Debian Codegraph @

manual/dependencies.xml

raw
1 <?xml version="1.0" encoding="utf-8"?>
2 <!DOCTYPE chapter SYSTEM "docbook-dtd-45/docbookx.dtd">
3
4 <chapter id="dependencies">
5 <title>Making use of Dependencies</title>
6
7 <para>
8 CG-3 can work with dependency trees in various ways. The input cohorts can have existing
9 dependencies; the grammar can create new attachments; or a combination of the two.
10 </para>
11
12 <section id="dep-setparent">
13 <title>SETPARENT</title>
14 <indexterm>
15 <primary>SETPARENT</primary>
16 </indexterm>
17 <screen>
18 [wordform] SETPARENT &lt;target&gt; [contextual_tests]
19 TO|FROM &lt;contextual_target&gt; [contextual_tests] ;
20 </screen>
21 <para>
22 Attaches the matching reading to the contextually targetted cohort as a child.
23 The last link of the contextual test is used as target.
24 </para>
25 <para>
26 If the contextual target is a scanning test and the first found candidate cannot be attached
27 due to loop prevention, SETPARENT will look onwards for the next candidate.
28 This can be controlled with rule option <link linkend="rule-options-nearest">NEAREST</link>
29 and <link linkend="rule-options-allowloop">ALLOWLOOP</link>.
30 </para>
31 <screen>
32 SETPARENT targetset (-1* ("someword"))
33 &nbsp; TO (1* (step) LINK 1* (candidate)) (2 SomeSet) ;
34 </screen>
35 </section>
36
37 <section id="dep-setchild">
38 <title>SETCHILD</title>
39 <indexterm>
40 <primary>SETCHILD</primary>
41 </indexterm>
42 <screen>
43 [wordform] SETCHILD &lt;target&gt; [contextual_tests]
44 TO|FROM &lt;contextual_target&gt; [contextual_tests] ;
45 </screen>
46 <para>
47 Attaches the matching reading to the contextually targetted cohort as the parent.
48 The last link of the contextual test is used as target.
49 </para>
50 <para>
51 If the contextual target is a scanning test and the first found candidate cannot be attached
52 due to loop prevention, SETCHILD will look onwards for the next candidate.
53 This can be controlled with rule option <link linkend="rule-options-nearest">NEAREST</link>
54 and <link linkend="rule-options-allowloop">ALLOWLOOP</link>.
55 </para>
56 <screen>
57 SETCHILD targetset (-1* ("someword"))
58 &nbsp; TO (1* (step) LINK 1* (candidate)) (2 SomeSet) ;
59 </screen>
60 </section>
61
62 <section id="dep-remparent">
63 <title>REMPARENT</title>
64 <indexterm>
65 <primary>REMPARENT</primary>
66 </indexterm>
67 <screen>
68 [wordform] REMPARENT &lt;target&gt; [contextual_tests] ;
69 </screen>
70 <para>
71 Sets the target as having no parent.
72 </para>
73 <screen>
74 REMPARENT targetset IF (p (cop)) ;
75 </screen>
76 </section>
77
78 <section id="dep-switchparent">
79 <title>SWITCHPARENT</title>
80 <indexterm>
81 <primary>SWITCHPARENT</primary>
82 </indexterm>
83 <screen>
84 [wordform] SWITCHPARENT [WITHCHILD &lt;child_set&gt;] &lt;target&gt; [contextual_tests] ;
85 </screen>
86 <para>
87 Sets the grandparent of the target as its parent and sets the target as the parent of the previous parent. If <link linkend="rule-options-withchild">WTIHCHILD</link> is present, any siblings of the target which match the set will also become children of the target.
88 </para>
89 <screen>
90 "&lt;and&gt;"
91 "and" CCONJ #1->2
92 "&lt;in&gt;"
93 "in" ADP #2->0
94 "&lt;the&gt;"
95 "the" DET #3->4
96 "&lt;house&gt;"
97 "house" NOUN #4->2
98 "&lt;.&gt;"
99 "." PUNCT #5->2
100
101 SWITCHPARENT WITHCHILD (*) - (PUNCT) (NOUN) IF (p (ADP)) ;
102
103 "&lt;and&gt;"
104 "and" CCONJ #1->4
105 "&lt;in&gt;"
106 "in" ADP #2->4
107 "&lt;the&gt;"
108 "the" DET #3->4
109 "&lt;house&gt;"
110 "house" NOUN #4->0
111 "&lt;.&gt;"
112 "." PUNCT #5->2
113 </screen>
114 </section>
115
116 <section id="dep-input">
117 <title>Existing Trees in Input</title>
118 <para>
119 Dependency attachments in input comes in the form of #X-&gt;Y or #X→Y tags where X is the number of the current
120 node and Y is the number of the parent node. The X must be unique positive integers and should
121 be sequentially enumerated. '0' is reserved and means the root of the tree, so no node may claim
122 to be '0', but nodes may attach to '0'.
