavas <notifications@github.commailto:notifications@github.com> wrote:

Extend the metadata templates to enforce/suggest the columns outlined in the MIMARKS standard for each environmental package.

This also means that when the environmental package of a study changes, we need to perform all the required changes on the DB storage regarding metadata plus do not allow to completely change it until the user provides the new columns.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/641.

[josenavas]

Thanks @rob-knight

Our current approach is to have it at study level, but the user can choose as many environmental packages as they apply to user's data. Is that ok?

[rob-knight]

As long as they don't have to fill all the fields for all the samples which is often impossible...

On Nov 19, 2014, at 9:54 AM, josenavas <notifications@github.commailto:notifications@github.com> wrote:

Thanks @rob-knighthttps://github.com/rob-knight

Our current approach is to have it at study level, but the user can choose as many environmental packages as they apply to user's data. Is that ok?

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/641#issuecomment-63651239.

[josenavas]

Thanks @rob-knight

@biocore/qiita-dev we should briefly discuss this today, to double-check that our schema fits our needs.

[josenavas]

The @biocore/qiita-dev group discussed this in today's meeting and we agreed on leave it at the study level and do not enforce them for the samples.
The 👎 of this solution is that we can potentially have sparse tables on the DB
The 👍 of this solution is that we don't force the user to choose the env package per sample (which can be horrible in a study with a lot of samples); the development burden is highly reduced and at least all the column names will be consistent across all the studies.