Extraction from various docstrings that contain Boundary Type

[mmdoja]

Is your feature request related to a problem? Please describe

Extraction of boundary type of the following two formats:

Must be between 0 and 1 
tuple (q_min, q_max), 0.0 < q_min < q_max < 100.0
The ElasticNet mixing parameter, with 0 <= l1_ratio <= 1
float > 0 and <= 1

non-negative float
Must be strictly positive
must be a positive float

Desired solution

Although some particular format can be solved with the help of Regex, the rest can be solved using Matchers in spaCy. Phrase matcher and Dependency matcher features of spaCy can also be looked into for solving this problem statement. The reason for using spaCy here is for the machine to process the natural language in which the description of the boundaries are given.

For example: non-negative float so non-negative here means inclusion of 0.

Possible alternatives (optional)

Screenshots (optional)

Additional context (optional)

[duklin]

Ideally, all the different docstrings would be first transformed using regex as in a pre-processing stage before using spacy to find the relations between the boundary and the parameter itself.
This issue seems very related to lars-reimann/api-editor#404