Group nucleotide databases based on admin-supplied keywords

[yannickwurm]

Some hurdles may be met if there are many databases. This is already becoming clear on fourmidable.

One idea could be to add a "Select all" button.

A more elegant approach could be to optionally group databases (based on keywords provided by the admin in config.yaml). For example, it would make sense if all databses with "genome" in the title were grouped together; and similarly that all databases with "transcript" or "cDNA" were grouped together. (under subheadings within the nucleotide section). Each subheading should have its own "select all" button.

[yeban]

Makes sense; +1.

[yannickwurm]

Maybe its easiest (but flexible) if the config.yaml file has a field such as: groupings: /(transcript)|(cDNA)/, /genome/

[yeban]

How about grouping database settings in config.yml, something like:

database:
  path: ~/db_path
  group_by:  /(transcript)|(cDNA)/, /genome/ #this can be a string or a regexp

[yannickwurm]

cool :)

grouping has to be optional though

On 16 Jun 2011, at 14:36, yeban wrote:

How about grouping database settings in config.yml, something like:

database:
 path: ~/db_path
 group_by:  /(transcript)|(cDNA)/, /genome/ #this can be a string or a regexp

Reply to this email directly or view it on GitHub: https://github.com/yannickwurm/sequenceserver/issues/29#issuecomment-1379307

[cmdcolin]

Are there any current ideas about adding some groupings like this?

[yeban]

Apart from the ideas above, there's this other one I had in mind and have experimented with a bit. If files under database dir are organised as:

|- genome/                    # => group / category
|  |- species_name_1/
|  |  ...
|  |- species_name_2/
|  |  ...
|  |- species_name_3/
|  |  ...
|- proteins/
|  |- species_name_1/
|  |  ...
`- transcripts/
   |- species_name_1/
   |  ...

the information from directory structure can be used to present the databases like:

People hosting a large number of databases could potentially benefit from this. But have to do it without getting in the way of people hosting just a handful of dbs. The grouping would be really flexible this way, allowing grouping by taxonomy, source, function, etc.

I have implemented a simple version of this in a separate branch, but it's gonna take some time to release even if others agree with the idea.

/cc @wwood

[cmdcolin]

Looks pretty cool! It isn't really urgent but I like this concept. Would definitely try it out.

But have to do it without getting in the way of people hosting just a handful of dbs

true.