Code for my MTech Project (MTP) on analysis of comments in C/C++ programs. (IIT Kharagpur, Spring 2019)
The aim of this project is to assign a usefulness score to each comment.
This project is written in Python 3.
The following python packages must be installed. (You can do this by pip3 install <package-name>.)
- editdistance
- nltk
The C/C++ code repository on which you want to run comment analysis must have a ProblemDomainConcepts.txt file in the top-level directory of the repository, containing the relevant problem domain words and phrases corresponding to that repository.
For example, for this GenePrediction repo based on computational biology, the concepts file might look somewhat like this:
amino acid
residue
hemoglobin
haemoglobin
taxonomic
beta chain
insulin
peptide
nucleotide
protein
gene
codon
genome
viterbi
genomic background
base pair
Run this command:
python3 analyze_comments.py <path-to-repo>
For example,
python3 analyze_comments.py repos/GenePrediction/
A comments.csv file will be generated in the current directory, containing the extracted comments along with rich metadata information, including the following fields:
- Filename
- Comment text, Start line, End line
- Number of words
- Program domain concepts extracted from a comment
- Problem domain concepts extracted from a comment
- Whether a comment contains one or more of the following kinds of data:
- Copyright/License information
- Build instructions
- Code author related info - name/email/contact
- Date related info - modified on/created on
- TODO information
- Junk (strings of symbols without any alphanumeric data)
- System requirements (OS, GPU, RAM, Cache, server etc)
- Bug/Version related information
TODO