seek4science / seek Goto Github PK

Hello,

when uploading a CSV file to seek 1.1.1 the explore button is enabled. However, this results in a crash (500 error) as the simple extractor cannot deal with csv files. The view option on the other hand is disabled even thought the tooltip says that csv files should be displayable.

Ideally csv and tsv files could both be explored, it would be easy enough to have those files be converted to spreadsheets.

cheers
Frank

Hi (again),

I have new users in my SEEK instance and they touch everything. This explains all my issues ;)

Thus, we observed that there are no version link concerning a data or a sop item. It's hard to explain so I'll describe it :

1. create a SOP
2. link the SOP to an assay
3. upload a new version of the SOP
4. in the related items and in the graph, it's the latest version (v2) of the SOP, and not the true SOP (v1)

I think this can be a real problem for the data reproductibility.

Putting this here since JIRA is down.

Add an option in the admin setting to completely disable the pop-up, or only show it when permissions have changed.

Hello everyone,

I have recently tried to upgrade our SEEK instance from 0.21 to 1.0.2 by following the commands listed in the docs at:
http://docs.seek4science.org/tech/earlier-upgrades.html
http://docs.seek4science.org/tech/upgrading.html
When I failed to get bundle registered as a command upon installing a newer ruby version with rvm install ruby-2.1.x, I (no ruby experience at all) realized after some time that I needed to run gem install bundler before continuing. Could you change the docs to mention that?

Hi, research team,

first of all, thank you very much for your new version which is great, especially the JSON API.

I want to use this API to extract information from my SEEK instance. I realized that an administrator can't see all the content (hidden content is not accessible).

So have you thought about creating a new role, such as a super-administrator, or improving the administrator's power to see all the hidden content?

Thank you very much,
Cyril

f I add an outgoing datafile then all the samples become outgoing. The XML and the Administration are correct. So it is the generation of the JSON

It's correct in the Administration form and in the database but not in the ISA graph.

We checked and incorrect JSON is being used for the graph

We have large RNAseq data stored on a SFTP server behind standard security.
Now we would like to link these datasets into our private SEEK platform however SFTP is not allowed.

Would it be possible to add SFTP support with or without authentication?

Putting this here since JIRA is down.

Currently you can choose permissions for either:

• Project members & registered users
• All users (registered & anonymous, regardless of project)

need a combo of both, so can easily specify permissions for project members (e.g. view, download, edit) and all users, anonymous or registered, (e.g. view only).

Hi,
I'm working on NGS data and we try to evaluate the use of SEEK.
I've got several files (fastq), each one corresponding to a strain (of Brassica napus).
I've performed different analysis on those files.
In the actual state of the SEEK project assays are linked to strains, but not files.

I'd like to link file to strain once and then deduce that an assay involved a strain by the file it uses.

Hi seek team,

first, happy new year with many seek successes :)

I have a question concerning Zenodo connection with snapshots. I configured with sandbox.zenodo.org in the admin panel but nothing append when I create a snapshot. The Zenodo URL is always "None".

Maybe I made mistakes during Zenodo or Seek configuration. Do you have a documentation?

Thanks,
Cyril

Trying to use seek for RNA-Seq data. How can I give the assocation between a list or individual sample and individual data file or set of files (with assay results)? I can't see any field on Sample to link to data files...and I can't see any field on Data Files to link to individual samples. I am starting with ISATab files...

In 'lib/seek/grouped_pagination.rb' and 'app/controllers/publications_controller.rb' there is some code for string normalization/transliterate:

        #handle the characters that can't be handled through normalization
        %w[ØO].each do |s|
          text.gsub!(/[#{s[0..-2]}]/, s[-1..-1])
        end

        codepoints = text.mb_chars.normalize(:d).split(//u)
        ascii=codepoints.map(&:to_s).reject{|e| e.length > 1}.join

There are some post on the internet discussing some normalization, first of all active_support's transliterate
also this normalization article which suggests

string.mb_chars.normalize(:kd).gsub(/[^\x00-\x7F]/n,'').downcase.to_s

One could discuss if for names of Persons (or authors) some canonical name (first, last) could be stored in addition to the actual name, so that active_record queries would have more hits.

Needs to be far more prominent and consistent.

Hi Seek team,

I've found a weard issue in sample type creation with the v1.4.1 release.

I have a user associated with only one project (as administrator).

When I want to create the sample type, the project is take by default but it seems the default policy is not applicated.

When I want to add a new attribute, nothing appends in the form.

