hmpf / easydmp Goto Github PK

On the Paas I cannot see the section introduction before each section for the NFR template.

If NIRD Storage chosen, ask for how much needed then quota a price.

The help text for the title field is a bit confusing. Perhaps: "The title of the final plan document itself, used as the topmost header in the generated file." should be modified to "This title will also be used as the title of the generated data management plan document."

A demo site should prominently warn people that it is a demo site, ditto for a review site.

One way is to load something extra on deploy, but that'll need to be hand made per type of data. Another is to use the Sites-framework and add data to that, a SiteConfig model which can hold any site-specific data. Review-sites are a bit odd, they don't exist for long, so a way to tell the site that it is a review-site would be useful.

A related thing: how to make it easier to customize which auth-backends to use. B2ACCESS and Dataporten is basically hardcoded today.

Instead of having a dropdown of all available templates (if more than one), use a wizard. First go to a "choose template"-page. Show the generic templates here, + one "domain specific/other", that leads to a new page with every single template.

When a template is chosen, then ask for plan name etc.

If I try to create a plan on the https://easydmp.paas2.uninett.no/plan/new/ I get a Server 500 error.

Would it be possible to add a link in the footer to the EUDAT terms of use document:
http://hdl.handle.net/11304/e43b2e3f-83c5-4e3f-b8b7-18d38d37a6cd

Also, should we have a link to the data privacy somewhere in the footer too?

See issue #20

Thanks for the info. We have an issue (I'm not sure what to label it so you can choose). As it stands the comment button is active when a person enters a question. They can fill in a response to a question and they can, at any time, also click the comment button which will result in the person losing the response they filled in.

Do you think you can either deactivate the comment button or hide it such that it only appears once the user has answered the question? But, then if the user goes back to a question that has already been answered how would we take care of that situation?

Perhaps it requires a rethink of the comments issue. I have some thoughts. I will try to articulate them in the google doc.

Philipp C says: "What documentation?", which might mean that he doesn't see this page as connected to the previous page. Also, what do we do if there are no urls to the documentation?

A plan that is cloned from another plan is "related" to the first plan. Any questions in the two plans that are identical (type and values), can be easily converted between. There will no doubt be questions in either plan that are not in the other, or that otherwise needs to be converted by hand (like essay question to integer..) so the conversion cannot be 100% but every question counts.

What's needed for this is that all clonable models have a field pointing back to the original row. (This should also make cloning easier..)

With that in: a way to run through the existing plan, look up the answers in the old template and map them to the new.

This issue is related to #36 and #33. The problem is that on the production easydmp it takes more than 10secs to move from one question to the next.

As per Philipp C: "What about INTERNATIONAL archives?"

When accessing the URL: https://easydmp.paas2.uninett.no/plan/ and trying to open the 'Test plan' by
clicking on the name 'Test plan' I get a Server 500 error. I'm not sure what is causing the problem. I think it is following the H2020 template.

Sometimes, it is well known for instance which file formats or metadata standards are in use for a discipline (or are supposed to be used..). It should be possible for a template designer to create a subset from existing external entities (and add some manual ones maybe).

This needs to be stored somewhere else than the EEStoreCache, but items copied from the EEStoreCache should point back probably.

This will also need a new question type: the curated list on top and a normal eestore lookup below.

The path finding algorithm runs every time Next or Prev is pressed. This is rather slow.

For a finished template, the actual paths never change. The correct way would probably be to store a dmpt.Section's possible paths in the section, and just look up that. The question is how best to do that.

The paths would need to be recalculated every time the template was changed, but that's not supposed to happen once a sepcific template is in use by end users.

I'm putting here one of Damien's comments:

"""
Q2.1.5a What services will you use to provide searchable metadata?*
Proposed list of services (B2find & OpenAIRE) look a bit weird. We should assume that most people don’t know these services and without any concrete pointers to these services it is difficult for the lay person to tell anything. In particular, there is no indication of the work that is required to get your MD harvested by b2find. Is it automatic? As for OpenAIRE, the specific service should be mentioned (i.e. not just OpenAIRE). Finally we should allow people to choose “other” as an option, and tell us which service. Not just indicate them as “Not listed”.
"""

I don't think we can sort this out for Friday, but can work on it for an imminent release?

Plan 8 has a complete set of answers for the H2020 template, but is very slow to load. Too slow for some timeouts on Kubernetes apparently.

