This project was created as part of the course TDT4250 Advanced Software Design at NTNU. The project uses a movie dataset, and can be used to edit and analyse the data, as well as showing connections between different kinds of data.
This project uses IMDb's datasets found at https://www.imdb.com/interfaces/. From these, we have generated some subsets using Python, which applies a region filter - showing only titles from the US, as these are more likely to be easily readable. An exception to this filtering is tvEpisode titles; these can be from any region, but must be connected to a tvSeries originating in the US. The remaining subsets are filtered by what is referenced from the region-filtered titles.
We recommend using these generated subsets, which can be downloaded from https://drive.google.com/open?id=1JdSAdLa_Nm8XGovUJyuULpH3YzL77Azv. The reason for this is that the main IMDb datsets use a different naming scheme, and the naming of the files in dataset.zip is already correct. In addition, we've also added a prepared XMI file - Imdb.imdb, as parsing the datasets and serializing to XMI uses a massive amount of memory, and can crash on some computers. If you want to run generate these subsets yourself, you should add an argument to JVM to increase the maximum available memory to 4.5-5 GB of RAM.
- Sirius: https://www.eclipse.org/sirius/download.html (follow the marketplace guidelines)
- Eclipse with Modeling tools: https://www.eclipse.org/downloads/ (select Eclipse Modeling tools during installation)
- Acceleo: https://marketplace.eclipse.org/content/acceleo
You can import the project by opening a new Eclipse Workspace - preferably located in the root folder of a clone of this repository, clicking Open Projects from File System, and then choosing the mentioned root folder.
First, you can choose to either use our subsets, or use the datasets available from IMDb. However, when downloading the IMDb datasets, they will need to be placed in /imdb/src/imdb/dataset/, and renamed to match the following:
name.basics.tsv->persons.tsvtitle.basics.tsv->titles.tsvtitle.ratings.tsv->ratings.tsvtitle.principals.tsv->involvements.tsvtitle.episode.tsv->episodesDetails.tsv
When these are placed correctly, you can run DatasetDeserializer.java. This will parse the .tsv files and serialize them to a single XMI file. Currently, the deserializer is set to randomly choose 10 tvSeries, 5 movie, 5 short and 5 video titles, and all the information, episodes and people that are connected to these. These numbers can be changed in the deserialize() function. We recommend not going above 50 of each, as many computers will struggle to show large amounts of data in the Sirius viewpoints. Remember that you might need to increase the amount of RAM available, as mentioned in the beginning. We have not attempted to run this on the full datasets.
After being run, the XMI will be named Imdb.imdb and placed in /imdb/src/imdb/dataset/. Move this file to /imdb.sirius.editor.model/.
If you're using our prepared XMI, move Imdb.imdb from the ZIP file to /imdb.sirius.editor.model/.
In both the above cases, you will have to overwrite the test XMI with the same name.
Now that the XMI dataset has been prepared, you are ready to show the data in the Sirius viewpoints. To do this, click
Run -> Run Configurations -> Eclipse Application -> Run.
When the runtime workspace has opened, navigate through the following:
Import Projects -> General -> Existing Projects into Workspace -> Browse -> *Root folder of the project directory*
Select imdb.odesign and imdb.sirius.editor.model and click Finish.
To create the first main IMDb diagram, expand imdb.sirius.editor.model -> Imdb.imdb -> Imdb, and right-click
Imdb -> New Representation -> Imdb Diagram -> Finish -> OK.
The diagram should now open. From here, you can navigate to all the other different diagrams, trees and tables that are available. See the flowchart at the bottom for the navigation options.
Sometimes, the TV Series Tree won't render all TV series; we believe this is due to the viewpoint rendering quicker than the Java services code can run. If the tree is empty, open odesign.odesign, and select
IMDb Viewpoint -> TV Series Tree -> Imdb -> TVseries branch -> Season branch.
In the Properties window, make some negligible changes to Semantic Candidates Expression - like adding a space after the closing parenthesis in [self.getSeasons()/], and save. Go back to the TV Series Tree, and collapse and re-expand the IMDb node; it should now properly render TV series with seasons and episodes.
Furthermore, some TV series can have thousands of episodes causing the TV Series Tree to take quite a long time to load.
This is the model used in the project:
In the IMDb Diagram, you can find tools for creating new objects. The tools are displayed in the Palette view on the right side of the window. Choose an object creation tool, click on an empty space in the IMDb Diagram and fill in the fields in the dialog box. It is also possible to create an object by hovering over an empty space in the IMDb Diagram, and clicking one of the diamonds that appear. Objects' properties can be edited in the Properties view.
The user can navigate through the different representations when using the Sirius editor in the runtime Eclipse. The boxes represent the different representations, and the small boxes inside represent the different objects which can be clicked in order to navigate. Depending on what kind of representation is open, the user can right-click these object boxes and choose Navigation -> *Chosen representation*, or right-click and choose New/Open -> *Chosen representation*, or right-click the background and choose New/Open -> *Chosen representation*.
When inside a node diagram, the user can press the 
button to display a dropdown of the filters available in that diagram. Three different filters are available in total:
- Filtering titles by title type
- Searching for title names
- Searching for person names