nulib/arch

Northwestern University institutional repository, built on Samvera's Hyrax gem.

Arch

Arch is our institutional repository and is a Hyrax 2 application.

Environments

• Production: https://arch.library.northwestern.edu/
• Staging: http://arch.stack.rdc-staging.library.northwestern.edu/
• Local development: https://devbox.library.northwestern.edu:3000/

Software Architecture

Arch is hosted on AWS. It's using NUL's shared Fedora 4 instance and has it's own Solr core on a shared Solr instance.

Developer Dependencies

These should match closely with the Hyrax requirements:

• Ruby 2.3.0 and rbenv, Bundler (https://github.com/rbenv/rbenv#homebrew-on-mac-os-x)
  • If you get permission errors running gem install bundler, try the following:
    • Set your global ruby environment by running: rbenv global 2.3.0
    • In your user's .bash_profile, include this line at the bottom: eval "$(rbenv init -)"
    • Open a new terminal window and run: gem install bundler
• Download and install LibreOffice https://www.libreoffice.org/download/download/
• Imagemagick brew install imagemagick --with-ghostscript --with-openjpeg
• ffmpeg brew install ffmpeg --with-fdk-aac --with-libvpx --with-libvorbis
• fits brew install fits
• vips brew install vips
• Install devstack according to the instructions in the README
• Follow the Dev Environment Setup instructions

Developer Installation

• Clone this repository git clone [email protected]:nulib/arch.git
• From inside the project directory run bundle install
• Start the docker stack with devstack up arch
• From inside the project directory run bundle exec rake arch:seed
  • You can include the optional arguments to create an admin user (such as yourself). Ex: bundle exec rake arch:seed ADMIN_USER=your_netid [email protected]
• In a separate tab, start the rails server bundle exec rails server
• You can see the app in a browser at https://devbox.library.northwestern.edu:3000

Running the tests

• Start the test stack devstack -t up arch
• Run the seed task for the test environment: bundle exec rake arch:seed RAILS_ENV=test
• Run the test suite: bundle exec rspec

Deploying

• Submit a PR in github to the environment's deploy branch
  • To deploy to staging submit a PR to deploy/staging
  • To deploy to production submit a PR to main

arch's Issues

Investigate SWORD compliance for SUFIA/Arch

Descriptive summary

SWORD is the protocol used by ProQuest for the automatic delivery of electronic theses & dissertations to institutional repositories. This is the desired functionality for getting dissertations produced at NU into the repository. As a workaround, we can batch import our dissertations from the using the current workflow in Metadisco.
SWORD Implimentation for ETD Administrator 5_20_14.docx

Done looks like

• Dissertations disseminated via SWORD can be automatically ingested into Hyrax.

ssh access nufiarepo-s and nufiarepo-p

ssh access to for karen, brendan, chris, mbk

nufiaweb-s intermittent Internal Server (500) Errors

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Staging Environment (nufiaweb-s) sometimes throwing Internal Server Errors instead of serving page.

Expected behavior

Navigating to a page displays the page in the web browser

Logs show a 200 response to user

Started GET "/concern/generic_works/new" for 165.124.200.29 at 2016-11-03 11:12:18 -0400
...
Completed 200 OK in 1560ms (Views: 1504.5ms | ActiveRecord: 6.9ms)

Actual behavior

Error page displayed to user

Logs show a 500 response for the request

Started GET "/concern/generic_works/new" for 165.124.200.29 at 2016-11-03 11:06:53 -0400
....
Completed 500 Internal Server Error in 314ms (ActiveRecord: 4.2ms)

Steps to reproduce the behavior

1. in your web browser, navigate around nufiaweb-s, particularly new works and new collections links
2. Observe intermittently the error page displayed instead of the desired page

Swap Thumbnail Figure Position with Metadata

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Swap the positions of the thumbnail figure with the metadata, so that the thumbnail displays on the left side of the content wrapper, and the metadata displays on the right side of the content wrapper.

Expected behavior

Here are two examples:

Source: https://era.library.ualberta.ca/files/xp68kh355#.WFRNXrIrKUk

Source: https://digitalhub.northwestern.edu/files/kh04dp681

retest after 7.2 update

@cpd3149 Can you run through this list and verify whether or not we got some "freebies" with the upgrade.

