File Relationships

[spficklin]

We are missing the ability to have relationships between files such that one could supercede the other or files could be grouped

[dsenalik]

I have wondered about this for files that naturally come in pairs, such as fastq forward + reverse reads, or vcf + index file.
With minimal changes we could add a "File Name" column to the download field.
For example

[dsenalik]

I really want this, but others may not, so I implemented this additional column as an optional setting, default is off.
That required building the whole administration page interface, so while I was in there I threw in a help page too!
See pull request #20

[dsenalik]

My Nov. 4 commit to pull request #20 resolves this by adding a way to enter a display_name value, which is the name displayed to the user. Comments welcomed!

[dsenalik]

To-do notes for a possible problem to fix. Under php 8 this becomes worse, but the new filename column in the chado.fileloc table seems to be missing its schema

Error observed is Warning: array_key_exists() expects parameter 2 to be array, null given in chado_insert_record() (line 609 of /var/www/2020-02-18/drupal-7.69/sites/all/modules/tripal/tripal_chado/api/tripal_chado.query.api.inc).

To test:

$x = chado_get_schema('fileloc');
var_dump($x['fields']);

... part of output omitted ...
  ["size"]=>
  array(3) {
    ["type"]=>
    string(7) "varchar"
    ["length"]=>
    int(1024)
    ["not null"]=>
    bool(false)
  }
  ["filename"]=>
  NULL
}

Found the bug in tripal_file.install line 808,
$schema['fields']['filename'] = $filename_schema;
forgot to define this variable!

[dsenalik]

pull #27

As far as relationships go, it is a very simple change to add a file_relationship table to this module. Then we can use the sbo__relationship field. And just a little more work to create the autocomplete for the widget for this field.

My first use case is for a vcf file, I would like to link it to the fasta file of the reference genome that was used for read mapping, using the sequence ontology term reference_genome (SO:0001505)

[dsenalik]

While there may be additional aspects to file relationships to consider, I am closing this with merged #27
Any further work on relationships will probably only happen in Tripal 4