search

nano-bio / IsotopeFitSharp

0 stars 0 forks source link

Calibration.NameList needs to be populated #6

Open johannespostler opened 6 years ago

johannespostler commented 6 years ago

When manuall filling a workspace with data the Calibration.NameList is not populated. I propose to add a custom setter for Cluster.MassOffset und Cluster.Resolution which checks if both values are present and then adds the Cluster ID to Calibration.NameList

enaixtu commented 6 years ago

I have made two changes. First one is, that when loading an IFD file, the NameList will not be populated. The thing is, that the concept of "NameList" is different in IFD file and in the GUI.

The second and very important change is in the way the Clusters dictionary is populated. Now, there is a Workspace.AddCluster(string id, IFData.Cluster cluster) method, that does this, while keeping the Clusters dictionary sorted by value. The type of this dictionary is actually SortedDictionary which allow access by indices. Sortedness and index access is required by numerical functions.

Regarding filling the NameList, when the AddCluster(...) method is called, it checks if the resolution and mass offset values of the supplied cluster are non-zero. If yes, the ID of the cluster is added to the NameList.

Important thing to note, that this is merely the first of the helper functions to provide a convenient way of populating the NameList to the GUI. We will definitely have to cater for the use case of adding a cluster to the NameList much later than when it was added to the Clusters dictionary and also the case of removing a cluster from the NameList. I suggest we discuss more about the possible user interactions with the NameList and continue from there.

enaixtu commented 6 years ago

There is currently a major problem with the NameList. The way is is built does not guarantee the entries are sorted by mass. As a consequence, when building the COMList, MassOffsetListand ResolutionList, the values might become garbage. There are two options:

  1. Either we provide methods for filling (and deleting) entries in the NameList that ensure the sortedness by mass.
  2. Or we sort the generated COMList along with the two other lists after they have been created.

I marginally prefer the first option, because it eliminates the problem at its origin.