tidysimple.xqm 12.7 KB
Newer Older
1
2
3
4
5
6
7
8
xquery version "3.1";

(:~
 : This modules handles the conversion of the Fontante-TEI/XML into TEI simplePrint
 : for the edited text. The resulting TEI simplePrint is the basis for the "Editerter
 : Text" (edited text) view on the website and the book. It represents the latest
 : layer of text.
 :
MRodz's avatar
MRodz committed
9
 : @author Michelle Weidling
10
11
12
13
14
15
 : @version 0.1
 : @since TODO
 :)

module namespace tidySimple ="http://fontane-nb.dariah.eu/tidysimple";

MRodz's avatar
MRodz committed
16

17
declare namespace tei="http://www.tei-c.org/ns/1.0";
MRodz's avatar
MRodz committed
18
declare namespace test="http://exist-db.org/xquery/xqsuite";
19

MRodz's avatar
MRodz committed
20
import module namespace config="http://textgrid.de/ns/SADE/config" at "../../config/config.xqm";
21
import module namespace console="http://exist-db.org/xquery/console";
22
import module namespace functx="http://www.functx.com";
23
24
import module namespace simpleHelpers="http://fontane-nb.dariah.eu/teisimplehelpers" at "teisimplehelpers.xqm";

MRodz's avatar
MRodz committed
25

26
declare variable $tidySimple:valid-hands :=
27
28
    for $res in collection("/db/sade-projects/textgrid/data/xml/data")
(:    for $res in collection("/db/apps/SADE/resources/xml"):)
MRodz's avatar
MRodz committed
29
30
31
32
    return
        $res//tei:handNote[@script = "contemporary"]/@xml:id/string();


33
declare function tidySimple:main($tei as node()*) {
MRodz's avatar
MRodz committed
34
35
36
    let $tidy := tidySimple:sort-out-invalid-hands($tei//tei:text) 
        => tidySimple:sort-out-surplus-elements()
        => tidySimple:whitespaces()
MRodz's avatar
MRodz committed
37
(:    let $text-with-sections := tidySimple:make-structure($clear-surplus-hands):)
MRodz's avatar
MRodz committed
38
    let $header := $tei/tei:teiHeader
MRodz's avatar
MRodz committed
39
40
    return 
        xmldb:store("/db/apps/SADE/resources/xml/", "tei-simple-tmp.xml", <TEI xmlns="http://www.tei-c.org/ns/1.0">{$header}{$tidy}</TEI>)
41
42
43
44
45
46
47
48
49
(:    let $doc := doc($config:app-root || "/resources/xml/tei-simple-tmp.xml"):)
(:    let $tidied-text := tidySimple:tidy($doc//tei:text):)
(:    let $final-tei:=:)
(:        <TEI xmlns="http://www.tei-c.org/ns/1.0">:)
(:            {$tei//tei:teiHeader}:)
(:            {$tidied-text}:)
(:        </TEI>:)
(::)
(:    return xmldb:store("/db/apps/SADE/resources/xml/", "tei-simple.xml", $final-tei):)
MRodz's avatar
MRodz committed
50
51
};

52
53
54
55
56
declare function tidySimple:sort-out-invalid-hands($nodes as node()*) 
as node()* {
    for $node in $nodes return
        typeswitch ($node)
        case text() return
MRodz's avatar
MRodz committed
57
58
            if($node/preceding::tei:milestone[@unit = "handshift"][1] 
            and not(simpleHelpers:is-hand-valid($tidySimple:valid-hands, $node/preceding::tei:milestone[@unit = "handshift"][1]))) then
MRodz's avatar
MRodz committed
59
60
61
                ()
            else
                $node
62
        
63
64
65
66
        (: all lines have to be preserved because of the editorial commentary
        which references the lines in the notebooks. if we omitted @unit = "line"
        referencing wouldn't work any longer :)        
        case element(tei:milestone) return
MRodz's avatar
MRodz committed
67
            let $prev-handshift := $node/preceding::tei:milestone[@unit = "handshift"][1]
68
                return
MRodz's avatar
MRodz committed
69
                if($node/@unit = "handshift" and
70
71
72
                    simpleHelpers:is-hand-valid($tidySimple:valid-hands, $node)) then
                    tidySimple:clear-handshift($node)  
                    