123 </para>
124 <para>
125 <emphasis>If the Y of a reading cannot be located, it will be reattached to itself.</emphasis>
126 <emphasis>If a reading contains more than one attachment, only the last will be honored.</emphasis>
127 <emphasis>If a cohort has conflicting attachments in its readings, the result is undefined.</emphasis>
128 </para>
129 <para>
130 For example:
131 <screen>
132 "&lt;There&gt;"
133 &nbsp; "there" &lt;*&gt; ADV @F-SUBJ #1-&gt;0
134 "&lt;once&gt;"
135 &nbsp; "once" ADV @ADVL #2-&gt;0
136 "&lt;was&gt;"
137 &nbsp; "be" &lt;SVC/N&gt; &lt;SVC/A&gt; V PAST SG1/3 VFIN IMP @FMV #3-&gt;2
138 "&lt;a&gt;"
139 &nbsp; "a" &lt;Indef&gt; ART DET CENTRAL SG @&gt;N #4-&gt;5
140 "&lt;man&gt;"
141 &nbsp; "man" N NOM SG @SC #5-&gt;0
142 "&lt;$.&gt;"
143 </screen>
144 </para>
145 </section>
146
147 <section id="dep-delimit">
148 <title>Using Dependency as Delimiters</title>
149 <para>
150 Cmdline flag -D or --dep-delimit will enable the use of dependency information to delimit windows.
151 Enabling this will disable <link linkend="keyword-delimiters">DELIMITERS</link> entirely,
152 but will not affect the behavior of <link linkend="keyword-soft-delimiters">SOFT-DELIMITERS</link>
153 nor the hard/soft cohort limits.
154 </para>
155 <para>
156 Windows are delimited if a cohort has a node number less than or equal to
157 the highest previously seen node number, and also if a cohort has a node number that seems like a discontinuous
158 jump up in numbers.
159 The discontinuous limit is by default 10 but you can pass a number to -D/--dep-delimit to set it yourself.
160 Some systems do not output dependency numbers for punctuation, so setting it too low may break those;
161 the default 10 was chosen since it is unlikely any real text would have 10 sequential cohorts not part of the tree.
162 </para>
163 <para>
164 For example:
165 <emphasis>#4→5 followed by #3→4 will delimit.</emphasis>
166 <emphasis>#4→5 followed by #4→4 will delimit.</emphasis>
167 <emphasis>#4→5 followed by #15→4 will delimit.</emphasis>
168 <emphasis>#4→5 followed by #5→3 will not delimit.</emphasis>
169 </para>
170 </section>
171
172 <section id="dep-grammar">
173 <title>Creating Trees from Grammar</title>
174 <para>
175 It is also possible to create or modify the tree on-the-fly with rules.
176 See <link linkend="keyword-setparent">SETPARENT</link>
177 and <link linkend="keyword-setchild">SETCHILD</link>.
178 Dependencies created in this fashion will be output in the same format as above.
179 </para>
180 <para>
181 For example:
182 <screen>
183 SETPARENT (@&gt;N) (0 (ART DET))
184 &nbsp; TO (1* (N)) ;
185
186 SETPARENT (@&lt;P)
187 &nbsp; TO (-1* (PRP)) (NEGATE 1* (V)) ;
188 </screen>
189 </para>
190 </section>
191
192 <section id="dep-contextual-tests">
193 <title>Contextual Tests</title>
194 <para>
195 Either case, once you have a dependency tree to work with,
196 you can use that in subsequent contextual tests as seen below.
197 These positions can be combined with the window spanning options.
198 </para>
199
200 <section id="dep-ct-parent">
201 <title>Parent</title>
202 <indexterm>
203 <primary>p</primary>
204 </indexterm>
205 <para>
206 The 'p' position asks for the parent of the current position.
207 </para>
208 <screen>
209 (-1* (ADJ) LINK p (N))
210 </screen>
211 </section>
212
213 <section id="dep-ct-ancestors">
214 <title>Ancestors</title>
215 <indexterm>
216 <primary>pp</primary>
217 </indexterm>
218 <para>
219 The 'pp' position asks for an ancestor of the current position,
220 where ancestor is defined as any parent, grand-parent, great-grand-parent, etc...
221 </para>
222 <screen>
223 (-1* (N) LINK pp (ADJ))
224 </screen>
225 <para>
226 <emphasis>The analogue of <link linkend="gotcha-cc-vs-cstar">difference between cc and c*</link> applies to pp vs. p*</emphasis>
227 </para>
228 </section>
229
230 <section id="dep-ct-children">
231 <title>Children</title>
232 <indexterm>
233 <primary>c</primary>
234 </indexterm>
235 <para>
236 The 'c' position asks for a child of the current position.
237 </para>
238 <screen>
239 (-1* (N) LINK c (ADJ))
240 </screen>
241 </section>
242
243 <section id="dep-ct-descendents">
244 <title>Descendents</title>
245 <indexterm>
246 <primary>cc</primary>
247 </indexterm>
248 <para>
249 The 'cc' position asks for a descendent of the current position,
250 where descendent is defined as any child, grand-child, great-grand-child, etc...
