Grades student answers based on a single or multiple regular expressions.

Regexmatch usage examples

This file contains some example regular expressions which can be used within Regexmatch

Some basic examples

Here are a few examples for simple regular expressions with only the default options enabled.

test//

Regexmatch Syntax

Regular expressions in Regexmatch consist of the regex, options and keys:

regex/OPTIONS/
separator=,
comment=text

The regex uses the default syntax of regular expressions in PHP (without the requirement of a delimiter or the ability to specify modifiers). The internet provides vast amounts of information on how to write regular expression (not specific to Regexmatch):

• A very good (but technical) explanation for different regex syntax can be found here.
• Test regexes directly in the browser here (Select PCRE2 flavor).

The OPTIONS described in Options.
The following is a valid regular expressions with no options changed that match abc:

abc//

The following is a valid regular expressions that matches abc and has some options set:

abc/I/

Additionally, all new lines and spaces before the options will be ignored. This means the following regular expressions are the same as the previous one:

abc
/I/

abc     /I/

abc

/I/

Keys

Regexmatch supports the keys comment= and separator=.

• comment= is a text field only visible inside the question edit form and has no other use.
• separator= is a field for the separator the student has to enter between his answers if the Match Any Order option (O) is enabled.

Options

All Options can be activated by a single capital letter and disabled by its small letter counterpart. Some options are enabled by default. These are called default options.

I: Ignore Case

This option makes the regular expression ignore case. For example the regular expression abc/I/ will match abc, Abc, ABC, aBc, etc. This option acts exactly as the PCRE option IGNORE_CASE.

D: Dot All

All Dots (.) in the regular expression will also match new lines. This option acts exactly as the PCRE option DOT_ALL.

P: Pipes and Semicolons

This is a shell specific option. All semicolons ; and escaped pipes \| will be replaced with ([ \t]*[;\n][ \t]*) and ([ \t]*\|[ \t]*) respectively. Thereby infinite spaces are allowed around these and the semicolon will also match a new line.

Examples:

The regular expression tee /P/ will match any of the following example answers

• cat test.txt|tee
• cat test.txt | tee
• cat test.txt | tee

To note is here, that any spaces in front and after the pipe in the regex must also be contained in the answer: The regular expression tee /P/ will match

• cat test.txt | tee
• cat test.txt | tee

but it will not match

• cat test.txt|tee.

The semicolon has the same properties, if this option is enabled. Additionally, the semicolon will also be matched by a new line´. Thus, the regular expression cat test.txt;tee /P/ will match any of the following example answers:

• cat test.txt;tee
• cat test.txt ; tee

and it will also match the following example answer:

cat test.txt
tee

R: Redirects

This is a shell specific option. All redirections (<,>,<<,>>) will be replaced for example with ([ \t]*<[ \t]*). If enabled redirections cannot be used in other regex-functions (eg.: lookbehind (?<=...)). Note: Any spaces in front and after the redirect inside the regex, must also be contained in the answer.

Examples:

The regular expression cat test.txt>2 /R/ will match any of the following example answers

• cat test.txt>2
• cat test.txt > 2
• cat test.txt > 2

Similar to the pipe option it is important to remember that any spaces in front and after the redirect in the regex must also be contained in the answer: The regular expression cat test.txt > tee /R/ will match

• cat test.txt > tee
• cat test.txt > tee

but it will not match

• cat test.txt>tee.

O: Match Any Order

This option is experimental and subject to change. If this option is enabled the regex must consist of multiple regexes. The answers (separated by the separator=) must match any of the regexes, but order is not important. Each regex can only be matched by a single answer. Wrong, too many or too few answers results in a point deduction evaluation.

Evaluation

The evaluation of the answer is calculated based on an internal point system. The maximum amount of points (maxPoints) is the same as the amount of regexes. The answer will be rated with points using the following rules:

• The rating starts with the value of maxPoints.
• If too little answers are given, a point is deducted for each missing answer .
• If too many answers are given, a point is deducted for each answer which is too much.
• If an answer is wrong (that means it does not match any regex), a point is deducted.
• A regex can only be matched once and an answer can only match a single regex
• An answer that is too much is not counted as a wrong answer.

The points are then converted to a percent based evaluation using rating/maxPoints (1.0 == 100%). This fraction is then used to calculate the actual points for the given answer (actualPoints*(rating/maxPoints)).

Examples:

The regular expression

cat dog alpaca /O/
separator=,

or (new lines are allowed)

cat
dog
alpaca
/O/
separator=,

will match any of the following example answers with 100% correctness.

cat,dog,alpaca

alpaca,cat,dog

If one line is not correct, too much or missing it will result in a point reduction. In this case the following answers will result in 66% correctness:

alpaca,cat

alpaca,cat,elephant

alpaca,cat,dog,elephant

I: Infinite Space

This option is enabled by default and can be disabled by specifying the letter I in the options. If this option is enabled all spaces will be replaced with ([ \t]+). Thereby they match one or more whitespace characters.

Examples

The regular expression some test sentence will match any of the following example answers:

• some test sentence
• some test sentence

But it will not match:

• sometestsentence
• some testsentence

T: Trim Spaces

This option is enabled by default and can be disabled by specifying the letter T in the options. If this option is enabled all trailing and leading empty lines in the answer, as well as all trailing and leading spaces of every line in the answer, will be ignored. Trailing empty lines will always be ignored, even if this option is disabled.

Examples

The regular expression test// or test/T/ will match any of the following example answers:

• test
• test

and also the following example answer

  test

Common Mistakes and special cases

• If the regular expression contains spaces at the end of a line these must also be contained in the answer.
• If you encounter a problem with a complex regular expression, try to disable the Infinite Space (i) option.