Informatics1-2017/Practice3

Regular expressions

Regular expressions are used to find complex paterns in text, or if we want to substitude these patterns for something else. We will use this site https://regex101.com/#python

• Special characters: These don't symbolize themselves, to find them in text we have to escape them with \ for example: \$, \^ etc.

. ^ $ * + ? { } [ ] \ | ( )

Character classes

For the time being we only use one character patterns.

• \d: arbitrary number, \D: arbitrary character that is not a number.
• \w: arbitrary alphanumeric character, character, number, or underline (_), \W: arbitrary non-alphanumeric character.
• \s: whitespace, which is tab, end of line, space, \S arbitrary non-whitespace character.
• We can create custom character classes: [xyz], or we can make exclusions, e.g. [^xyz]. The former matches x, y or z, the latter matches any character that is not x, y or z. Using a dash we can specify intervals, e.g. [a-z] this matches all lower case characters, but for example [A-Za-z0-9] maches all uppercase, lowercase and numeric characters.
• ^: beginning of line, $, end of line.
• A . matches any character.

Recurrence

Notation	Recurrance number	Example
*	0,1, or however many	\d* matches '123', and it even matches the empty string, as well
+	at least 1	\d+ matches any number of numeric characters
?	0 or 1	the?an matches 'then' and 'than' as well
{m,n}	At least m, at most n number of something, both of them are optional	:D{4,10} does not match ':DDDDDDDDDDDDDD'

Choice

• The pattern a|e|i|o|u matches any vowel. Try the GetValue|Get|Set|SetValue expression. What do we get for the text SetValue?

Grouping

We can specify groups within the expression. The following example matches any string that repeats once:

(.*)\1

We can match for HTML tags:

<([A-Z][A-Z0-9]*)\b[^>]*>.*?</\1>

We can specify multiple groups, the sequence of the opening parenthesis specifies the number. Replace the ending if email addresses to .hu!

(\w+)@((\w+)\.)+(\w+)

Tasks

• date formats: yyyy.mm.dd
• Mobile numbers starting with +36 20, +36 30, +36 70
• Link tags (<a>anything here</a>)
• Webpage adresses
• Find the BME logo with patterns on: http://www.bme.hu/?language=en
• 2 digit numbers divisible by 4
• leap years

Advanced tasks

• Roman numerals written with capital latters

Millenium: M{0,4}, century: CM|CD|D?C{0,3}, decade: XC|XL|L?X{0,3}, year: IX|IV|V?I{0,3}.

• Positive integers, really long numbers might contain spaces when grouped by 3 digits (1 000, 435 000 000).
• Decimal color code in HTML (3 or 6 hexa number)