transform2teisimple.xqm 18 KB
Newer Older
MRodz's avatar
MRodz committed
1
2
3
4
5
xquery version "3.1";

(:~ 
 : This modules handles the conversion of the Fontante-TEI/XML into TEISimple
 : for the edited text. The resulting TEISimple is the basis for the "Editerter
MRodz's avatar
MRodz committed
6
 : Text" view on the website and the book and represent the latest layer of text.
MRodz's avatar
MRodz committed
7
8
9
10
11
12
13
14
15
16
17
 : 
 : The resulting TEISimple orientates itself toward the so-called "SUB-Schema"
 : which is a subset for TEISimple that is recommended for future projects that
 : are held by the Göttingen State and University Library (SUB).
 : 
 : @author Michelle Rodzis
 : @version 0.1
 :)

module namespace fontane-simple="http://fontane-nb.dariah.eu/teisimple";

MRodz's avatar
MRodz committed
18

MRodz's avatar
MRodz committed
19
20
21
declare namespace tei="http://www.tei-c.org/ns/1.0";
declare namespace test="http://exist-db.org/xquery/xqsuite";

22
import module namespace config="http://textgrid.de/ns/SADE/config" at "../config/config.xqm";
MRodz's avatar
MRodz committed
23
import module namespace console="http://exist-db.org/xquery/console";
MRodz's avatar
MRodz committed
24
import module namespace functx = "http://www.functx.com";
MRodz's avatar
MRodz committed
25
26
27
28
29
30
31

(:~ 
 : The main function initiates the transformation of a given notebook.
 : 
 : TODO: adapt to several input files?
 : 
 : :)
32
declare function fontane-simple:main($file as xs:string) as xs:string? {
MRodz's avatar
MRodz committed
33
34
35
36
37
38
39
  let $doc := 
    try {
      (doc("/db/sade-projects/textgrid/data/xml/data/" || $file))
    } catch * {
      (console:log("It was not possible to open the requested file " || $file))
    }
    
MRodz's avatar
MRodz committed
40
41
42
43
44
    let $front-covers := $doc//tei:sourceDoc/tei:surface[contains(@n, "front_cover")]/*
    let $back-covers := $doc//tei:sourceDoc/tei:surface[contains(@n, "back_cover")]/*
    let $content := $doc//tei:sourceDoc/tei:surface[not(contains(@n, "cover") 
        and matches(@n, "spine"))]
    
45
    let $tei := <TEI xmlns="http://www.tei-c.org/ns/1.0">
MRodz's avatar
MRodz committed
46
        {$doc//tei:teiHeader}
MRodz's avatar
MRodz committed
47
48
49
50
51
        <text>
            <front>{fontane-simple:transform($front-covers)}</front>
            <body>{fontane-simple:transform($content)}</body>
            <back>{fontane-simple:transform($back-covers)}</back>
        </text>
MRodz's avatar
MRodz committed
52
      </TEI>
53
    return xmldb:store("/db/apps/SADE/resources/xml/", "tei-simple.xml", $tei)
MRodz's avatar
MRodz committed
54
55
};

MRodz's avatar
MRodz committed
56
57
58
59
60
61
62
63
64
65
66
67
68
69
(:~ 
 : Recursivly iterates the passed nodes and converts them according to the 
 : requirements for the "Edierter Text". 
 : 
 : The requirements are part of the encoding documentation, c.f.
 : https://fontane-nb.dariah.eu/doku.html.
 : 
 : 
 : @author Michelle Rodzis
 : @param $nodes
 : @return $node()* a TEISimple element
 : 
 :)
declare function fontane-simple:transform($nodes as node()*) as node()* {
MRodz's avatar
MRodz committed
70
71
72
    for $node in $nodes
      return
        typeswitch ($node)
MRodz's avatar
MRodz committed
73
74
75
76
77
78
79
        case text() return
            local:prepare-text($node)
        
        case element(tei:lb) return
            if($node[@break = "keepHyphen"]) then
                ()
            else
80
                local:keep-element($node)
MRodz's avatar
MRodz committed
81
82
83
84
85
                
        case element(tei:g) return
            if($node[@ref = "#vds"]) then
                ()
            else if($node/@ref ="#rth" or $node/@ref ="#hb") then
86
                local:keep-element($node)
MRodz's avatar
MRodz committed
87
88
            else
                fontane-simple:transform($node/node())
89
            
90
91
        case element(tei:graphic) return
            local:keep-element($node)
92
                
