2

I'm building a chatbot subset of RiveScript and trying to build the pattern matching parser with regular expression. Which three regexes match the following three examples?

ex1: I am * years old
valid match:
- "I am 24 years old"
invalid match:
- "I am years old"

ex2: what color is [my|your|his|her] (bright red|blue|green|lemon chiffon) *
valid matches:
- "what color is lemon chiffon car"
- "what color is my some random text till the end of string"

ex3: [*] told me to say *
valid matches:
- "Bob and Alice told me to say hallelujah"
- "told me to say by nobody"

The wildcards mean any text that is not empty is acceptable.

In example 2, anything between [ ] is optional, anything between ( ) is alternative, each option or alternative is separated by a |.

In example 3, the [*] is an optional wildcard, meaning blank text can be accepted.

Rodia
  • 1,335
  • 7
  • 19
  • 27
MiP
  • 481
  • 1
  • 6
  • 14

1 Answers1

2

  1. https://regex101.com/r/CuZuMi/4

    I am (?:\d+) years old

  2. https://regex101.com/r/CuZuMi/2

    what color is.*(?:my|your|his|her).*(?:bright red|blue|green|lemon chiffon)?.*

  3. https://regex101.com/r/CuZuMi/3

    .*told me to say.*

I am using mostly 2 things:

  1. (?:) non-capture groups, to group things together like the parenthesis use on math.
  2. .* match any character 0 or more times. Could be replaced by {1,3} to match between 1 and 3 times.

You can exchange * by + to match at least 1 character, instead of 0. And the ? after the non-capture group, makes that group optional.

These are golden place for you to start:

  1. http://www.rexegg.com/regex-quickstart.html
  2. https://regexone.com/
  3. http://www.regular-expressions.info/quickstart.html
  4. Reference - What does this regex mean?
Community
  • 1
  • 1
user
  • 5,816
  • 7
  • 53
  • 105
  • What about this scenario `(.+ |.*)told me to say (.+)`. If there are characters before `told me to say `, there must also have a space in between ` ` to prevent word colliding, e.g. `Alice told me to say hi` and `told me to say hi` should work but `Alicetold me to say hi` doesn't. – MiP Feb 26 '17 at 05:36
  • 1
    Just remove the space as in `(.*)told me to say (.+)` and `Alicetold me to say hi` will match. – user Feb 26 '17 at 05:40