Hi Nate,

 

thanks for your reply. My solution will break down in such a case, indeed. So 
here is what I need/have:

 

I have a tool which allows the user to select a specific workflow and 
appropriate input data from his/her

history. On execution a script is used to do several preprocessing steps and at 
last the selected workflow

will be executed via the API functions (in batch mode if the user selected more 
than one input (pair).

As far as I have seen from the documentation one needs to specifiy a history 
name or id which will hold

the output data. Logically the output should be submitted to the users current 
history which serves as the

input source. That?s why I need the id of the users current history.

 

Of course, workflows can be executed directly within Galaxy but we need to 
perform these preprocessing

steps like copying files and, more important, we have paired end data which 
isn?t that good supported for

batch processing yet.

 

Thanks for your help!

 

Cheers,

Sascha

 

Original Message

processed by CONSOLIDATE

Subject:

Re: [galaxy-dev] Some questions about histories

Sent:

Mittwoch, 15. August 2012 17:07

From:

Nate Coraor (n...@bx.psu.edu)


On Aug 15, 2012, at 10:08 AM, Sascha Kastens wrote:

> Hi again!
> I was able to implement a quite dirty solution:
>  
> first I edited /lib/galaxy/model/mapping.py and extended
>  
> assign_mapper( context, User, User.table,
>     properties=dict( histories=relation( History, backref="user",
>                                          
> order_by=desc(History.table.c.update_time) ),
>                      active_histories=relation( History, primaryjoin=( ( 
> History.table.c.user_id == User.table.c.id ) & ( not_( 
> History.table.c.deleted ) ) ), order_by=desc( History.table.c.update_time ) ),
>                      galaxy_sessions=relation( GalaxySession, order_by=desc( 
> GalaxySession.table.c.update_time ) ),
>                      stored_workflow_menu_entries=relation( 
> StoredWorkflowMenuEntry, backref="user",
>                                                             cascade="all, 
> delete-orphan",
>                                                             
> collection_class=ordering_list( ’order_index’ ) ),
>                      _preferences=relation( UserPreference, backref="user", 
> collection_class=attribute_mapped_collection(’name’)),
> #                     addresses=relation( UserAddress,
> #                                         primaryjoin=( User.table.c.id == 
> UserAddress.table.c.user_id ) )
>                      values=relation( FormValues,
>                                       primaryjoin=( 
> User.table.c.form_values_id == FormValues.table.c.id ) ),
>                      api_keys=relation( APIKeys, backref="user", 
> order_by=desc( APIKeys.table.c.create_time ) ),
>                      ) )
>  
> by this:
> currentHistory=relation( History, primaryjoin=( ( History.table.c.user_id == 
> User.table.c.id ) & ( not_( History.table.c.deleted ) ) & 
> (GalaxySession.table.c.current_history_id == History.table.c.id) ), 
> order_by=desc( GalaxySession.table.c.update_time ) )
>  
> As far as I am correct, this selects all current histories of all sessions 
> per user. These histories get sorted by the latest session change
> which should always lead to the current history as the first entry.
>  
> Inside my tool I use this to get the current history id and hash it to an id 
> which can be used for API calls:
>  
> ${__app__.security.encode_id( ’%s’ % 
> $__app__.model.User.get($__user_id__).currentHistory[0].id )}
>  
> Until now everything seems to work as it should.

Hi Sascha,

This will break down if you have a user with more than one active sessions (in 
two browsers, for example).  The most recent session and its current history 
will always be used, rather than the one that the job was initiated from.

Jobs are intentionally decoupled from histories since histories are not the 
only context in which a job can run.  There may be accurate ways of getting to 
the parent history at the job context, but it?d help to know what you?re trying 
to do.

Thanks,
--nate

>  
> Finally I have another question about the histories and API. Referring to 
> this thread: 
> http://gmod.827538.n3.nabble.com/How-to-use-scripts-api-update-py-td3657803.html#a3700436
>  
> Are you working on this? Updating history items via the API would be really 
> helpful for my usecase. If not, can you give me some hints where I can have a 
> look at to enable it by my own?

No, nothing?s been written for this yet.  Have a look at 
lib/galaxy/web/api/history_contents.py, and you may find the update() method in 
other API controllers to be a useful template.  One that is fairly well sorted 
is the quota API (quotas.py).

--nate

>  
> Thanks a lot!
>  
> Cheers,
> Sascha
>  
> Original Message
> processed by CONSOLIDATE
> Subject:
> [galaxy-dev] Some questions about histories
> Sent:
> Dienstag, 14. August 2012 12:10
> From:
> Sascha Kastens (s.kast...@gatc-biotech.com)
> Hi all!
>  
> I have two questions about histories in Galaxy:
>  
> 1. Is it possible to mark/set a history as published via the API?
> 2. How can I get the ID of the current users history via a tool? Something 
> like
>     ${__app__.model.User.get($__user_id__).histories[0].id} ?!
>  
>  
> Thanks in advance!
>  
> Cheers,
> Sascha
> ___________________________________________________________
> Please keep all replies on the list by using "reply all"
> in your mail client.  To manage your subscriptions to this
> and other Galaxy lists, please use the interface at:
>
>  http://lists.bx.psu.edu/

___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:

  http://lists.bx.psu.edu/

Reply via email to