MRodz's avatar
MRodz committed
73
                else if($node/@unit = "handshift") then
74
75
76
77
78
79
80
81
82
83
84
85
86
                    ()
                    
                else if($prev-handshift
                and $node/@unit = "line" 
                and not(simpleHelpers:is-hand-valid($tidySimple:valid-hands, $prev-handshift))) then
                    tidySimple:copy-element($node, "post")
                
                else if($prev-handshift
                and not(simpleHelpers:is-hand-valid($tidySimple:valid-hands, $prev-handshift))) then
                    ()
                
                else
                    tidySimple:copy-element($node, "post")
87
88
                
        default return
MRodz's avatar
MRodz committed
89
            let $prev-handshift := $node/preceding::tei:milestone[@unit = "handshift"][1]
90
91
92
93
94
95
            return
                if($prev-handshift
                and not(simpleHelpers:is-hand-valid($tidySimple:valid-hands, $prev-handshift))) then
                    ()
                else
                    tidySimple:copy-element($node, "post")
96
97
98
};


99
declare function tidySimple:sort-out-surplus-elements($nodes as node()*) 
100
101
102
103
104
105
as node()* {
    for $node in $nodes return
        typeswitch ($node)
        case text() return
            $node
        
106
        case element(tei:milestone) return
MRodz's avatar
MRodz committed
107
            if($node/@unit = "handshift") then
108
109
110
111
112
                if(simpleHelpers:is-prev-hand-same($node)
                or not(tidySimple:has-hand-text($node))) then
                    ()
                else
                    tidySimple:clear-handshift($node)
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
                    
            else if($node/@unit = "line") then
                let $next-pb := $node/following::tei:pb[1]
                let $nodes-on-page := $node/preceding::tei:pb[1]/following::node()[. << $next-pb][functx:node-kind(.) = "text"]
                let $contains-text :=
                    for $node-pb in $nodes-on-page return
                        if(normalize-space($node-pb) != "") then
                            true()
                        else 
                            false()
                return
                    if(functx:is-value-in-sequence(true(), $contains-text)) then
                        element {QName("http://www.tei-c.org/ns/1.0", $node/name())} {
                            $node/@*,
                            tidySimple:sort-out-surplus-elements($node/node())
                        }
                    else
                        ()
                
132
            else
133
134
                element {QName("http://www.tei-c.org/ns/1.0", $node/name())} {
                    $node/@*,
135
                    tidySimple:sort-out-surplus-elements($node/node())
136
                }                
137
138
                
        default return
139
140
            element {QName("http://www.tei-c.org/ns/1.0", $node/name())} {
                $node/@*,
141
                tidySimple:sort-out-surplus-elements($node/node())
142
            }
143
144
};

145
declare function tidySimple:has-hand-text($node as element(tei:milestone))
146
as xs:boolean {
MRodz's avatar
MRodz committed
147
    let $next-handshift := $node/following::tei:milestone[@unit = "handshift"][1]
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
    let $nodes-between := $node/following::node()[. << $next-handshift]
    let $is-text-node :=
        for $node-between in $nodes-between
        return 
            if ($node-between[self::text()]
            and not(normalize-space($node-between) = "")) then
                true()
            else
                false()
    return 
        if($next-handshift 
        and functx:is-value-in-sequence(true(), $is-text-node)) then
            true()
        else if(not($next-handshift)) then
            true()
        else
            false()
};

MRodz's avatar
MRodz committed
167

168
declare function tidySimple:copy-element($node as node(), $flag as xs:string)
169
 {
170
171
    element {QName("http://www.tei-c.org/ns/1.0", $node/name())}{
        $node/@*,
172
173
        if($flag = "structure") then
            tidySimple:make-structure($node/node())
174
175
176
177
        else if($flag = "post") then
            tidySimple:sort-out-invalid-hands($node/node())        
        else if($flag = "surplus") then
            tidySimple:sort-out-invalid-hands($node/node())
178
179
        else
            tidySimple:tidy($node/node())
180
181
182
    }
};

183
(:~
MRodz's avatar
MRodz committed
184
 : Purges surplus attributes from tei:milestone[@unit = "handshift"].
185
 : 
MRodz's avatar
MRodz committed
186
 : @author Michelle Weidling
MRodz's avatar
MRodz committed
187
188
 : @param $node the current tei:milestone[@unit = "handshift"]
 : @return the purged tei:milestone[@unit = "handshift"]
189
 :  :)
190
191
declare function tidySimple:clear-handshift($node as element(tei:milestone))
as element(tei:milestone) {
MRodz's avatar
MRodz committed
192
    element tei:milestone {
MRodz's avatar
MRodz committed
193
        attribute unit {"handshift"},
194
        if($node/@subtype = "") then
195
196
            ()
        else
197
198
            $node/@subtype,
        if($node/@rend = "") then
199
200
            ()
        else
201
            $node/@rend
202
203
    }
};
204
205


206
207
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
235
236
237
238
239
240
241
242
243
244
(:~ 
 : 
 :  :)
declare function tidySimple:make-structure($nodes as node()*) as node()* {
    for $node in $nodes return
        typeswitch ($node)
        case text() return
            if(tidySimple:is-in-section($node)) then
                ()
            else
                $node
                
        case element(tei:milestone) return 
            if($node/@unit = "section") then
                tidySimple:make-section($node)
                
            else if($node/@unit = "paragraph") then
                tidySimple:copy-element($node, "structure")
            
