Class Dependency Graph

The Class Dependency Graph takes a plain text file, parses it into nodes, and gives the user the ability to select each node and identify the nodes linked to the node they selected.

Getting Started

Clone the repository
git clone git@github.com:apipkin/DependencyGraph.git

Change into the directory
cd DependencyGraph

Install important packages
npm install

Start the server
npm start

Testing

Unit Tests

test npm test
test coverage npm run test:cover
test coverage in browser test:cover:browser

Test Plan

Attempt with a simple file and select the node. Upload:

ClassA

Select: ClassA
Expected: ClassA
Result: ClassA

Attempt with a more complex file and select ClassA Upload:

ClassA ClassC ClassE
ClassB ClassD ClassG
ClassE ClassB ClassF ClassH
ClassI ClassC

Select: ClassA
Expected: ClassA ClassE ClassH ClassF ClassB ClassG ClassD ClassC
Result: ClassA ClassE ClassH ClassF ClassB ClassG ClassD ClassC

Attempt with a more complex file and select ClassC Upload:

ClassA ClassC ClassE
ClassB ClassD ClassG
ClassE ClassB ClassF ClassH
ClassI ClassC

Select: ClassC
Expected: ClassC
Result: ClassC

Attempt with a cyclic graph and select ClassA Upload:

ClassA ClassB
ClassB ClassA

Select: ClassA
Expected: Error: CircularDependency
Result: Error: CircularDependency

Attempt an upload of a non .txt file Upload: (image file)

Expected: Error: Invalid File
Result: Error: Invalid File

Attempt to select an invalid option Upload:

ClassA ClassC ClassE
ClassB ClassD ClassG
ClassE ClassB ClassF ClassH
ClassI ClassC

Select: ClassN (modify list with web inspector)
Expected: Error: NodeRetrievalError
Result: Error: NodeRetrievalError

UML Diagram

Lessons learned

One of the main lessons learned, and the longest portion of research for very little gain was uploading the asynchronous file to the server and parsing it directly. The biggest portion of this had to do with understanding how hapi loaded files with streams. This is done through a configuration object in the route.

Working with a one way dependency graph worked really well with lessons learned in previous tree and list structures. Graphs were a bit of a mystery going into this lesson, but after reading how they are able to be broken down with verticies and edges it begin to make more sense. Verticies become nodes, and edges become a list of nodes within the node. This gives the abilty to traverse the graph recursively.

Possible Improvements

One possible improvement would be to graphically display the graph that is presented rather than a string. This graph would not have to an image file, but rather an ASCII tree.

As with other projects, custom dialog boxes would be enhance the UI for application feedback to the user.