A DSL for describing questionnaires with a simple terminal interpreter.
Example:
Q[type]: Project type
1[lang]. Language project
2[gen]. Generator project
Q[type=lang, extension]: File extension (e.g. "*.que"):
___ /\*\.[a-z0-9]+/
Q[package]: Package name
___ /[a-z][a-z0-9_]*$/
Q: Author name
___
Q: Author email
___
Q: This question is to
test multiline feature and indenting.
1. Working
2. Not working.
This is also to test multiline in choices.
Each question is of the form Q: or Q[<comma separated terms>]. If [] with
terms is given, each term is either a question ID or a boolean expression of the
form <lhs>=<rhs> where <lhs> is a reference to previous question ID while
<rhs> is a value of the previous question which must be satisfied for this
question to be used. This is handy if the questions that follows depends on the
answer. See File extension question above which is asked only for language
projects (thus type=lang, the type is a reference to the previous question).
The type of the question is either free-form (specified by ___) or choice if
enumerated options are given.
Both question and option can span multiple lines (like in the last question).
For free-form question an optional regular expression can be given inside //
(see Package name above). This regex is an input validator.
This package provides an interpreter of the questionnaire
txquestionnaire.questionnaire_interpreter(model, data=None) which accepts the
model created by parsing of questionnaire description using this language and
optionally a dictionary of default answers keyed by either question ID if given
or question number. The interpreter will run the questionnaire on the console
and return a dictionary of collected data. The values for options are ordinal
numbers or IDs if given (like in Project type above -- lang, gen).
Please see tests for the detailed usage.