MRodz's avatar
MRodz committed
93
94
95
96
97
98
        case element(tei:del) return
            if($node/parent::tei:restore) then
                fontane-simple:transform($node/node())
            else if($node/descendant::tei:restore) then
                fontane-simple:transform($node/descendant::tei:restore)
            else
MRodz's avatar
MRodz committed
99
                ()
100
                
MRodz's avatar
MRodz committed
101
102
103
104
105
106
        case element(tei:restore) return
            if(count($node/child::*) = 1 and $node/child::tei:del
            and $node/ancestor::tei:del) then
                ()
            else
                fontane-simple:transform($node/node())
107
                
MRodz's avatar
MRodz committed
108
109
110
111
112
113
114
115
        case element(tei:retrace) return
            if($node/@rend) then
                element tei:seg {
                    $node/@rend,
                    fontane-simple:transform($node/node())
                }
            else
                fontane-simple:transform($node/node())
116
117
                
        case element(tei:add) return
118
119
            if($node/@type = "edited_text" 
            or $node/child::tei:seg[@type = "multiphrase"]) then
120
                local:keep-element($node)
MRodz's avatar
MRodz committed
121
122
            else if($node/@cause ="catchword" or $node/@cause ="unclear") then
                ()
MRodz's avatar
MRodz committed
123
124
            else if($node/@rend ="|") then
                fontane-simple:transform($node/node())
125
126
127
128
129
130
131
132
133
134
135
136
            else if(not($node/@xml:id)) then
                fontane-simple:transform($node/node())
            else if(fontane-simple:is-transposed($node)) then
                let $corresp := //tei:metamark[matches(@target, $node/@xml:id)]
                return
                    (fontane-simple:transform($corresp/node()),
                    fontane-simple:transform($node/node()))
            else
                fontane-simple:transform($node/node())
                
        case element(tei:addSpan) return
            if($node/@type = "edited_text") then
137
                local:keep-element($node)
138
139
140
            else
                ()       
                
141
142
        case element(tei:fw) return
            ()
143
                
144
145
146
147
148
149
150
        case element(tei:line) return
            if(count($node/*) = 1 and $node/child::tei:fw) then
                ()
            else if(not($node/@type = "item")) then
                fontane-simple:transform($node/node())
            else if($node/@type = "item" and not($node/@xml:id)) then
                element tei:item {
151
                    fontane-simple:transform($node/node())
152
153
154
155
156
157
                }
            else if($node/@type = "item" 
            and fontane-simple:is-transposed($node)) then
                ()
            else
                fontane-simple:transform($node/node())
158
                        
159
160
161
162
163
164
165
166
167
        case element(tei:handShift) return
            (: tei:handShift without @new denote the duct of the current 
            writer or a change of writing medium :)
            if($node/@new and (fontane-simple:is-hand-contemporary($node/@new) or
            $node/@new = "#Friedrich_Fontane") or not($node/@new))
                then
                    local:keep-element($node)
                else
                    ()
MRodz's avatar
MRodz committed
168
169
170
171
172
173
                
(:                if($node/following-sibling::*[1][count(./*) = 1 :)
(:                    and child::tei:fw or child::tei:stamp]):)
(:                    then:)
(:                        ():)
(:                    else:)
174
(:                        element tei:handShift {:)
MRodz's avatar
MRodz committed
175
176
(:                            $node/@*:)
(:                        }:)
177
178
        case element(tei:stamp) return
            ()
179
                
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
        case element(tei:seg) return
            if(count($node/*) = 1 and 
            ($node/child::tei:stamp or $node/child::tei:metamark[@function = "caret"]))
                then
                    ()
            else if(matches($node/@style, "underline") 
            and not(matches($node/@style, "vertical-align"))) then
                fontane-simple:transform($node/node())
            else if(contains($node/@style, "vertical-align")) then
                local:keep-element($node)                   
            else if($node/@type = "initials" or $node/@type = "monogram" 
            or $node/@style or $node/@type = "multiphrase")
                then
                    local:keep-element($node)
            else if($node/@type = "auction_number" 
            or $node/@type = "cancel"
MRodz's avatar
MRodz committed
196
197
            or $node/@type = "abort"
            or $node/@function ="unknown")
198
199
200
201
                then
                    ()
            else
                fontane-simple:transform($node/node())
MRodz's avatar
MRodz committed
202
203
                    
