0
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
1 package de.mpg.mpiwg.berlin.mpdl.donatus.xmlrpc;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
2
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
3 import java.io.UnsupportedEncodingException;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
4 import java.util.ArrayList;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
5 import java.util.Collection;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
6 import java.util.HashMap;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
7 import java.util.Iterator;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
8
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
9 import de.mpg.mpiwg.berlin.mpdl.exception.ApplicationException;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
10
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
11 public class DonatusMorphologyDocument {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
12 private String docUri;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
13 private String language = DonatusConstants.DEFAULT_LANGUAGE;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
14 private HashMap<String, DonatusVariant> variants; // hold this variable for performance reasons: a key on each lemma variant
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
15 private HashMap<String, DonatusLemma> lemmas;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
16 private ArrayList<DonatusContextForm> contextForms;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
17 private byte[] documentBytes;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
18 private byte[] wtagBytes;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
19
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
20 public DonatusMorphologyDocument(String docUri) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
21 this.docUri = docUri;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
22 this.variants = new HashMap<String, DonatusVariant>();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
23 this.lemmas = new HashMap<String, DonatusLemma>();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
24 this.contextForms = new ArrayList<DonatusContextForm>();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
25 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
26
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
27 public ArrayList<DonatusLemma> getLemmas() {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
28 Collection<DonatusLemma> values = lemmas.values();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
29 ArrayList<DonatusLemma> retArrayList = new ArrayList<DonatusLemma>();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
30 Iterator<DonatusLemma> iter = values.iterator();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
31 while(iter.hasNext()) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
32 DonatusLemma lemma = iter.next();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
33 retArrayList.add(lemma);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
34 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
35 return retArrayList;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
36 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
37
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
38 public String getDocUri() {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
39 return docUri;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
40 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
41
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
42 public ArrayList<DonatusContextForm> getContextForms() {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
43 return contextForms;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
44 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
45
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
46 public void setWtagBytes(byte[] wtagBytes) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
47 this.wtagBytes = wtagBytes;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
48 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
49
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
50 public byte[] getWtagBytes() {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
51 return wtagBytes;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
52 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
53
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
54 public byte[] getDocumentBytes() throws ApplicationException {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
55 String contentXml = getXmlString();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
56 try {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
57 documentBytes = contentXml.getBytes("UTF-8");
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
58 } catch (UnsupportedEncodingException e) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
59 throw new ApplicationException(e);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
60 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
61 return documentBytes;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
62 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
63
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
64 public String getLemmaFormsString() {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
65 String result = "";
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
66 ArrayList<DonatusLemma> lemmas = getLemmas();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
67 for (int i=0; i<lemmas.size(); i++) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
68 DonatusLemma lemma = lemmas.get(i);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
69 String lemmaFormString = lemma.getForm();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
70 result += lemmaFormString + " ";
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
71 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
72 return result.trim();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
73 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
74
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
75 public ArrayList<DonatusVariant> getVariants() {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
76 ArrayList<DonatusVariant> result = new ArrayList<DonatusVariant>();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
77 ArrayList<DonatusLemma> lemmas = getLemmas();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
78 for (int i=0; i<lemmas.size(); i++) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
79 DonatusLemma lemma = lemmas.get(i);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
80 ArrayList<DonatusVariant> lemmaVariants = lemma.getVariants();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
81 result.addAll(lemmaVariants);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
82 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
83 return result;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
84 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
85
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
86 public String getLanguage() {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
87 return language;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
88 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
89
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
90 public void setLanguage(String language) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
91 this.language = language;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
92 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
93
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
94 public DonatusVariant getVariant(String variantForm) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
95 String hashKey = language + "###" + variantForm;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
96 DonatusVariant variant = variants.get(hashKey);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
97 return variant;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
98 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
99
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
100 public DonatusVariant putVariant(DonatusVariant variant) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
101 String variantForm = variant.getForm();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
102 String hashKey = language + "###" + variantForm;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
103 DonatusVariant putReturn = this.variants.put(hashKey, variant);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
104 return putReturn;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
105 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
106
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
107 public DonatusLemma getLemma(String lemmaForm) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
108 String hashKey = language + "###" + lemmaForm;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
109 DonatusLemma lemma = lemmas.get(hashKey);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
110 return lemma;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
111 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
112
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
113 public void putLemma(DonatusLemma lemma) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
114 String lemmaForm = lemma.getForm();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
115 String lang = lemma.getLanguage();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
116 String hashKey = lang + "###" + lemmaForm;
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
117 lemmas.put(hashKey, lemma);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
118 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
119
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
120 public void addContextForm(DonatusContextForm form) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
121 contextForms.add(form);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
122 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
123
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
124 public String getXmlString() {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
125 StringBuffer xmlString = new StringBuffer("<morphology>");
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
126 ArrayList<DonatusLemma> lemmas = getLemmas();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
127 for (int i=0; i<lemmas.size(); i++) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
128 DonatusLemma lemma = lemmas.get(i);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
129 String lemmaXmlString = lemma.getXmlString();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
130 xmlString.append(lemmaXmlString);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
131 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
132 for (int i=0; i<contextForms.size(); i++) {
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
133 DonatusContextForm contextForm = contextForms.get(i);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
134 String contextFormXmlString = contextForm.getXmlString();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
135 xmlString.append(contextFormXmlString);
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
136 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
137 xmlString.append("</morphology>");
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
138 return xmlString.toString();
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
139 }
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
140
|
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff
changeset
|
141 }
|