Publishing is the process by which you can make any item accessible to the greater precisionFDA community. By using publishing, you can contribute reference data, software, results, or pretty much anything else, so that others can use them. The process of publishing changes the access permissions of an item from 'private' to 'public'. NOTE: Currently there is no way to "unpublish" an item (other than deleting it).
Since publishing an item means that others will be able to access it, it's very important to understand the exact implications, to ensure that you make a conscious decision. The following table summarizes what it means to publish items of a given type:
|Type||What does it mean to publish an item of that type?|
|App||The app will appear in everyone's list of available apps to run. Users will be able to run the app, and view the app's specification (input and output spec, VM environment details and app shell script).|
|App asset||The asset will appear in everyone's list of available assets. Users will be able to use the asset when making apps, view the asset's information (readme, and list of files), and download the asset.|
|Comparison||The comparison will appear in everyone's list of available comparisons. Users will be able to view comparison details (including statistics such as sensitivity and positive predictive value), visualize the comparison, and download any of the comparison output files.|
|File||The file will appear in everyone's list of available files. Users will be able to download the file or use it as input in their jobs.|
|Job||Users will be able to view job details, such as which parameters/values were provided as inputs. NOTE: users will not see the job's log.|
|Note||The note will appear in everyone's list of notes. Users will be able to read the note.|
The publishing interface presents you with the item you chose to publish at the top, and preselects that item for publication. If the item has related items, then those are also offered for publication. Among the related items, those that are already public are shown as such and cannot be selected for publication. Those that are owned by you and can be published have a checkbox next to them. Tick that checkbox to enable publication. If the item you tick contains additional related items, those will be presented as well, etc. IMPORTANT: For items with a long chain of dependencies (such as a file produced by a series of consecutive jobs) this interface can get very deep. For that reason, prior to publishing any item, we strongly suggest that you first track the item to understand exactly its dependencies and their access level. The items offered for publication are derived in the same way as the "source items / ingredients" shown in the tracking graph. The tracking graph offers a more concise way to summarize the full landscape of relationships.
For example, if you publish a note, the publishing interface will ask if you would like to also publish its attachments. If you publish a file, it will offer to publish the job it was generated from, and the app that was used for that job, etc. We recommend publishing related items, to ensure that others have access to complete information and are able to reproduce your work. To better understand the implications of NOT publishing a related item and instead keeping it private, consult the following table:
|If you publish a...||But not a related...||Then|
|Note||Attachment||When reading the note, users will see the attachment by its id and not its name (i.e. file-Bk0kjkQ0ZP01x1KJqQyqJ7yq) and will not be able to access the attachment.|
|App||App asset||Users will still be able to run the app on precisionFDA, but they will not be able to access its private app asset. As a result, when users fork the app, the private asset will not be included in the list of assets, essentially leading to a nonfunctional fork. Moreover, users won't be able to run the app outside of precisionFDA, since they can't download and extract the private asset.
Keeping an asset private may be needed in the case where you don't want people to access sensitive software information, such as any code that is not open-source and whose executables you don't want to openly distribute.
|Comparison||Input file||When viewing comparison details, users will see the input files by their ids and not their names (i.e. file-Bk0kjkQ0ZP01x1KJqQyqJ7yq) and will not be able to access the input files. However, they will still be able to access the comparison outputs, which pretty much provide for a copy of the comparison inputs (since they simply split the input in terms of true positives and false positives, etc.)|
|File||Job||In the file listing, users will see a job id (i.e. job-BjgGbqQ02308XVYgV4px5Zy2) in the "Origin" column, and will not be able to see further information as to how the file got generated.
Keeping a job private may be needed in the case where the job accessed sensitive information. For example, when you bring a file into precisionFDA by running the "Fetch file from URL", and later publish the file, you will be offered to publish the job. If the input URL is sensitive information, then keep the job private.
|Job||Input file||When viewing job details, users will see the input files by their ids and not their names (i.e. file-Bk0kjkQ0ZP01x1KJqQyqJ7yq) and will not be able to reproduce the results starting from the same inputs, since those are inaccessible to them.|
|Job||App||When viewing job details, users will see the app by its id and not its name (i.e. app-Bk7jzXj02902969jx5q2pfzq) and will not be able to reproduce the work, since they can't run that app.|