• Facets for Limiting Search Results is Missing bug ready ux #46
• Contact Form Doesn't Deliver Messages bug ready #42
• Adjust width constraints to main content wrapper high priority ready #41
• Resize top bar CSS dimensions to 1583 x 60 enhancement ready ux #40
• Batch Deleting private files reappear as deposits and attachments bug high priority ready #33
• Batch Create results in duplicate files bug high priority ready #32
• nufiaweb-s intermittent Internal Server (500) Errors devops ready #27

Add comments to the issues themselves, not here. This will act as just a checkbox.

Revisit development environment for ARCH

Descriptive summary

Lock down development environment for ARCH/Sufia. We spent a lot of time setting up vagrant for development boxes, but it may not be worth it for development. Following Avalon's lead, Docker may make more sense. We need to discuss this a bit with Devops.

Done

Team has agreed upon, documented and is using a development environment for ARCH.

• Investigate Avalon's Docker solution, can we roll with that
• Meeting setup with Devops to discuss if necessary
• Document process to ensure that requirements get into the hands of devops (e.g. at each change in requirement, a github issue is written documenting the change and assigned)
• Publish development environment environment setup instructions/ assets

Add "Download" Action to file records

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Please add "Download" as an Action users can perform on file record pages.

For example, users of this file (https://scholarsphere.psu.edu/files/2514nk570) can click on the Download link beneath the file thumbnail.

Can this be added to Arch?

Expected behavior

This will make downloading files easier.

Actual behavior

Users need to select "Download" from a drop-menu.

Steps to reproduce the behavior

Related work

Similar request to #38

fedora down on nufiarepo-s

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Fedora isn't responding: http://nufiarepo-s.library.northwestern.edu:8080/fcrepo/rest (you'll see a tomcat error).

Related work

#22 can we get monitoring set up for fedora on nufiarepo-s too maybe?

Well done

Integrate FAST autosuggest Subject Headings in SUBJECT field

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Adds a controlled vocabulary to the SUBJECT metadata field. OCLC FAST documentation is here: http://experimental.worldcat.org/fast/assignfast/

Expected behavior

A user begins entering information into the SUBJECT field and FAST Subject Headings appear as suggestions. The User selects from those suggestions.

Actual behavior

This is a free-text field, redundant with the KEYWORD field.

Change "Create Work" label to "Upload

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

On the new works form, change label "Create Work" to "Upload" and "View Works" (in the dashboard) to "View Files" to match DigitalHub's (Northwestern Medicine) repository.

Expected behavior

Actual behavior

Steps to reproduce the behavior

1. Do this
2. Then do this...

Related work

Link to related tickets or prior related work here.

Remove 'Lease' visibility option from New Work form

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

We don't want users to have the option to "Lease" their deposits. Please remove.

can't login to arch-s

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

To Reproduce:

• Go to arch-s.library.northwestern.edu
• Click Login
• Set Creds and click go
• See "Incomplete response received from application" Response

Relates to #54 #48

Change "Relationships" Label to "Collections"

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

When creating a new work or viewing the record of an existing work, the label "Relationships" should be changed to "Collections".

Create new work: https://arch.library.northwestern.edu/concern/generic_works/new

Example record: https://arch.library.northwestern.edu/concern/generic_works/hx11xf24j

test

Severity

Is the production site running?

• [ x] yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Include any relevant tracebacks if you're reporting a bug.

Expected behavior

Actual behavior

Steps to reproduce the behavior

1. Do this
2. Then do this...

Related work

Link to related tickets or prior related work here.

Add Bootstrap JS tooltip to metadata create/edit form

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Users are having difficulty with some of the metadata elements on the add new work metadata page, such as first name/ last name order in the Creator field and whether or not keywords can be comma-separated or separate fields.

I recommend using the Bootstrap JS Tooltip plugin (http://www.w3schools.com/bootstrap/bootstrap_ref_js_tooltip.asp), but I'm open to other suggestions.

Here's the box note for the text to be used in the Tool Tips: https://northwestern.box.com/s/1wqycu1atcvahk317njg2w12lwm81t99

Expected behavior

Users will know what to enter in each metadata field. Here's an examples from DigitalHub:

Facets for Limiting Search Results is Missing

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Perhaps there are too few items, but the facets that general appear on the left side of a search results page is missing.

Expected behavior

Actual behavior

Steps to reproduce the behavior

1. Do a search

Related work

Link to related tickets or prior related work here.

Link Thumbnail image to file download

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Please allow users to download files by clicking on the thumbnail image. Here's an example: https://digitalhub.northwestern.edu/files/e13451f1-01ff-4462-acbf-2a87eb2312e2

Expected behavior

When a user clicks on the thumbnail image on an item record, the file will download.

