DRY

This tool helps you abstract away common or repeated chunks of code by finding similar pieces. It allows you to avoid repeating yourself with less effort on your side.

The tool makes use of Semantic's json output. (Because as of now, Semantic isn't published to hackage yet).

It supports all the languages that Semantic supports

How to Build AST in JSON

Before one can run DRY, it's required to first have an AST to analyze. DRY expects one in particular json format, the one that Semantic outputs.

1. Pull Semantic's container.

2. Run it, passing a command like this:

docker run -v /root/path/to/a/project:/r --entrypoint sh semantic -c "/usr/local/bin/semantic -- parse --json \`find /r/internal/path -name *.js\`" | tail -n 1 > /path/to/output.json

Where:

• /root/path/to/a/project is the root path to the directory of the project one wants to analyze.
• /r is its root path within Docker's container.
• find /r/internal/path -name *.js is the command to find all files that have extension ".js" and reside in /root/path/to/a/project/internal/path. Notice that the command is wrapped in escaped backticks. Sematic accepts a list of filepaths to be parsed as an argument. find returns them, and backticks are required in order to pass find's output as an argument to semantic. Finally, the backticks have to be escaped in order to make the command run within Docker's container (as opposed to the shell where the whole docker run ... command is invoked).

Alternatively, one can provide the root path to a source file to be analyzed, or a number of them. find is only used to enumerate them automatically.

How to Build DRY

1. git clone https://github.com/ch3rn0v/dry.git
2. cd dry
3. stack build

How to Execute DRY

stack exec dry-exe <path_to_the_json_file> <path_to_the_output_csv_file>

Where:

• <path_to_the_json_file> is an absolute path in double quotes
• <path_to_the_output_csv_file> is an absolute path in double quotes

Example:

stack exec dry-exe "/Users/**/dev/parsed.json" "/Users/**/dev/analysis/results.csv"

Instead of stack build followed by stack exec you can use stack run, like this:

stack run dry-exe "/Users/**/dev/parsed.json" "/Users/**/dev/analysis/results.csv"

Example of DRY Output's Visualization

Source code for visualizations

Acknowledgements

• This tool wouldn't have been possible without awesome Haskell community, @haskellru in particular.
• Semantic's staff members helped a lot to run their tool.
• And a great thank you to @python273 who sped up the process by helping with Docker.

You can find the list of libraries used in package.yaml