X-Frame-Options: ALLOW-FROM https://bibsys.no/

In the production easyDMP when a user tries to make a change to an existing response in a data management plan the changes are not 'saved'. If the user goes to the plan summary and then views the response they see that the changes are missing.

A plan (on the review-system) currently has two types of users:

• viewers
• editors

Being an editor implies also being a viewer.

There's no easy way to convert between the two. All changes to user-access to a plan need to be collected in one place. That is:

• invitations, and everything connected to that
• changing the status of an existing user, including removing that user. But who gets to remove users?
• user should be able to lower or remove their own access to a plan, so that an editor can change themself into a viewer instead, or leave the plan altogether.

Is it asking for years? Months? Days? Needs a help text or a better stating of the question.

If a user doesn't fill in a question and clicks on the overview they get an error telling them to answer the question

It must be easier to create a new template.

The first time clicking "next' on a branching question is broken.

If I own a plan and try and share it with myself with the same email address the service doesn't detect that I am trying to do this. Should we prevent a user from trying to share with herself? Perhaps detect and say you already own the plan or this plan is already shared with this user?

Put in a link to the Sigma2 userguide in the footer:
https://www.sigma2.no/easydmp/how-to

See issue #20

See issue #20

Pre-save the calculations in the template so that going next-prev is a matter of looking up instead of calculating real time.

Q3,1a,b,c are optional so it should be possible for the user to just skip the questions. I'm not sure how we can allow the user to do that. I think we may need a Q3.1 perhaps. I think it could be "Q3.1 Do you plan to manage your data yourself or do you intend to outsource the management to some infrastructure?" then we have 3.1a,b,c.

Thoughts?

We need a log to see who has done what to what.

• Invitations sent, accepted, revoked (revoke=delete invitation, not rescind access)
• Plans created, locked, published
• Users created, probably

There might not be any other metadata standards in use after all. Perhaps a new type, fill-in-urls or toggle "cannot fill in urls"?

In the review version if I click on 'comment' I am presented with the text field, but my answers to the question are not saved. Is it possible to somehow save the question?

Since it is now possible to skip between obligatory questions, reaching the last question no longer guarantees that all questions have been answered.

Therefore, we must manually validate a plan.

• Backend support for validating is needed, issue #21
• UI to show what is validated is needed, issue #22
• Some way to mark a plan as valid is needed in the REST API, issue #23

After the commits in #33, if a user visits for instance the summary page of a plan without being logged in, instead of beuing redirected to the login page, a 500-error is generated. This is due to the shuffling of when and how several database lookups were done are incompatible with the LoginRequiredMixin.

Usually, a data management plan will be created before a project exists. There can be cases where a user wishes to create a data management plan for an existing project. What we probably want is the ability to choose if this plan is for a new project or an existing one. I'll update the template with the information needed.

I created a plan following the H2020 renumbering. I filled in some of the questions and then went to 'My plans'. I then clicked on the plan title link and get a server 500 error (https://easydmp.paas2.uninett.no/plan/124/) seems to be the page giving the error. I also get a server 500 error if I click on the html link.

This question needs a branch.

Reported by Philipp Conzett.

Next-button was broken, now works.

Going back from 2.3.1 ends up at 2.2.6.a regardless.

Two of the choices have identical canned texts.

I created a linear template and I had on one page one question that was optional. When I started to fill in the plan and then clicked on 'next' to go to the page with the optional question I got a 'page not found' error. I then changed the question to mandatory and was able to go to the next page.

For the H2020 linear template and the NIBIO template which have more than one question on a page when I fill in plan details and click 'next' the details are not saved. If I then click on the plan summary I cannot see the information that I filled in.

When I try to delete a plan I get a server 500 error after confirming the delete. This is for the review easydmp.

Please feel free to split these remarks up into multiple (GitHub) issues.

GENERAL REMARKS
(For more specific feedback, see section SPECIFIC REMARKS below.)

• Only one, predefined template is available.
• One cannot define own templates.
• One cannot provide guidance for researchers.
• No institutional access to employees' DMPs.
• One cannot define different templates for different stages in project/data lifecycle.
• In many parts of the template there is no clear distinction between the different phases of the data lifecycle. It is, e.g., often unclear whether the term "data" refers to active data og archived data / data to be archived.
• Many questions are hardly understandable by "common" researchers.
• Some questions are superfluous since they are based on the answer of other questions.
• Some topics are covered/asked for through multiple questions resulting in redundancy.
• Some of the answer alternatives are very long containing identical text apart from the one main issue at stake. Rather use only the main issue as alternatives.
• Some pulldown menues lack crucial alternatives.
• System is slow; it takes ca. 10 secs. to go to next step.
• Misguiding progression information at the bottom of the tool; cf. "Question X/Y". In the FAIR section, progression starts from 0 several times.
• Cannot export plans to PDF, doxc, etc.
• Section breaks disappear in exported version.

