(DRAFT)
The imbNLP library introduces quite unique data structure designed to facilitate robust regular expression queries over different facets of tokenized content graph model. It is called textual map, a textual representation pivoted collection, keeping record on references to the graph elements on char position level. When the textual map, shown below in conceptual example, is queried for range from 5 to 9, it will return: words T1 and T2, sentence TS1 and phrase P1.
To build textual map, we use contentTokenSubjectRender.render method, passing as parameters: element of the content graph and mode of rendering that we want to query later.
| Value | |
|---|---|
| currentForm | Putting out the current form of the token |
| initialForm | Putting out the initial form of the token |
| posTypeTagForm | The POSPart-of-speech, is very frequently used to provide linguistic information to NER and CR in form of features in statistical approaches... type tag form: |A V|N|Part|Conj| |
| posTypeAndGramTagForm | The POSPart-of-speech, is very frequently used to provide linguistic information to NER and CR in form of features in statistical approaches... type and grammatical tag form: [Amspf:Nmsps]|ADJ[fs1f]| |
| flagsForm | The flags form: |phoneOfficeNeedle|symbol|number phone phoneNumber| |
| flagsFullForm | The flags form: |dat_business.phoneOfficeNeedle|tkn_contains.symbols|tkn_contains.number dat_business.phone dat_business.phoneNumber| |
| descriptive | The descriptive form: |”kompanijom”:”kompanija”:N,common,f,s,instrumental:lowerCase,letter,onlyLetters| |
| lemmaForm |