New Tag manager methdos

[linuxscout]

Add new tag manager functionalities to handle POS tag merging:

Need: When we replace a word 'w1' by another word 'w2', we need to inflect the new word according the replaced word tag.

POS Tag Structure:

3 parts:

• word class and category
• Conjugation ( suffixes used to conjugate words in dual form, feminin form).
• Affixes ( prefixes and enclitics)

Example:

• Example: replace "يعودون" by "يرجعون",
  • target lemma is رجع
  • lemma postag is different from source postag
    • verb عاد : "VW1;--------;---"
    • verb رجع : "V31;--------;---"
  • Inflected source postag: يعودون "VW1;M3H-faU;WS-"
    • The wanted target يرجعون"V31;M3H-faU;WS-"

Process

Same word type

if postags have the same word class, it can be used to replace :

• a verb by new verb : يعتبرونه by يعدونه
• a preposition by a preposition: يتردد عليه by يتردد إليه
• a noun by a noun: بالمستهتر by بالمستهين
  • replace Conjugation Part and Affixes part from the source postag to target postag
  • merge(source: "VW1;M3H-faU;WS-", target: "V31;--------;---") => "V31;M3H-faU;WS-"

Different word type

if postags haven't the same word class, it can be used to replace :

• a verb by new preposition : تحاشاه by تحاشى منه
• a preposition by a verb : تحاشاه by
  • replace and Affixes part(only compatible flags) from the source postag to target postag
    • change pronoun suffix
    • verb specific features aren't borrowed
  • merge(source: "VW1;M1H-pa-;--H", target: "PRD;---;---") => "PRD;---;--H"

[linuxscout]

Implemented