            (: lines in sections have to be ignored since they are copied into 
            the section with get-section-chunk :)    
            else if($node/@unit = "line"
            and tidySimple:is-in-section($node)) then
                ()
            else
                tidySimple:copy-element($node, "structure")
                
        case element(tei:anchor) return
            if (matches($node/preceding::tei:milestone/@spanTo, $node/@xml:id)) then
                ()
            else
                tidySimple:copy-element($node, "structure")
                
        default return 
            if(tidySimple:is-in-section($node)) then
                ()
            else
                tidySimple:copy-element($node, "structure")
};
245

246
247
248
249
(:~
 : Checks if a given node is between a tei:milestone[@unit = "section"] and its
 : corresponing tei:anchor.
 : 
MRodz's avatar
MRodz committed
250
 : @author Michelle Weidling
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
 : @param the current node
 : @return xs:boolean
 :  :)
declare function tidySimple:is-in-section($node as node()) as xs:boolean {
    let $prev-section-marker := $node/preceding::tei:milestone[1][@spanTo]
    let $target-id := replace($prev-section-marker, "#", "")
    let $next-anchor := $node/following::tei:anchor[matches(@xml:id, $target-id)]
    return
        if($prev-section-marker 
        and $next-anchor) then
            true()
        else
            false()
};


(:~
 : Creates a tei:div from a given tei:milestone[@unit = "section"] that contains 
 : all elements from a given tei:milestone to its corresponding tei:anchor.
 : 
 : Since the nodes between the current and the target one may also contain 
 : structural information (e.g. tei:milestone[@unit = "paragraph"]), these nodes
 : cannot simply be copied but have to be transformed as well.
 : 
MRodz's avatar
MRodz committed
275
 : @author Michelle Weidling
276
277
278
 : @param $node the current tei:milestone[@unit = "section"]
 : @return element(tei:div)
 :  :)
279
280
declare function tidySimple:make-section($node as element(tei:milestone)) 
as element(tei:div) {
MRodz's avatar
MRodz committed
281
    element tei:div {
282
283
284
285
286
287
288
289
290
291
        attribute type {"section"},
        tidySimple:get-section-chunk($node)
    }
};


(:~
 : Returns all nodes between a tei:milestone[@unit = "section"] and its 
 : corresponding tei:anchor.
 :
MRodz's avatar
MRodz committed
292
 : @author Michelle Weidling
293
294
295
296
297
298
299
300
 : @param the current tei:milestone element
 : @return all nodes between the current element and the respective tei:anchor
 :)
declare function tidySimple:get-section-chunk($node as element(tei:milestone)) 
as node()* {
    let $target-id := substring-after($node/@spanTo, "#")
    let $target := $node/following::tei:anchor[matches(@xml:id, $target-id)]
    return
301
        $node/following-sibling::node()[. << $target]
302
303
304
};


305
306
declare function tidySimple:tidy($nodes as node()*) as node()* {
    ()
MRodz's avatar
MRodz committed
307
308
};

MRodz's avatar
MRodz committed
309
declare function tidySimple:whitespaces($nodes as node()*) as node()* {
MRodz's avatar
MRodz committed
310
311
312
313
314
    for $node in $nodes return
        typeswitch($node)
        case text() return
            $node
            
MRodz's avatar
MRodz committed
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
        default return
            element {QName("http://www.tei-c.org/ns/1.0", $node/name())} {
                $node/@*,
                if($node/@next) then
                    let $target-id := replace($node/@next, "#", "")
                    let $target := $node/ancestor::tei:text//*[@xml:id = $target-id]
                    return
                        (if(not(ends-with($node//text()[last()], "@P"))
                        and ($node/following::*[. << $target][self::tei:milestone[@unit = "line"]]
                            or $target/following::*[. << $node][self::tei:milestone[@unit = "line"]])) then
                            attribute break {"true"}
                        else
                            (),
                        if($node/following::text()[. << $target][not(normalize-space(.) = "")]
                            or $target/following::text()[. << $node][not(normalize-space(.) = "")]) then
                            attribute text-inbetween {"true"}
                        else
                            ())
                else if($node/@prev) then
                    let $target-id := replace($node/@prev, "#", "")
                    let $target := $node/ancestor::tei:text//*[@xml:id = $target-id]
                    return
                        if($node/following::text()[. << $target][not(normalize-space(.) = "")]
                            or $target/following::text()[. << $node][not(normalize-space(.) = "")]) then
                            attribute text-inbetween {"true"}
                        else
                            ()        
                else
                    (),
                tidySimple:whitespaces($node/node())
            }
346
};