SPECIFIC REMARKS

Q1.4 Will your project reuse data produced by another project?

Or rather "... existing data"?

Q1.5 Please specify where the data originates from. Derived covers data from simulations, models, etc. Experimental covers data arising from experiments made on objects (such as particle physics, chemical reactions etc). Observational covers data from observation of objects (such as earthquakes, sunspot activity, etc).

What about research on human beings? This explanation is somewhat natural science biased.

Q1.6 Please estimate the total amount of data your project will produce.

Missing option "Not sure / do not know."

Q2.1.1 Will you use metadata to describe the data?

The answer depends on funder requirements. Many funders demand documentation of data.

2.1 Making data findable, including provisions for metadata Please select all metadata standards that will be used

Researchers may be aware of subject-specific metadata standards, but hardly know Dublin Core citation metadata. Such information will be available automatically if one first specifies the archive one is going to use.

2.1 Making data findable, including provisions for metadata Where will the metadata be stored? Please provide the URL for the registry that you intend to host your metadata.

Is "registry" the same as archive? Do researchers understand this?

2.1 Making data findable, including provisions for metadata Please select all relevant standardized vocabulary

Where are the ISO standards?

Q2.1.1b Will you make the metadata available free of charge?

Depending on choice of archive.

Q2.1.1c Will your metadata be harvestable?

Does a "common" researcher know what harvesting is?

Making your metadata harvestable by providing, for example, an OAI-PMH interface makes it easier for other metadata registries to store your metadata increasing the exposure of your work.

I doubt that a "common" researcher knows what OAI-PMH is.

Q2.1.3 Will you provide clear version numbers for your data?

Is this referring to active or archived data?

Q2.1.4 Will you provide persistent identifiers for your data? Persistent identifiers are independent of the physical location of the data. Using persistent identifiers means you have the freedom to move the data to new storage without impacting your users.

1. "Common" researchers do not know what PIDs are.
2. The answer to this question is based on the choice of archive.

The data will be issued with DOI identifiers once the data has reached an approved level of maturity for consumption by interested parties.

Clumsy wording. Usually, data get a PID when they are deposited into an archive.

Q2.1.5 Will you provide searchable metadata for your data?

The answer depends on the choice of archive.

Q2.1.5a What services will you use to provide searchable metadata?

Not understandable for a "common" researcher. Depends on choice of archive. Researchers choose archives, not "service[s] [...] to provide searchable metadata".

Q2.1.6 Will you use standardized formats for some or all of your data?

How does this question relate to the earlier question(s) on file format?

Q2.1.7 Are the file formats you will use open?

1. Do researchers know what this is, and if the formats they use are open?
2. The question is rather whether they archive data in open formats?

Q2.1.9 Will you provide metadata describing the quality of the data?

Shouldn't this come right after the other questions about metadata / documentation?

Q2.2.1 Are there ethical or legal issues that can impact sharing your data? You should consult your research office to understand if your data is subject to these constraints.

Here, one could have integrated a tool like DataTags.

Q2.2.2 Will all your data be openly accessible?

Depends on funder requirements.

Q2.2.3 How will the data be made available?

Shouldn't this be related to the question of where metadata will be made available?

2.2 Making data openly accessible Q2.2.4 Is the storage sufficiently secure for your data? Sufficiently secure storage will provide access control for non-public data and will provide backup and recovery procedures. Ensuring your data is secure will ensure your data remain accessible during your project and afterwards. The data will be stored in a repository that is sufficiently secure, but does not provide backup or recovery of the data. The data will be stored in a repository that is sufficiently secure and that will ensure the data is backed up and can be recovered. The data will be stored in a repository that is not sufficiently secure and that does not provide backup or recovery procedures. The data will be stored in a repository that is not sufficiently secure, but that provides backup and recovery procedures.

The section is about accessibility. But safe storage is mostly about active data? Differentiate between storage of active data, and archiving of data at the end of the project. Are those two mixed up here?

Q2.2.5a Please provide links describing the documentation for accessing your data.

Not sure if I understand this question.