        case element(tei:hi) return
204
            local:keep-element($node)
205
206
        
        case element(tei:anchor) return
207
            local:keep-element($node)
208
                        
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
        case element(tei:surface) return 
            if(matches($node/@n, "cover"))
                then
                    (if(not($node/@n = "outer_front_cover"))
                        then
                            element tei:pb {
                                if($node/@n = "inner_front_cover" 
                                or $node/@n = "outer_back_cover")
                                    then
                                        attribute type {"flyleaf"} 
                                    else
                                        attribute type {"cover"}
                            }
                        else
                            (),
                    element tei:div {
                        attribute facs {$node/@facs},
                        attribute xml:id {$node/@xml:id},
                        attribute type {$node/@n},
                        fontane-simple:transform($node/node())
                    })
            else if(matches($node/@n, "[0-9]{1,4}[rv]{1}"))
                then
                    (element tei:pb{
                        attribute n {$node/@n},
                        if($node/@facs)
MRodz's avatar
MRodz committed
235
                            then
236
                                attribute facs {$node/@facs}
MRodz's avatar
MRodz committed
237
                            else
238
239
240
241
242
243
244
245
246
247
248
249
250
251
                                ()
                    }, 
                    (: if all descendant text nodes shoudln't be serialized, z.g. parent::fw or parent::stamp :)
                    if($node)
                        then
                            element tei:div{
                                attribute type {"page"},
                                fontane-simple:transform($node/node())
                            }
                        else
                            ())
            else if($node/@type = "label" and
                (contains($node/@subtype, "Fontane")
                or contains($node/@subtype, "Hersteller"))    
MRodz's avatar
MRodz committed
252
                )
253
254
255
                then
                    element tei:div {
                        attribute type {"label"},
MRodz's avatar
MRodz committed
256
                        fontane-simple:transform($node/node())
257
258
259
260
261
262
263
264
265
266
                    }
            else if($node/@type = "fragment" 
            and $node/descendant::text()[not(ancestor::tei:fw or ancestor::tei:stamp)])
                then
                    fontane-simple:transform($node/node())
                else
                    ()
                    
        case element(tei:gap) return
            local:keep-element($node)
267
                
268
269
270
271
272
273
        case element(tei:metamark) return
            if($node/@function = "integrate" 
            or $node/@function = "authorial_note") then
                element tei:ab {
                    $node/@*
                }
274
275
276
277
278
279
            else if($node/@function = "placeholder" 
            or $node/@function ="etc."
            or $node/@function ="caret"
            or $node/@function ="footnotes"
            or $node/@function ="ellipsis"
            or $node/@function = "paragraph") then
MRodz's avatar
MRodz committed
280
281
282
                element tei:ab {
                    attribute type {$node/@function},
                    fontane-simple:transform($node/node())
283
                }            
284
285
            else
                ()
286
287
288
289
290
        
        (: TODO: check if correct :)
        case element(tei:surplus) return
            ()
            
MRodz's avatar
MRodz committed
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
        case element(tei:zone) return
            if(matches($node/@style, "border-style:solid") 
            and not(matches($node/@style, "border-radius"))
            and not($node/@rend = "border-style:house")) then
                element tei:div {
                    attribute type {"frame"},
                    fontane-simple:transform($node/node())
                }
            else if(matches($node/@rend, "border-bottom-style:brace")) then
                (fontane-simple:transform($node/node()),
                element tei:ab {
                    attribute type {"bottom-brace"}
                })                    
            else if($node/@type = "cancel") then
                ()
            else if($node/@type = "marked_off") then
                element tei:seg {
                    $node/@type,
                    $node/@xml:id,
                    fontane-simple:transform($node/node())
                }
MRodz's avatar
MRodz committed
312
313
314
315
316
317
318
319
            else if($node/@type = "highlighted") then
                if($node/child::tei:zone[@type = "highlighted"]) then
                    fontane-simple:transform($node/node())
                else
                    element tei:hi {
                        attribute type {"vertical-mark"},
                        fontane-simple:transform($node/node())
                    }            
320
321
322
323
            else if(not($node/@xml:id)) then
                fontane-simple:transform($node/node())
            else if($node/@xml:id and fontane-simple:is-transposed($node)) then
                ()
MRodz's avatar
MRodz committed
324
325
            else
                fontane-simple:transform($node/node())
326
                
