• Keine Ergebnisse gefunden

4.7 Platform usability

5.1.3 View concept

Finally, theView entity is the last, but the most essential cornerstone of the basic BioIMAX data model. The idea of the View concept is, that any analysis result can be considered as a special view or perspective on an original ImageStack imported into the system and is comparable to the ImageStack transformation principle mentioned above (see Figure 5.1(a,b) for a graphical illustration of both concepts). One of the major objectives of theView entity is to connect the analysis data category to the meta data category in a distinct way, thereby avoiding unnecessary redundancies and complex links within the database. A View instance collects and aggregates all analysis results and their data structures regarding a particular ImageStack transformation. ImageStack transformations or projections inBioIMAX are rep-resented by singleView instances, each of them characterized by a specificView type, which match the type of the respective transformation. Each View instance is clearly associated to an ImageStack instance by using the unique ID of the respective ImageStack instance as foreign key in the View instance. All results of one transformation regarding a particular ImageStack are associated to a respective View instance using a unique View ID, inden-dently of how the results are stored on the server or how they are distributed. Thus, all data of the analysis category is connected to the respective sourceImageStack entities through theView entity.

The handling of raw images from an original ImageStack plays a decisive role in the View concept. As mentioned before, original images are also associated to the analysis data cat-egory. Since these images are not a result of a specific transformation step, there exists no direct View type equivalent to a transformation. Thus, an additional View type called original view has been defined, which allows the assignment of raw images of an original ImageStack to a particular View instance. For each newly imported ImageStack the raw

38 Chapter 5 Architecture

Figure 5.2:The basicBioIMAX data model. Illustration of the core entities forming the basic conceptual scheme for the relational database model and their functional roles within theBioIMAXsystem. TheBioIMAX data model consists of four basic entities: User,Project,ImageStack, andView. A registeredBioIMAX user uploads a multivariate ImageStack to theBioIMAXdatabase. Afterwards he/she is uniquely associated to the newly integrated image data as its owner and has the exclusive rights to edit the ImageStack, to control its availability for other users, or to delete the ImageStack. All data generated based on an ImageStack, either automatically or via analysis tools withinBioIMAX, is represented by respective Views, which describe a specific perspective on the original ImageStack. The major objective of the View concept is to connect the entities of the meta data category (User,Project,ImageStack) to the analysis data category. In order to share data with a selected group of individuals, a user can create and manage a project instance with which data in the form of Views as well as invited users can be brought together. Using projects as communication channel, collaborating users get access to data from other users, can generate new project related content, and discuss relevant datasets, e.g., regarding specific biological analysis questions.

5.1 Database design 39

images are stored on the server without any previous processing of the image data and are associated to the respective View instance via a unique View ID. Together with the type original view all data of the analysis data category can completely be connected to the meta data category of theBioIMAX data model.

In sum, the BioIMAX data model has a simple but powerful structure, since it is based on just a few essential entities, which allows a fast and transparent access to all data stored in the database. In particular, the View concept plays a key role in this data model. The essential benefits of the View concept with respect to the entire data model are pointed out in the following.

• First, the View concept allows an easy and flexible integration of novel algorithms, tools, or interfaces, which potentially generate new types of data. These new data types are usually stored using appropriate data structures, e.g., database tables or even more complex data structures, that are not yet included in the BioIMAX data model.

However, the BioIMAX data model can be extended with these new data structures without changing the rest of the data model, which applies in particular to the meta data category. The developer only has to define a new View type, which should describe and identify the newly embedded datasets within the data model. In this way, each dataset is connected to the basic entities of the meta data category. The same procedure also refers to modifications and updates of existing tools or functionalities, which can be performed without adapting the entire data model.

• Second, the way the system deals with complex and even distributed analysis data structures is simplified considerably. This refers to data handling aspects such as searching for specific data, displaying the search results, and assigning analysis results to particular projects. In particular, the View concept reduces the amount of data transferred from theBioIMAXmain application to dedicated analysis tools (see Chapter 5.2.4 for detailed explanation of the BioIMAX interface concept). Thus, working with abstract Views withinBioIMAXrather results in a faster and a more performant system than by using and providing the total number of corresponding analysis datasets at any time the system requests the data.

• Finally, the association of additional meta information to specific analysis data is easily possible via View instances in a central manner independently of the type of meta information.

In Figure 5.2 the conceptional data model based on the essential entities User, Project, ImageStack, andView and their functional roles within theBioIMAX system are illustrated, whereas Figure 5.3 depicts the detailed overview of the entire data model showing all entities and their interconnectivity.

40 Chapter 5 Architecture

User

+ id+ name+ email+ login+ password+ question+ answer+ registered+ last_login

Private_messages

+ id+ receiver_id+ sender_id+ date+ subject+ message+ read+ response+ responded+ receiver_storing+ sender_storing ImageStack

+ id+ user_id+ name+ acquisition_date+ num_images+ width+ height+ meta_info Project

+ id+ creator_id+ name+ date+ meta_info

Project_members

+ id+ project_id+ user_id+ write_access Project_views

+ id+ project_id+ view_id

Result

+ id+ view_id+ name+ meta_info Antidict

+ id+ antibody_id+ specificity+ clone+ hostspecies+ reactivespecies+ supplier

Label_chat

Antidict_selection

+ id+ user_id+ selection_name+ antidict_xml Labels

+ id+ user_id+ view_id+ image_id+ type_id+ chat_existent+ category+ value+ color+ description+ x_pos+ y_pos+ geometry_type+ geometry_parameters+ size+ meta_info Label_types

+ id+ parent_id+ name

Project_invitations

+ id+ project_id+ user_id

Figure5.3:CompleteoverviewoftheBioIMAXdatamodel.ThisschematicdiagramdepictsadetailedoverviewaboutofthecompleteBioIMAX

datamodelincludingallentitiesandtheirinterconnections.Eachentitycontainsanumberofattributesdescribingtheirproperties.Keyattributesare

italicized,wherebytheprimarykeysareadditionallywritteninbold.TheentitiesUser,Project,ImageStack,andViewaretheessentialcornerstones

ofthedatamodelandarehighlightedwitharedcoloredheader.Relationshipsbetweenentitiesarerepresentedbylinesconnectingtheentities.