A shell-based tool that resorts to Topic Modelling and Clustering techniques in order to identify microservices from projects built with monolithic architectures. Currently only Java projects are supported given the structural and lexical information being extracted by the Java Parser. However the core functionality is language agnostic and if provided with structural and lexical data should work equally.
- Clone the repo
- Give execution permissions to setup script
chmod u+x setup.sh - Execute
setup.sh
All the interaction with the tool is done through the python script main.py located in /app/ directory.
main.py [-h] --project PROJECT [--stop_words STOP_WORDS]
[--k_topics K_TOPICS] [--resolution RESOLUTION] [--metrics]
[--draw] [--lda-plotting]
optional arguments:
-h, --help show this help message and exit
--project PROJECT, -p PROJECT
Project path
--stop_words STOP_WORDS, -s STOP_WORDS
Path to stopwords file
--k_topics K_TOPICS, -k K_TOPICS
Number of topics for given project
--resolution RESOLUTION, -r RESOLUTION
Resolution number parameter in Louvain community
detection. A range between 0.3 and 1 is advised. A
smaller resolution will identify smaller communities (smaller microservices)
and vice versa. By default the whole range is tested
and communities for each community saved.
--metrics, -m Execute metrics for a given project name after normal
parsing and execution (At the current time it does NOT work independently
from the identification process)
--draw, -d Enable plotting of graphs
--lda-plotting, -l Enable plotting of LDA topics which might help identifying the appropriate number of topics
Automated K-Topics and Resolution
python3 main.py -p <path_to_project>
Automated K-Topics and Resolution with Metric execution
python3 main.py -p <path_to_project> -m
Custom K-Topics and Resolution
python3 main.py -p <path_to_project> -k <num_topics> -r <resolution>
Custom list of stop words
python3 main.py -p <path_to_project> -s <path_to_stop_words_file>