MRodz's avatar
MRodz committed
327
        case element(tei:figure) return
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
            if(count($node/child::*) = 1 and $node/child::tei:figDesc) then
                (: genealogy lines probably shouldn't be displayed, but I still have to 
                check that. in case they should be serialized, I leave the code :)
(:                if(matches($node/descendant::tei:ref, "Stammbaumverbindungslinie")) then:)
(:                    element tei:seg {:)
(:                        $node/@*,:)
(:                        fontane-simple:transform($node/node()):)
(:                    }:)
(:                else if(matches($node/descendant::tei:ref, "Schlusslinie")):)
                if(matches($node/descendant::tei:ref, "Schlusslinie"))
                    then
                        element tei:ab {
                            switch ($node/descendant::tei:ref)
                                case "horizontale einfache Schlusslinie" return 
                                    attribute type {"long-end-line"}                                    
                                case "Schlusslinie; horizontale Halbschleife von links oben nach rechts" return 
                                    attribute type {"long-end-line"}
                                case "horizontale einfache Schlusslinie (gewellt)" return 
                                    attribute type {"long-end-line-wavy"}
                                case "Schlusslinien; horizontale Schleife von links oben nach rechts unten" return 
                                    attribute type {"bottom-brace-short"}
                                default return 
                                    attribute type {"end-line"}  
                        }
                else if(matches($node/descendant::tei:ref, "Absatzlinie")
                (: in case of double paragraph lines the single lines are 
                encoded with "oberer" resp. "unterer Teil", but we only 
                serialize the encoding for the upper line :)
                and not(matches($node/descendant::tei:ref, "unterer Teil"))
                and not(matches($node/descendant::tei:figDesc, "unsicher"))) then
                    element tei:ab {
                        if(matches($node/descendant::tei:ref, "doppelt"))
                            then
                                attribute type {"short-paragraph-line-double"}
                            else
                                attribute type {"short-paragraph-line"}
                    }
                    else
                        ()
                else
368
                    local:keep-element($node)
369
                    
MRodz's avatar
MRodz committed
370
        case element(tei:note) return
371
372
            if($node/@type = "authorial" 
            and not($node/@subtype = "footnote")) then
MRodz's avatar
MRodz committed
373
374
                ()
            else 
375
                local:keep-element($node)
MRodz's avatar
MRodz committed
376
                    
MRodz's avatar
MRodz committed
377
        case element(tei:figDesc) return
378
            local:keep-element($node)
379
            
MRodz's avatar
MRodz committed
380
        case element(tei:ref) return
381
            local:keep-element($node)
MRodz's avatar
MRodz committed
382
383
        
        case element(tei:space) return
384
            local:keep-element($node)
MRodz's avatar
MRodz committed
385
386
387
388
389
390
391
392
393
            
        case element(tei:choice) return
            local:keep-element($node)
            
        case element(tei:abbr) return
            local:keep-element($node)
            
        case element(tei:expan) return
            local:keep-element($node)
394
395
396
397
398
399
            
        case element(tei:ptr) return
            let $target-id := substring-after($node/@target, "#")
            let $target := $node/ancestor::tei:TEI//*[matches(@xml:id, $target-id)]
            let $stripped-target := fontane-simple:strip-element($target)
            return fontane-simple:transform($stripped-target)
MRodz's avatar
MRodz committed
400
            
MRodz's avatar
MRodz committed
401
        default return
402
            fontane-simple:transform($node/node())
MRodz's avatar
MRodz committed
403
404
405
406
407
408
409
410
411
412
413
414
415
};

(: TODO: adapt to general XML :)
declare function fontane-simple:is-hand-contemporary($hand as xs:string) as xs:boolean {
    let $hand := replace($hand, "#", "")
    let $file := doc("/db/sade-projects/textgrid/data/xml/data/16b00.xml")
    let $handNote := $file//tei:handNote[@xml:id = $hand]
    return 
        if($handNote/@script = "contemporary")
            then
                true()
            else
                false()
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
};

declare function fontane-simple:is-transposed($node as node()) as xs:boolean {
    let $root := $node/ancestor::tei:TEI
    return
        if($root//tei:ptr[contains(@target, $node/@xml:id)]) then
            true()
        else
            false()
};

declare function fontane-simple:strip-element($node as node()) {
    element {$node/name()} {
        $node/@type,
        fontane-simple:transform($node/node())
    }
MRodz's avatar
MRodz committed
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
};

declare function local:prepare-text($node as text()) as text() {
    let $cleared-hyphen := 
        if(ends-with($node, "-") and not(local:keep-hyphen($node))) then
            text {functx:substring-before-last($node, "-")}
        else if(ends-with($node, "⸗") and not(local:keep-hyphen($node))) then
            text {functx:substring-before-last($node, "⸗")}
        else
            $node
    let $cleared-round-s := replace($cleared-hyphen, "ſ", "s")
    return text {$cleared-round-s}
};

declare function local:keep-hyphen($node as text()) as xs:boolean {
    if($node/parent::tei:line/following-sibling::*[1][self::tei:lb[@break = "keepHyphen"]])
        then
            true()
    else
        false()
};

declare function local:keep-element($node as node()) as node() {
    element {QName("http://www.tei-c.org/ns/1.0", $node/name())}{
        $node/@*,
        fontane-simple:transform($node/node())
    }
MRodz's avatar
MRodz committed
459
};