There is a javascript error :

Uncaught TypeError: Cannot set property 'permissions' of undefined
    at Object.applyPolicy (application-89f951cc6bd45245a2edff099d5ab70a6aa46f1fc9eed5be9569b30c9592db28.js:21)
    at Ue.add (new:420)
    at Ue.n [as add] (application-89f951cc6bd45245a2edff099d5ab70a6aa46f1fc9eed5be9569b30c9592db28.js:17)
    at HTMLDocument.<anonymous> (new:442)
    at u (application-89f951cc6bd45245a2edff099d5ab70a6aa46f1fc9eed5be9569b30c9592db28.js:3)
    at Object.fireWith [as resolveWith] (application-89f951cc6bd45245a2edff099d5ab70a6aa46f1fc9eed5be9569b30c9592db28.js:3)
    at Function.ready (application-89f951cc6bd45245a2edff099d5ab70a6aa46f1fc9eed5be9569b30c9592db28.js:3)
    at HTMLDocument.a (application-89f951cc6bd45245a2edff099d5ab70a6aa46f1fc9eed5be9569b30c9592db28.js:2)

When I delete the project from the form, a clone is generated in the list (cf image).

If I choose one of the duplicate project in the list, nothing appends and the field keeps None value.

I've tried with another user having several projects and it works even if there is the duplication project when I remove it from the list.

I hope I'm clear in my explanations :/

Anyway, thanks a lot for your great work.

We have an internal excel sheet to capture information related to samples which map to a study, because we can have multiple assays on the same sample (e.g. DNA/RNA sequencing, proteomics, metabolomics, microscopy, etc..).

In this case a sample has a location (long/lat), environment, country, biosafety level...

It would be nice if we can capture these fields within the SEEK environment so that users can add additional fields to the corresponding studies, which, we think is currently not possible?

Hello again,

as a follow-up to my problems with upgrading (see #7), I ended up reinitializing our (mostly empty anyway) SEEK instance at 1.0.2. As an administrator, I can now add organisms, but I cannot remove them or even find any way to see which organism are already registered. I have noticed that organisms is the only category for which browsing is not possible - is that by design?

Hi Seek team,

I've found an issue on my Seek instance when I want to create a study.

In the form, when I choose my investigation, it's not the list of my project users but it seems to be another users from a different project.

When I save my study (with a wrong user) and I edit it, there is the good list of users in the field.

Thanks a lot

Hi,

I observed a connection problem with Firefox 59.0.1 and SEEK v1.5.2 (and 1.6.0-dev).

After completing the fields, the user is redirected to the required page but is not logged in.

There are no error messages in the console or in the search logs.

It works when the "Remember me" button is checked.

Have you observed this problem?

It seems that Model.all.each is used quite often (29 matches in current master)
Since this retrieves all rows in that table at once and converts that into model objects that need to be kept in memory, this does not scale well
find_each might be a better choice in respect to scale-ability
It is well discussed here: http://guides.rubyonrails.org/active_record_querying.html#retrieving-multiple-objects-in-batches

The check for whether a sample type can be edited or destroyed, according to whether it has associated samples, should be included in the can_edit? and can_destroy? methods

The controller can then also call the generic find_and_authorize_requested_item

In my project, I did the same analysis on different strains.
I created an "Assembly" Investigation, then I need to create one "assembly strainX" study per strain, then one assay per prog used.
This is a lot of similar action, only the strain will change.
I'd like to do it by command line to easily loop on strains.
This can take to form of a web service (or stg else).

Not just "all in the project". Need WP leaders, possibly WPn.

Hello again,

I have a problem when I want to create a new project.

When I click on create, I have an Internal Server Error (500).

In the production.log, I have this message:

ERROR - NoMethodError (undefined method `set_attributes_with_sharing' for nil:NilClass)

I've observed that in the privacy sharing, I've an empty project automatically in the table.

Any ideas?

Cheers,
Cyril

Hi all,

Currently, a user not in any project might be at a loss as to how to join / ask to create a new one.

The current way is to directly contact the seek admin or a project admin by mail, which might not be the most evident way for a new user.

Would it be possible to add a "join" request on each project page? (Basically sending a mail to the project admin for validation).
It might also be good to add a "Create new project" button, leading to a form where the user must fill out the details. The seek admin will then be able to validate/deny the request.

Hi Seek team,

there is a problem when I want to subscribe to any content (assay, data, etc.).