Actual behavior

Thumbnail images are not linked.

Related work

Link to related tickets or prior related work here.

Hide Table Headings on Metadata Records

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Please hide the table headings "Attribute Name" and "Values" on metadata records:

Header links don't work

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Northwestern link and Northwestern University Libraries link in header don't resolve

Expected behavior

Northwestern --> http://www.northwestern.edu/
NUL --> http://www.library.northwestern.edu/

Actual behavior

Links don't do anything

Steps to reproduce the behavior

1. Do this
2. Then do this...

Related work

Link to related tickets or prior related work here.

Thumbnails not generating, files not downloading

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

On staging, thumbnail images are not being generated on work metadata records, and the files are not downloadable. It was thought that this may be related to batch processes, but it is affecting single deposits too. Weirdly, this does not occur with all works. The works where the thumbnail images are generating are downloadable. This issue also appears to only affect the following two examples.

Here's an example from a single file deposit: https://nufiaweb-s.library.northwestern.edu/concern/generic_works/rj4304528

Clicking the "Download the file" link returns this error:

Here's an example from a batch deposit experiencing the same problems: https://nufiaweb-s.library.northwestern.edu/concern/generic_works/nv9352841

Expected behavior

Deposited files should generate thumbnails and be downloadable.

Actual behavior

Deposited files (via both single and batch) are not generating thumbnails and are not downloadable. With regard to batch deposits, this appears to affect the last file in the batch.

Related work

Possibly related to #32 wherein duplicate files are being generated off the last file in a batch that is processed.

Change "Repository | Institutional" to "Arch"

On the Header please change "Repository | Institutional" to "Arch"

Change visibility type label

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

On the Create New Work page, change "Your Institution" to "Northwestern University":

LibreOffice on nufiaweb-s and nufiaweb-p

puppet install yum package: libreoffice-headless.x86_64 : LibreOffice Headless plug-in (need which soffice to return path to the binary like /bin/soffice

Done looks like

libreoffice-headless.x86_64 is puppetized

Sidekiq start/stop/restart scripts on nufiaweb-s (staging)

sidekiq start/stop/restart (see slack notes from Images)

Adjust width constraints to main content wrapper

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

There appears to be a problem with the main style sheet, whereby the content width across the site expands as wide as the browser window.

Expected behavior

Content should be constrained based on the Arch navigation bar. See this page as an example: http://www.library.northwestern.edu/research/tools/index.html

Actual behavior

See:

And:

Related Work

This card is related to another CSS request: #40

Change Arch Header Text to LIBRARIES | Arch

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Arch's header text currently says "Arch" and Marketing/ Communications has asked us to change it to "LIBRARIES | Arch", following the blog as an example: http://sites.northwestern.edu/northwesternlibrary/

Expected behavior

Something like this:

Actual behavior

Contact Form Doesn't Deliver Messages

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

The Contact Form does not deliver messages: https://arch.library.northwestern.edu/contact

I tested this on Chrome and Firefox, both logged in and not logged in.

Expected behavior

Form submissions should email [email protected]

Actual behavior

Steps to reproduce the behavior

1. Go to https://arch.library.northwestern.edu/contact
2. Complete the form

This is a test issue, training.

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Include any relevant tracebacks if you're reporting a bug.

Expected behavior

Actual behavior

Steps to reproduce the behavior

1. Do this
2. Then do this...

Related work

Image for "Preserve and Share your work" box

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Need to distinguish Arch from the NUL website. Could we use one of the images here: https://northwestern.box.com/s/hryyhtgpr43ozbwzhnjfk8jjhtbbgazr in the box that currently has that shades of purple design thingy?

Expected behavior

Main element on https://arch.library.northwestern.edu is image of people doing lab research.

Actual behavior

Main element on https://arch.library.northwestern.edu is that shades of purple design thingy?

Steps to reproduce the behavior

1. Do this
2. Then do this...

Related work

Link to related tickets or prior related work here.

add google analytics

Done looks like:

• GA code is added to page. UA-797260-32
• Jon Dorr is trained on google analytics.
• add filter for separate arch and images views

ssh access for ksd927 on nufiarepo-s

I have access on nufiarepo-p but have lost access to nufiarepo-s.

change font color of title on work view

The work title is black on purple and unreadable. Can we tweak it.

files not attaching to work

To Recreate:

• Go to arch.library.northwestern.edu
• Login
• click "deposit your work"
• fill out required stuff
• attach files (pdfs, images)
• note that items are never attached

Batch Deleting private files reappear as deposits and attachments

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

After testing the batch create function with dummy files, I attempted to delete all of My Works. All of the works I have uploaded have been marked as Private. Deleting the files does not result in any errors (I receive a message saying "Batch Delete Complete"; however, the files are reappearing automatically as deposits and attachments. Here are files I deleted yesterday, nearly 24 hours ago:

Expected behavior

Batch delete should delete files permanently.

