Skip to main content

A conservative interactive spell checker for source code.

Project description

scspell is a spell checker for source code. It does not try to be particularly smart–rather, it does the simplest thing that can possibly work:

  1. All alphanumeric strings (strings of letters, numbers, and underscores) are spell-checked tokens.

  2. Each token is split into one or more subtokens. Underscores and digits always divide tokens, and capital letters will begin new subtokens. In other words, some_variable and someVariable will both generate the subtoken list {some, variable}.

  3. All subtokens longer than three characters are matched against a set of dictionaries, and a match failure prompts the user for action. When matching against the included English dictionary, prefix matching is employed; this choice permits the use of truncated words like dict as valid subtokens.

When applied to code written in most popular programming languages while using typical naming conventions, this algorithm will usually catch many errors without an annoying false positive rate.

In an effort to catch more spelling errors, scspell is able to check each file against a set of dictionary words selected specifically for that file. Up to three different sub-dictionaries may be searched for any given file:

  1. A natural language dictionary. (scspell provides an American English dictionary as the default.)

  2. A programming language-specific dictionary, intended to contain oddly-spelled keywords and APIs associated with that language. (scspell provides small default dictionaries for a number of popular programming languages.)

  3. A file-specific dictionary, intended to contain uncommon strings which are not likely to be found in more than a handful of unique files.


To begin the spell checker, run

$ scspell source_file1 source_file2 ...

For each spell check failure, you will see output much like this:

filename.c:27: Unmatched 'someMispeldVaraible' -> {mispeld, varaible}

In other words, the token “someMispeldVaraible” was found on line 27 of filename.c, and it contains subtokens “mispeld” and “varaible” which both failed the spell-checking algorithm. You will be prompted for an action to take:


Skip to the next unmatched token, without taking any action.

(I)gnore all

Skip over this token every time it is encountered, for the remainder of this spell check session.


Enter some text to use as a replacement for this token, and replace only the token at this point in the file.

(R)eplace all

Enter some text to use as a replacement for this token, and replace every occurrence of the token until the end of the current file.

(a)dd to dictionary

Add one or more tokens to one of the dictionaries (see below).

show (c)ontext

Print out some lines of context surrounding the unmatched token.

If you accidentally select a replacement operation, enter an empty string to cancel.

If you select the (a)dd to dictionary option, then you will be prompted with the following options for every subtoken:


Return to the previous menu, without taking any action.


Skip to the next subtoken, without taking any action.

add to (p)rogramming language dictionary

Add this subtoken to the dictionary associated with the programming language of the current file. scspell uses the file extension to determine the language, so you will only see this option for files which have an extension.

add to (f)ile-specific dictionary

Add this subtoken to the dictionary associated with the current file. scspell identifies unique files by scanning for an embedded ID string, so you will only see this option for files which have such an ID. See Creating File IDs for details.

add to (n)atural language dictionary

Add this subtoken to the natural language dictionary.

Creating File IDs

If you would like scspell to be able to uniquely identify a file, thus enabling the creation of a file-specific dictionary, then you must insert a unique ID somewhere in the contents of that file. scspell will scan each file for a string of the following form:

scspell-id: <unique ID>

The unique ID must consist only of letters, numbers, underscores, and dashes. scspell can generate suitable unique ID strings using the --gen-id option:

$ scspell --gen-id
scspell-id: e497803c-523a-11de-ae42-0017f2ee0f37

(Most likely you will want to place a file’s unique ID inside a source code comment.)

Sharing a Dictionary

A team of developers working on the same source tree may wish to share a common dictionary. You can permanently set the location of a shared dictionary by executing

$ scspell --set-dictionary=/path/to/dictionary_file.txt

The dictionary is formatted as a simple newline-separated list of words, so it can easily be managed by a version control system if desired.

The current dictionary can be saved to a file by executing

$ scspell --export-dictionary=/path/to/output_file.txt

You can also override the dictionary location for a single spell check session, by using the --override-dictionary option:

$ scspell --override-dictionary=/path/to/dictionary_file.txt source_file1 ...


If you have setuptools installed, then you can install scspell via:

$ easy_install scspell

Alternatively, download and unpack the source archive, switch to the archive root directory, and run the installation script:

$ python install

On a UNIX-like system, you may need to use sudo if installing to a directory that requires root privileges:

$ sudo python install


scspell is Free Software, licensed under Version 2 of the GNU General Public License; see COPYING.txt for details.

The English dictionary distributed with scspell is derived from the SCOWL word lists . See SCOWL-LICENSE.txt for the myriad licenses that apply to that dictionary.

Bugs, etc.

scspell is hosted on Launchpad; this would be a great place to file bug reports and feature requests or track development via bzr. If that’s not your style, just send an email to Paul Pelzl <pelzlpj at gmail dot com> .

Patches adding the most common keywords/APIs for popular programming languages would be most welcome.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

scspell-0.1.0.tar.gz (276.2 kB view hashes)

Uploaded source

Built Distribution

scspell-0.1.0.win32.exe (320.1 kB view hashes)

Uploaded any

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page