I've a internal error with message :

ActiveModel::ForbiddenAttributesError in SubscriptionsController#create

I tried on your seek-1.5 branch and I've the same issue. Maybe something is broken on my instance?

Best regards,
Cyril

Hello Seek team,

first thank you for this new release and your work. It's a pleasure to see your active development.

Then, on my SEEK instance, I've seen some troubles during permissions (new UI) when I want to add "members of a project" in an assay or a study. It works good for an investigation.

When I valid the form with a project, there is no new entry in the table and I've a javascript error:

Uncaught TypeError: Cannot read property 'selected' of undefined
at Object.addPermission (application-724beab00326adf1fdb0972c7d0c139d.js:301)
at HTMLAnchorElement. (new:476)
at HTMLAnchorElement.dispatch (application-724beab00326adf1fdb0972c7d0c139d.js:60)
at HTMLAnchorElement.g.handle (application-724beab00326adf1fdb0972c7d0c139d.js:59)

It works when I add a project associated with an institution or just an institution.

I work on my own branch (#41) so maybe it's because of that.

Thanks,
Cyril

Hi,

it seems that for a project subscription, I do not receive immediately email after a content creation.

However, I receive a mail when a content was updated.

Do you have the same problem?

So can easily choose

Would love to be able to import my existing ISATabs into SEEK through file upload. Either individual files or zip or tar archives (we use zip).

I promise this is the last one for today. 😄

Is it possible to use e.g. rvm upgrade 2.1.6 2.1.7 instead of rvm install 2.1.7? According to rvm's documentation, that will migrate existing gemsets and environments to the new version, and I actually faced some problems with the environment settings. I guess it perhaps also negates the need to re-install every single gem from scratch with every upgrade?

I would have tested rvm upgrade myself but right now I do not feel like tinkering any further with our instance.

Another follow-up to #7: After I upgraded our instance, I relaxed a bit - and then deleted all of the old ruby versions in rvm that had accumulated. Unfortunately, I had completely forgotten about passenger's integration and thus left Apache with a stale passenger path, which killed access to our instance.
Perhaps a link to http://docs.seek4science.org/tech/install-production.html at the end of each upgrade description is in order?

Hello all,

I've observed a bad issue on my instance.

1. Only the administrator of the instance can modify the default policy for a project. The administrator of the project can't.

2. When we change permissions on investigation / study / assays / etc. , there are no changes after update. I've observed that the behavior is very weard:

• try to update (Bio-imarie project have been added on a previous update but not appears in the table)
  

• and just after, when I try to watch my changes:

Thanks

Ability to give the sample type more info, with a description.
Description should also be indexed for search

Hi again :),

I'm trying all functionnalities and I have questions concerning controlled vocabularies during sample type creation.

It seems that a user can view, edit and remove (if not used in a sample type) controlled vocabs, even if he isn't the creator.

In the same scope, samples created in other projects are visible during my assay creation.

Is it normal?

Thanks a lot for all your fixes ;)

Dear SEEK team,

I just upgraded our SEEK instance from 1.0.2 to the latest commit (9cd2e9f) to take advantage of the previously elusive organism browser. I also upgraded ruby with the experimental rvm upgrade 2.1.7 2.1.8, recompiled and set up the passenger module, and recompiled the assets as described in the upgrading guide.

I then deleted many of the default bacterial strains that SEEK comes with, as well as a duplicate for organism "Human" that I had set up while oblivious to the fact that there was a pre-defined one with the same name. For the two "Humans", I also switched our only project from the deletion candidate to the other one (which was necessary for the "Delete" button to become available). Afterwards, I added some new organisms. The final state looks like this:

Shortly after finishing that work, I received 23 SEEK error notifications that started like this:

A RuntimeError occurred in background at 2016-04-23 15:57:13 UTC :

  Strain#ncbi_uri delegated to organism.ncbi_uri, but organism is nil: #<Strain id: 3, title: "168", organism_id: 548322508, created_at: "2016-03-02 18:23:29", updated_at: "2016-03-02 18:23:29", parent_id: nil, synonym: nil, comment: nil, provider_id: nil, provider_name: nil, is_dummy: false, contributor_type: nil, contributor_id: nil, policy_id: 3, uuid: nil, first_letter: nil>
  /home/seek-op/seek/app/models/strain.rb:33:in `rescue in ncbi_uri'

The first message with the full backtrace is here:
SEEK-backtrace.txt

So far nothing bad has come from that, but you might want to investigate. Let me know if you need more details.