galterlibrary / rdcat Goto Github PK

Currently, the source for the organization field is http://www.feinberg.northwestern.edu/research/clinical-trials/. This should be changed to include all affiliations within Feinberg.

Come up with an alternative to DOI.

In addition to 'creator' and 'maintainer', this would allow us to give credit to those who helped create the data set in some way. We also probably need a list of 'contribution types' to go along with this.

We're on Rails 5.0.0-rc1, most other libraries are way out of date. It's time to update everything and then fix the resulting issues. Initially, I'll update to rails 5.0.x, I'll open another ticket for Rails 5.1 update.

This needs to be a part of the deployment.

Needs #17 to happen first.

Not sure if this is the best way (since .zip files will be identified as such and actual file types could not be captured. Is there a better feasible way?

Run a background job to identify MIME type display in a user-friendly manner.

chosen is too slow for close to 30K terms. And it lacks features such as displaying descriptions with the matched word.

Similar to LinkedIn, user profile pages should show...

• Who looks at your dataset
• Who has made similar searches
• Choose to be anonymous OR open level - possibly a middle level(?)

For now this can just be a plain text field.

Or an object store that supports RDF, or something

See the example for 'Grants and Funding' in the screenshot below:

Datasets currently can be associated with any number of distributions. We want to to be able to create dataset + distribution pairs that we could 'freeze' (not allow edits) and version. Any edit would result in a new version. The ultimate goal is to be able to mint DOIs for dataset-distribution pairs.

• Allow creation of fixed Dataset + [Distribuitions] sets (collections?) that can be frozen
• You can edit the Dataset in a set but not the associated distribution
• You can mint a DOI for the set #18
• Version the Dataset associated with the set (optional and might be annoying to implement)
• Create a GUI that will allow DOI minting. Ask user to chose the dataset and the associated distributions and inform them that they won't be able to change the association once the set is created and DOI minted.

• Is it PHI? SPHI?
• If it's clinical trial data, is it consented or unconsented?
• Is an IRB required to use data?

It's not needed anymore and it needs to be removed.

See http://www.oclc.org/research/themes/data-science/fast.html for more details on FAST

Add MESH subjects and MESH pubtypes (https://www.nlm.nih.gov/mesh/pubtypes.html). Improve autocomplete so the completion starts at the beginning of a word instead of a phrase.

• If a data set shows up in a search by another user, notify the owner of that data set that their data set showed up in that search.

• Possibly an opt-out option for this feature

User profiles (data from Symplectic/FASIS?) would help other users identify those with expertise in their area of interest. Useful features would include:

• list of publications, or number of publications
• data set owners need evaluation tools to see if match would be beneficial

I realize that this is vague. We probably just need to talk about this.

Currently we're running a version that's missing bunch of features compared to:
https://rdcat-staging.nubic.northwestern.edu/datasets

Currently it just throws an error

There's no deployment scripts right now that I can find.

Related to #38 in one way or another...

Replace "CDSI Data Indexing Project", "Dataset Indexing Project"...etc.

This could be acquired from InfoEd or possibly NIHReporter. Each of these sources has their own benefits and limitations.

Get this done as soon as we get the servers from FSMIT

Right now it just selects from the DB

• Data catalog entry creation date
• Beginning and ending dates of the associated research project
• Publication date(s)
• Embargo period

Dataset is still visible to anonymous on the show page not index.

Anyone can delete a dataset also those admin links look pretty bad.

Allow users to set up recurring searches within their profile.

Similar to the field used for DigitalHub files.

On both the "Create Dataset" form and the "show" dataset page.

Make the home page google.com-like? Keep It Simple approach.