Codelyzr

Analyses any git repo to determine the top 10 hotspots based on how frequently files are changed and how many lines of code they contain. It looks for large files changed frequently. That is what Adam Tornhill explains in his research as the place where developers should focus their attention in order to maximize the effort of paying-off the debt on parts of the software that are more subject to changes.

Getting started

docker run -v <full_path_of_local_git_repo>:/data -it riccardomerlin/codelyzr:latest <start_date> <tab_size>

<start_date>: format YYYY-MM-DD, it will restrict the analysis starting from the specified date. <tab_size>: number, indicates the tab length in spaces used in the code base (typically 2 or 4). The value is used to calculate white-space complexity for Complexity trends.

Results

All results are saved in the hotspots folder in the analyzed repo.

Run Codelyzr from local repo

git clone https://github.com/riccardomerlin/codelyzr.git
cd codelyzr
docker build --tag codelyzr:latest .
docker run -v <full_path_of_local_git_repo>:/data -it codelyzr <start_date> <tab_size>

Exclusions to speed up the analysis

It is possible to exclude some paths from the analysis if not relevant so that the processing will take less time. To specify what paths should be excluded, create the file hotspots/.pathExclusions and add one path to be excluded per line.

Example:

dist/*
coverage/*
*/obj/*

Exclusions for data normalization

To determine the top 10 hotspots, non relevant files must be excluded. The CHANGELOG for example, is definitely frequently changed and could easily become the top hotspot, but it is not something we are interested in analyze. To remove files or paths that do not contribute to the analysis, create the file hotspots/.fileExclusions. This uses regular expressions, you can add one regex per line.

Example:

.*\.csproj.*
.*package(-lock)?\.json.*
.*yarn\.lock.*
.*(app|Web|packages|Local)\.Config.*
.*\.sln.*
.*CHANGELOG\.md.*

Web server

In order to visualize enclosed diagrams for hotspots or complexity trends there is a web server available that runs on nodeJs.

node server.js

It will run a local webserver that respond to http://localhost:9000 To visualize enclosed diagrams go to

http://localhost:9000/<ANALYSIS_FOLDER_PATH>/hotspots.html

To visualize complexity trends of the first 10 hotspots got to

http://localhost:9000/<ANALYSIS_FOLDER_PATH>/complexity-trends/complexity-file-trend.html?file=hotspot[number]

where [number] ranges from 1 to 10.

Extras

Interact with the docker container bash

You can run the container and execute individual commands on the local repo. To access the container form bash use the following:

docker run -v <full-path-of-local-git-repo>:/data --entrypoint=/bin/bash -it codelyzr

See commands.sh for the possible commands you can run.

Credits

This tool has been inspired by Adam Tornhill and uses some of the free tools made available in his book Software Design X-Ray like Code Maat.