This project is an assignment project for new potential developers of the connectivity registry team. The goal is to get to know you better as a developer, how you write your code and what your approach to writing algoritms is. Additionally, it is a good check for you whether you would like to work in our team since it is a good reflection of the daily work in the connectivity registry team.
What we value:
- Quality over quantity
- DRY, SOLID, readable code
- Optimized algorithms
- We love tests, but for this assignment time is short, pseudocode and test cases will suffice.
Some tips:
- Pseudocode is also acceptable and expected given the amount of work.
- We love tests, but for this assignment there's simply no time, please also write your test cases in pseudocode
- The assignments are consecutive, a good solution for the first assignment will help with the second assignment and so on
- Try to find a solution for all three assignments even if it is not perfect
Please request an invite to this repo so that you can push branches. Then, when you are ready to start, create and instantly push a new branch. Please make a fork, and then create a pull request when you are done. There are three consecutive assignments, try to timebox each assignment to half an hour.
NOTE: The assignment template is given in Java, but the following languages are also allowed:
- Python
- C#
- Kotlin
If you want to use a different language, please contact us to verify that that is also an acceptable language.
Good luck!
Connectivity is another word for how elements are connected. In the electricity and gas net in the Netherlands, this plays a vital role for:
- Crisis management
- Calculating load on the net
- Planning/modeling future net
The Alliander net is represented by nodes in several layers. For this exercise we will be simplifying the model to two layers:
The connectivity is represented as a graph with nodes and relationships.
Nodes have:
- an ID which is represented as a Long
- 0...N Relationships of any type (label)
- a label indicating what kind of node it is:
- If it is a topology node, it will always have the label "topology"
- If it is an asset node, it will have one of the following labels: "veld", "rail", "kabel", "transformator"
A relationship describes the relation between nodes, it must have at least the following properties:
- an ID which is represented as a Long
- a label, indicating the type of relationship, between topology nodes it is: "CONNECTS_TOPOLOGY", between a topology node and an asset node it is "DESCRIBES_ASSET".
- a way of telling which nodes are connected by the relationship, a relationship never has more than 2 connections.
Assignment 1: Create a data structure that can represent the connectivity as pictured above.
The requirements of a data structure describing this graph are:
- that you can efficiently traverse over the graph i.e. move from 1 topology node to another
- that given a layer name (topology, asset) you can efficiently find data on that layer
- that you can efficiently move between layers i.e. move from 1 asset node to a topology node or the other way around
- that the data structure does not cost a lot to keep in memory
(Optional) Try to create a data structure that will still be applicable when graph contains an arbitrary number of layers
if you cannot find an efficient data structure then for the following assignments you can use a simple data structure of a list of nodes and a list of relationships
To fulfill the wishes of each of our stakeholders, we need to be capable of offering different views on the topology. We do this by densifying the topology as shown in the picture below:
We call this a recipe, as there are many ways you can perform a densify.
The recipe for this densify we call a "ContainerRecipe".
Assignment 2: Create a solution that can perform a ContainerRecipe that will densify a collection of nodes to a container
- The input for the densify method is in the format of the data structure defined in assignment 1
- The output should be compatible with the data structure as defined in assignment 1.
- this should include a "container" node
- this should include the asset nodes pointing towards the container node.
- The container node should have the label "container"
- The container node should have outgoing relationships
- The outgoing relationship should have the same ID as the last relationship of the original node (indicated in red in the images)
- The original nodes and relations, which are densified after this recipe, are not part of the output.
- Only nodes from which the asset nodes meets a certain criteria, should be densified.
Another form of densification is the Split Container recipe. In this recipe, containers are made, until there is a split in the graph.
If this split is reached, stop the container, create a single topology node connected to its assets, and the start 2...N containers after this split node.
Assignment 3: Create a solution that can perform a SplitContainerRecipe that will densify a collection of nodes to a split container:
- The input for the densify method is in the format of the data structure defined in assignment 1
- The output should be compatible with the data structure as defined in assignment 1.
- this should include a "split" node for each split in the input
- this should include the asset nodes pointing towards the split nodes.
- this should include a "container" node for each chain of consecutive topology nodes that does not split
- this should include the asset nodes pointing towards the container nodes.
- (hint) If there is no split in the graph, it behaves the same as a normal container
- The split node is always one topology node with its original connected asset nodes, it should have the same ID's as the input graph
- The split node has the label "split"
- The relations between the split node on the topology layer, and the container nodes on the topology layer, have the same ID as the relations from the input graph
- Outgoing relations are handled the same as for the Container Recipe
- The original nodes and relations, which are densified after this recipe, are not part of the output.
- Only nodes from which the asset nodes meets a certain criteria, should be densified.