1. Links do not describe, but link / refer to documentation.
2. Such documentation is available in the archive where data is archived? Can we combine this question with the question(s) about where metadata and data will be archived?

Q2.2.5b Please provide links describing the tools for accessing the data.

See previous remark.

Q2.3.1 Will you use a standard vocabulary for your data types?

1. In what way does this question differ from earlier questions about metadata vocabulary?
2. Does a "common" researcher know what this means?

Q2.3.1.a Will you provide a mapping to more commonly used ontologies?

See previous remark.

Q2.4.1 What internationally recognised licence will you use for your data?

Depends on the choice of archive and/or funder requirements. Alternatives "None", "Do not know" are missing. Crucial licenses like CC0 are missing.

Q2.4.2 When do you plan to make your data available for reuse? Even after your project completes your data may still have value to fellow researchers. You should consider providing access to other researchers once the data is no longer of primary value to you.

1. Depends on funder requirements.
2. Strange wording "Even after ...". Sharing data at the end of a project will for many researchers be one of the most important points in following good practice for open science. The way this explanation is formulated, one might get the impression that the previous questions are dealing with something quite different.

Q2.4.3a Please input url (s) to your documented QA procedures.

Shouldn't this question be related to the questions about where data and metadata will be made available?

Q2.4.4 Will you provide any support for data reuse? Providing documentation and a contact person can help researchers make better use of your data and increase its usefulness.

Belongs to the question(s) about documentation.

Q3.1a Will the project make use of national infrastructure? Using national infrastructure for managing your data and metadata can often help to offset the cost of managing your data. The infrastructure will be familiar with the requirements for managing data which should reduce the burden on your project.

This question + choice of archive + info on funder should be included earlier. This is information that in many cases determines the answer to many of the other questions that are asked in this form.

Q3.1b Will the project make use of institutional infrastructure? Institutions increasingly provide data management resources (for example, storage with backup, metadata databases) for research projects that may help reduce the burden of data management for your project.

See previous remark.

Q3.3 How do you intend to ensure data reuse after your project finishes? Storing your data in an archive where it will be safely managed and findable will ensure researchers can use the data once your project completes. It can offset the burden of managing your data over the long-term.

Haven't we answered this previously?

Q4.1 What do you plan to do with research data of limited use? Some of the data you collect will not be of interest to researchers intending to use your data. Typically these data are raw data (such as raw signals from an instrument) that need to be transformed with software into more meaningful objects.

1. I'm not sure about how well founded the claim is that raw data often are uninteresting for other researchers.
2. Strange wording "of limited use". Do you mean "usefulness"?

Q5.1 Are there other ethical aspects that your data are subject to that have not been covered by previous questions? You should consult your institution and funding agency guidelines on ethical aspects and document whether your data are subject to those guidelines.

Similar questions popped up a little earlier in the form. Why not group questions regarding ethics and privacy in together?

Q6.1.1 Do you make use of other procedures for data management? Your project may also be subject to further guidelines defined by your institution or funding agency.

That's precisely why one should ask for this kind of information early in the form.

We used these fields to retrieve a users plans

NSD fetches project names from CRISTIN and uses them to suggest plan names. This doesn't need to be in the backend. But if we also link to the CRISTIN project description we need to store (url, maybe? eestore id? json blob?) to the project on the backend. This hooks into storing project info in the plan.

The *_by-fields on a plan are linked directly to a person. If that person deletes his user, unless we change things, any plan with that user in a *_by-field will be deleted. This might be fine for unpublished plans, but not quite so for doi-ed, published plans and certainly not for plans with more than one editor.

The easiest fix to preserve plans is to set the *_by-fields to empty. We could also remove all *_by-fields and use a logging-mechanism instead. But you're not supposed to ever change a log.

Importing a plan depends on the template already being in, so import/export of templates take priority.

The main problem is the use of autogenerated primary keys.

Ideas:

• Add natural keys to everything in dmpt so templates can be exported.

  Turns out, CannedAnswer/ExplicitBranch is a little unspecified so nothing added for them.

• Export of templates

  • Serializer for exports
  • API/link for exports
  • CLI
  • Maybe admin?

• Import of templates, in order of dependencies:

  • 1. A function/method for import
  • 2. A mangement command to use from the CLI
  • 3. Admin interface
  • 3. API interface #241
  • 4. Frontend interface #242

• Plan import/export: map question id's to natural keys.

Natural keys:

Originating site should be part of it. URIs?