Running the module can be done by executing it with simple python code python mooring_dbscan.py.
This executes all the parts of the pipeline.
Individual parts of the pipeline can be executed if so desired.
The preprocessing part of the module reads the data from either a local csv file or a MonetDB database instance. The preprocessing filters the data and calculates the aggregate data to be used in model training.
Running this step individually requires the user to import the DataObject object and executing it as follows
from mooring_dbscan import DataObject
dataobj = DataObject()
dataobj.run()
This will run all the preprocessing steps and the data will be available in dataobj.train
The modeling object takes the training data created by the preprocessing and detects clusters from it. The modeling object then creates polygons out of the cluster points based on the dimensions of ships that form the cluster point. The modeling object also detects quay areas that are formed from multiple individual berths.
Running this step individually requires the user to import the Model object and executing it as follows
from mooring_dbscan import Model
modelobj = Model()
modelobj.run()
The detected berths will be available in model.clustering_results and the detected quays in model.combined_quays
The code includes a visualizer that maps the found polygons to a folium map. The map is saved to a HTML file.
Running this step individually requires the user to import the Visualizer model and executing it as follows
from mooring_dbscan import Visualizer
Visualizer(modelobj)
This will map the berths and quays detected by modelobj and store the results to an HTML file.
The execution of the code can be configured using the config.py file.
Most important configuration variables are:
LOCAL_FILE: the name of the local file. If missing or left blank, the data will be fetched from a database
HTML_FILE: the name of the html file where the results will be saved.