populse / populse_mia

Multiparametric Image Analysis
Other
10 stars 9 forks source link

[Input_filter node] user friendly access to the filter in the Input_Filter brick #248

Open servoz opened 2 years ago

servoz commented 2 years ago

Currently, to define the filter, used in this brick, to perform the output completion, it is necessary to right click on the brick and click on Open filter. It would be nice to have a more obvious way for the user to do this.

*For more details, please see #243

denisri commented 2 years ago

If I may add a comment on the input filter node: this node is not operated through "standrad" node/pipeline handling (using the traits API for parameters, and the controller edition GUI). Thus it needs a special GUI to define filters, and its API is also different: the node contains a filter variable (of type Filter) which is not a trait, thus does not appear on a pipeline signature, cannot be exported to a parent pipeline, and is not saved with the pipeline. This contradicts most of the pipeline API genericity concepts. This is why a specific GUI is needed, actually. To set the filter, the user must have knowledge of it, and, if it is integrated in an existing pipeline, the user has to display the pipeline structure, select this node, and perform a specific action (the "open filter" menu up to now). Thus this node may be used "on the fly" by a "user" who builds his own pipeline with no intention to reuse it later, but it cannot be used in reusable pipelines. Moreover a programmer who wants to write his pipeline and use it in a scriptable way, will have difficulties to do it. By the way this leads to another discussion area, which is that MIA processes need the Mia infrastructure to work (database, pipeline manager, project and settings) which have not be designed to be used as a library (most of the code is in "user_interface" and need the GUI).

servoz commented 2 years ago

Your remark illustrates perfectly the difficulty to develop a common project when we don't have time to meet to define common bases, commonly understood goals, concepts, etc. We all have good wills, we try to do our best to solve the problems that arise through the issues in github, but it must be said that we develop things each of us in our own corner. It is not certain that your evidence is the same as ours, especially if we do not talk about it. I must confess that this brick is very old (very close to the beginning of Mia, because if I remember well it corresponded to the need to be able to automatically filter the entries of a brick from the database) and dates from a time when our vision of things was quite different from the one of today. I am ready to make the necessary modifications for this brick but I fear that we don't really have the same expectations or needs, simply because we may never have talked about it.