kodespel

spellchecker for source code


Keywords
spellcheck, developer-tools, source-code
License
BSD-3-Clause
Install
pip install kodespel==0.1.1

Documentation

kodespel

kodespel is a spellchecker for source code. It is implemented as a small Python script with all the real work done in a library (package kodespel).

kodespel's nifty trick is that it knows how to split common programming identifiers like 'getAllStuff' or 'DoThingsNow' or 'num_objects' or 'HTTPResponse' into words, feed those to ispell, and interpret ispell's output.

Requirements & installation

kodespel requires Python 3.6+ and ispell. To install ispell, use your OS-specific package manager (e.g. apt, dnf, brew, ...).

To install kodespel itself, use pip:

pip install --user kodespel

(Or install it in a virtualenv if you prefer.)

Usage

Basic usage is to run kodespel on one or more individual files or directories:

kodespel foo.py main.go README.md

kodespel uses a collection of dictionaries to spellcheck each file. It always uses the base dictionary, which is a set of words common in source code across languages and platforms. Additionally, there is a language-specific dictionary for each language the kodespel knows about. Language-specific dictionaries are automatically chosen for you.

In this example, kodespell will spellcheck each file with:

  • foo.py: dictionaries base and python
  • main.go: dictionaries base and go
  • README.md: dictionary base only (no language dictionary for Markdown)

If run on a directory, kodespel will recurse into that directory and spellcheck every file that it recognizes:

kodespel src/

will search for *.py, *.c, *.h, and any other extension that kodespel has built-in support for. (Currently: Python, Perl, Go, C, C++, and Java.) Unsupported files are ignored, but if you pass those filenames explicitly, they will be checked.

kodespel ships with several other common dictionaries. For example, if the program you are spellchecking uses a lot of Unix system calls, you would add the unix dictionary:

kodespel -d unix foo.py main.go README.md

The -d option applies to every file being checked.

To see the list of all builtin dictionaries, run

kodespel --list-dicts

Finally, you can create your own dictionaries, and use as many of them as you like. A dictionary is a plain text file with one word per line:

$ cat myproject.dict
nargs
args

You can specify your person dictionaries with -d, just like kodespel's builtin dictionaries:

kodespel -d unix -d myproject.dict foo.py ...

See also

A tool with similar goals but a different implementation is codespell.

The main advantage of codespell is that it seems to have many fewer false positives.

The main advantage of kodespel is that it checks identifiers, not just comments and strings, so can find a lot more errors. And more false positives too, unfortunately.