251 </para>
252 <screen>
253 (-1* (N) LINK cc (ADJ))
254 </screen>
255 <para>
256 <emphasis><link linkend="gotcha-cc-vs-cstar">Difference between cc and c*</link></emphasis>
257 </para>
258 </section>
259
260 <section id="dep-ct-siblings">
261 <title>Siblings</title>
262 <indexterm>
263 <primary>s</primary>
264 </indexterm>
265 <para>
266 The 's' position asks for a sibling of the current position.
267 </para>
268 <screen>
269 (-1* (ADJ) LINK s (ADJ))
270 </screen>
271 </section>
272
273 <section id="dep-ct-self">
274 <title>Self</title>
275 <indexterm>
276 <primary>S</primary>
277 </indexterm>
278 <para>
279 The 'S' option allows the test to look at the current target as well.
280 Used in conjunction with p, c, cc, s, or r to test self and the relations.
281 </para>
282 <para>
283 Be aware that BARRIER and CBARRIER will check and thus possibly stop at the current target when 'S' is in effect.
284 This can be toggled on a per context basis with modifier <link linkend="dep-ct-no-barrier">N</link>.
285 </para>
286 <screen>
287 (cS (ADJ))
288 </screen>
289 </section>
290
291 <section id="dep-ct-no-barrier">
292 <title>No Barrier</title>
293 <indexterm>
294 <primary>N</primary>
295 </indexterm>
296 <para>
297 The 'N' option causes barriers to ignore the self cohort.
298 If <link linkend="grammar-options-self-no-barrier">self-no-barrier</link> is enabled,
299 then instead it forces barriers to respect the self cohort.
300 </para>
301 </section>
302
303 <section id="dep-ct-deep">
304 <title>Deep Scan</title>
305 <indexterm>
306 <primary>*</primary>
307 </indexterm>
308 <para>
309 The '*' option behaves differently when dealing with dependencies.
310 Here, the '*' option allows the test to perform a deep scan.
311 Used in conjunction with p, c, or s to continue until there are no deeper relations.
312 For example, position 'c*' tests the children, grand-children, great-grand-children, and so forth.
313 </para>
314 <screen>
315 (c* (ADJ))
316 </screen>
317 <para>
318 <emphasis><link linkend="gotcha-cc-vs-cstar">Difference between cc and c*</link></emphasis>
319 </para>
320 </section>
321
322 <section id="dep-ct-leftof">
323 <title>Left of</title>
324 <indexterm>
325 <primary>l</primary>
326 </indexterm>
327 <para>
328 The 'l' option limits the search to cohorts that are to the left of the current target.
329 </para>
330 <screen>
331 (lc (ADJ))
332 </screen>
333 </section>
334
335 <section id="dep-ct-rightof">
336 <title>Right of</title>
337 <indexterm>
338 <primary>r</primary>
339 </indexterm>
340 <para>
341 The 'r' option limits the search to cohorts that are to the right of the current target.
342 </para>
343 <screen>
344 (rc (ADJ))
345 </screen>
346 </section>
347
348 <section id="dep-ct-leftmost">
349 <title>Leftmost</title>
350 <indexterm>
351 <primary>ll</primary>
352 </indexterm>
353 <indexterm>
354 <primary>lll</primary>
355 </indexterm>
356 <indexterm>
357 <primary>llr</primary>
358 </indexterm>
359 <para>
360 The 'll' option limits the search to the leftmost cohort of the possible matches.
361 Note that this cohort may be to the right of the current target;
362 use 'lll' if you want the leftmost of the cohorts to the left of the current target,
363 or 'llr' if you want the leftmost of the cohorts to the right of the current target.
364 </para>
365 <screen>
366 (llc (ADJ))
367 </screen>
368 </section>
369
370 <section id="dep-ct-rightmost">
371 <title>Rightmost</title>
372 <indexterm>
373 <primary>rr</primary>
374 </indexterm>
375 <indexterm>
376 <primary>rrr</primary>
377 </indexterm>
378 <indexterm>
379 <primary>rrl</primary>
380 </indexterm>
381 <para>
382 The 'rr' option limits the search to the rightmost cohort of the possible matches.
383 Note that this cohort may be to the left of the current target;
384 use 'rrr' if you want the rightmost of the cohorts to the right of the current target,
385 or 'rrl' if you want the rightmost of the cohorts to the left of the current target.
386 </para>
387 <screen>
388 (rrc (ADJ))
389 </screen>
390 </section>
391
392 <section id="dep-ct-all">
393 <title>All Scan</title>
394 <indexterm>
395 <primary>ALL</primary>
396 </indexterm>
397 <para>
398 The 'ALL' option will require that all of the relations match the set.
399 For example, position 'ALL s' requires that all of the siblings match the set.
400 </para>
401 <screen>
402 (ALL s (ADJ))
403 </screen>
404 </section>
405
406 <section id="dep-ct-not">
407 <title>None Scan</title>
408 <indexterm>
409 <primary>NONE</primary>
410 </indexterm>
411 <para>
412 The 'NONE' option will require that none of the relations match the set.
413 For example, position 'NONE c' requires that none of the children match the set.
414 </para>
415 <screen>
416 (NONE c (ADJ))
417 </screen>
418 </section>
419 </section>
420
421 </chapter>