Actual behavior

Batch deleted files reappear automatically several hours after the deletion occurs.

Steps to reproduce the behavior

1. Batch delete a number of (Private) files.
2. Check 2-3 hours later to see if they have, in fact, been deleted.

Related work

This issue is related to #32

Create Test Works for Staging

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Create a batch of test works for the staging environment and share in Box.

Related work

Related to #36 #37 #33 #32 #46

Batch Create results in duplicate files

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

The batch create function for uploading files from a folder is duplicating files.

Expected behavior

Batch create should add one copy of each file in the selected folder.

Actual behavior

I tried this twice using a folder of two PDFs (I deleted the files from My Works between tests). Each time I tested the function, the result would be two to five copies of each file in My Works.

Steps to reproduce the behavior

1. Create a folder with 2 or more PDFs.
2. Upload the folder using batch create.

Sufia layout issues

see here for more details! https://github.com/nulib/repodev_planning_and_docs/wiki/Arch-Technical-Documentation

update footer using library footer with tweaks

Please update footer to be the same as the main library's website with two exceptions:

1. Email address should be [email protected] (not [email protected])
2. Get rid of the FAQs link. Instead we will link to the Terms of Use.

Here¹s the issue: Terms of use is a word document on box. Can we format it
to be a web page within the Sufia app?

change <title>

Currently the <title> on the homepage is Sufia

Please change to Arch : Northwestern University Institutional Repository

Change Nufia-s to Arch-s

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Change https://nufiaweb-s.library.northwestern.edu/ to http://arch-s.library.northwestern.edu/

Expected behavior

Actual behavior

Steps to reproduce the behavior

1. Do this
2. Then do this...

Related work

Relates to #24

update sufia to latest 7.x

we're running the 7.0 beta, now is a good time to bump up to the latest version

IR OAI PMH is turned on

FINAL:

• The ResourceSync specification seems to supersede the OAI-PMH, and it is auto-enabled in Sufia: http://www.openarchives.org/rs/1.1/resourcesync

• IR is registered with SHARE (https://osf.io/share/registration/) and can harvest metadata

NOTES
Note, we're not sure whether this needs to be turned on in Fedora... there is a library.

We need to investigate how to deal with the Fedora server. We may need to open the tomcat/fedora server to the world. We don't know how to do this right.

Project github repo and README: https://github.com/fcrepo4-labs/fcrepo4-oaiprovider

Installation wiki: https://wiki.duraspace.org/display/FEDORA40/Setup+OAI-PMH+Provider

make sure sidekiq workers are restarted on deploy

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Include any relevant tracebacks if you're reporting a bug.

Expected behavior

Actual behavior

Steps to reproduce the behavior

1. Do this
2. Then do this...

Related work

Link to related tickets or prior related work here.

Resize top bar CSS dimensions to 1583 x 60

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

Marketing \ Communications asks that we resize the top-bar to 1583 x 60 pixels. Right now, the top-bar has a min-height of 50px; whereas, the other top-bars on Northwestern website have min-heights of 60px.

I'm not finding a max-width specification in the CSS, but 1583px is recommended.

Add Analytics action to file records

Severity

Is the production site running?

• yes
• no

Are staff blocked from performing their work?

• yes
• no

Descriptive summary

The Analytics feature allows users to see views and download statistics at the file-level: https://scholarsphere.psu.edu/files/6682x4006/stats

If this is available, please add as a default for all works.

Related work

Similar request to #36

Customize generic work fields if necessary

Add "sort by creator" to sort drop-down menu

Customization. Add a "sort by creator" to drop down menu.

Tomcat nufiarepo-s and nufiarepo-p

investigate tomcat on server restart / tomcat monitoring

Log Rotation for Arch

Could we have log rotation setup for the Institutional Repository (nufiaweb-s and nufiaweb-p) the same way it was recently configured on Images?

• Logs rotated daily and archived for 60 days
• Logs are in /var/www/nufia/shared/log
• There is a sidekiq.log and a production.log (or staging.log)