Apertium-recursive/Cookbook - Revision history

Trondtr at 06:11, 1 June 2023

2023-06-01T06:11:05Z

Unhammer: /* Reordering */

2021-10-06T13:22:49Z

Reordering

Unhammer: /* Reordering */

2021-10-06T13:18:35Z

Reordering

Unhammer: indices on blanks is deprecated

2021-09-23T13:20:26Z

indices on blanks is deprecated

Popcorndude: syntax update

2019-08-06T16:48:37Z

syntax update

Popcorndude: cases and prepositions

2019-08-06T14:57:43Z

cases and prepositions

Popcorndude at 16:12, 5 August 2019

2019-08-05T16:12:10Z

Popcorndude: Created page with "This page is intended to be a collection of solutions to common transfer problems. It is sorted vaguely be complexity, so it could potentially also serve as a tutorial. Functi..."

2019-08-05T15:46:56Z

Created page with "This page is intended to be a collection of solutions to common transfer problems. It is sorted vaguely be complexity, so it could potentially also serve as a tutorial. Functi..."

New page

This page is intended to be a collection of solutions to common transfer problems. It is sorted vaguely be complexity, so it could potentially also serve as a tutorial. Functioning rulesets for each of these examples can be found at https://github.com/apertium/apertium-recursive/tree/master/tests/cookbook.

=== Noun-Adjective Gender Agreement ===

NP -> adj n { 1(gender=2.gender) _1 2 } ;

This will copy the gender tag of the noun to the adjective.

=== Reordering ===

NP -> adj n { 2 _1 1 } ;

This will reverse the order of the noun and adjective.

=== Agreement across Multiple Levels ===

NP -> n.$gender.$number { 1 } |
adj NP.$gender.$number { 2 _1 1(gender=2.gender) } ;
DP -> det NP { 1(number=2.number) _1 2 } ;

These rules will match a determiner, 0 or more adjectives, and a noun and make all the adjectives agree with the noun in gender and the determiner agree with the noun in number. The adjectives will also be output after the noun in the reverse of their input order.

← Older revision		Revision as of 06:11, 1 June 2023
Line 64:		Line 64:

	Now things can be added to the <code>NP</code> rule to deal with adjectives and the <code>DP</code> rule could match just an <code>NP</code> and insert determiners and so on.		Now things can be added to the <code>NP</code> rule to deal with adjectives and the <code>DP</code> rule could match just an <code>NP</code> and insert determiners and so on.

			[[Category:Recursive transfer]]

@@ Line 20: / Line 20: @@
 the casings will be switched around.
-If instead of adj/n you have non-leaf categories like A NP in NP -> A NP { … } then you need to do something else (WHAT?)
+If instead of adj/n you have non-leaf categories like A N in NP -> A N { … } then you also need to propagate the case manually (% will not do it for you), e.g.
+ N -> %n { 1[lemcase=$lemcase] } ;
+ A -> %adj { 1[lemcase=$lemcase] } ;
+ NP -> A %N { 2[lemcase=1.lemcase] _ 1[lemcase=2.lemcase] } ;
 === Agreement across Multiple Levels ===

@@ Line 12: / Line 12: @@
 This will reverse the order of the noun and adjective.
+This will fail to move the titlecasing at the start of a sentence. If you do
+ NP -> adj n { 2[lemcase=1.lemcase] _ 1[lemcase=2.lemcase] } ;
+the casings will be switched around.
+If instead of adj/n you have non-leaf categories like A NP in NP -> A NP { … } then you need to do something else (WHAT?)
 === Agreement across Multiple Levels ===

@@ Line 3: / Line 3: @@
 === Noun-Adjective Gender Agreement ===
- NP -> adj n { 1[gender=2.gender] _1 2 } ;
+ NP -> adj n { 1[gender=2.gender] _ 2 } ;
 This will copy the gender tag of the noun to the adjective.
@@ Line 9: / Line 9: @@
 === Reordering ===
- NP -> adj n { 2 _1 1 } ;
+ NP -> adj n { 2 _ 1 } ;
 This will reverse the order of the noun and adjective.
@@ Line 16: / Line 16: @@
  NP -> n.$gender.$number { 1 } |
-       adj NP.$gender.$number { 2 _1 1[gender=2.gender] } ;
+       adj NP.$gender.$number { 2 _ 1[gender=2.gender] } ;
- DP -> det NP { 1[number=2.number] _1 2 } ;
+ DP -> det NP { 1[number=2.number] _ 2 } ;
 These rules will match a determiner, 0 or more adjectives, and a noun and make all the adjectives agree with the noun in gender and the determiner agree with the noun in number. The adjectives will also be output after the noun in the reverse of their input order.

@@ Line 3: / Line 3: @@
 === Noun-Adjective Gender Agreement ===
- NP -> adj n { 1(gender=2.gender) _1 2 } ;
+ NP -> adj n { 1[gender=2.gender] _1 2 } ;
 This will copy the gender tag of the noun to the adjective.
@@ Line 16: / Line 16: @@
  NP -> n.$gender.$number { 1 } |
-       adj NP.$gender.$number { 2 _1 1(gender=2.gender) } ;
+       adj NP.$gender.$number { 2 _1 1[gender=2.gender] } ;
- DP -> det NP { 1(number=2.number) _1 2 } ;
+ DP -> det NP { 1[number=2.number] _1 2 } ;
 These rules will match a determiner, 0 or more adjectives, and a noun and make all the adjectives agree with the noun in gender and the determiner agree with the noun in number. The adjectives will also be output after the noun in the reverse of their input order.
@@ Line 25: / Line 25: @@
  v_inf: _.<inf>;
  v_fin: _.tense.person.number;
- v: (if (1.tense = inf) 1[v_inf]
+ v: (if (1.tense = inf) 1(v_inf)
-     else 1[v_fin]);
+     else 1(v_fin));
 When outputting a verb, this will not try to add person and number tags if the tense is <code><inf></code>.
@@ Line 40: / Line 40: @@
              { 1 }
           else
-             { *[pr](lemh=1.case>preps) _ 1 } ) ;
+             { *(pr)[lemh=1.case>preps] _ 1 } ) ;
 === Case to Preposition on Another Level ===
@@ Line 48: / Line 48: @@
  NP -> n [$preps=1.case>preps] { 1 } ;
  DP -> det NP.$preps { 1 _1 2 } ;
- PP -> DP ?(1.preps not = NOpr) { *[pr](lemh=1.preps) _ 1 } ;
+ PP -> DP ?(1.preps not = NOpr) { *(pr)[lemh=1.preps] _ 1 } ;
 Now things can be added to the <code>NP</code> rule to deal with adjectives and the <code>DP</code> rule could match just an <code>NP</code> and insert determiners and so on.