Flere ord som bør med.

[text-mekanikerord.git] / make-glossary

diff --git a/make-glossary b/make-glossary
index 51c94bf291b092896226ead78fe549168ba3c88f..6e36003a21f81ba14ded2f503322c535c7ce6353 100755 (executable)
--- a/make-glossary
+++ b/make-glossary
@@ -5,6 +5,8 @@ import locale
 from lxml import etree
 from lxml.etree import tostring
 
+import json
+
 list_topic = False
 
 filemakerxml = 'meksme-utf8.xml'
@@ -16,7 +18,7 @@ root = tree.getroot()
 #print(tostring(tree))
 
 cols = (
-    'topic', 'sme', 'desc-sme', 'desc-nb', 'nb', 'sv', 'fi', 'en', 'is',
+    'topic', 'se', 'desc-se', 'desc-nb', 'nb', 'sv', 'fi', 'en', 'is',
 )
 
 topicmap = {
@@ -55,11 +57,15 @@ for row in resultset.getchildren():
         index += 1
     #print(d)
     words.append(d)
+
+    with open('meksme-utf8.json', 'w') as f:
+        json.dump(words, f)
+
 def langsort(lang, e):
     if lang in e:
         return locale.strxfrm(e[lang])
     else:
-        return locale.strxfrm(e['sme'])
+        return locale.strxfrm(e['se'])
 
 def make_glossary_docbook(lang, desccodes, langcodes, output='glossary.xml'):
     import lxml.builder
@@ -69,6 +75,16 @@ def make_glossary_docbook(lang, desccodes, langcodes, output='glossary.xml'):
         }
     )
 
+    def word2id(word):
+        return word \
+                .replace('[', '_') \
+                .replace(']', '_') \
+                .replace('(', '_') \
+                .replace(')', '_') \
+                .replace('/', '_') \
+                .replace('\'', '_') \
+                .replace(' ', '_')
+
     def indexit(entry, wlist, lang=None):
         for w in wlist.split(","):
             if "" != w:
@@ -76,6 +92,7 @@ def make_glossary_docbook(lang, desccodes, langcodes, output='glossary.xml'):
                     w += "[%s]" % lang
                 entry.append(E.indexterm(E.primary(w)))
     ids = {}
+    redirects = {}
     glossary = E.glossary()
     for e in sorted(words, key=lambda x: langsort(lang, x)):
         ldesc = 'desc-%s' % lang
@@ -83,19 +100,52 @@ def make_glossary_docbook(lang, desccodes, langcodes, output='glossary.xml'):
             e['topic'] = topicmap[lang][e['topic']]
         if lang in e:
             w = e[lang].split(',')
-            id = w[0] \
-                .replace('[', '_') \
-                .replace(']', '_') \
-                .replace('(', '_') \
-                .replace(')', '_') \
-                .replace('/', '_') \
-                .replace(' ', '_')
+            id = word2id(w[0])
             while id in ids:
                 id = id + 'x'
             ids[id] = True
-            if ldesc not in e:
-                print("warning: %s missing %s description" % (e[lang], lang))
+
+            # First handle redirections with not extra info
+            if -1 != e[lang].find('>') and ldesc not in e:
+                p = e[lang].split(' > ')
+                if p[0] in redirects: # Skip if already added
+                    continue
+                if -1 == p[1].find(','):
+                    if '-' == p[1][-1]:
+                        print("warning: Skipping dangling reference %s -> %s" %
+                              (p[0], p[1]))
+                    else:
+                        seeentry = E.glossentry()
+                        seeentry.append(E.glossterm(p[0]))
+                        id = word2id(p[1])
+                        seeentry.append(E.glosssee(otherterm=id))
+                        glossary.append(seeentry)
+                        redirects[p[0]] = id
+                else:
+                  print("warning: skipping split refererence %s -> %s" %
+                        (p[0], p[1]))
+                  if False: # Not allowed in docbook
+                    seeentry = E.glossentry()
+                    seeentry.append(E.glossterm(p[0]))
+                    for s in p[1].split(','):
+                        s = s.strip().lstrip()
+                        seeentry.append(E.glosssee(otherterm=word2id(s)))
+                    glossary.append(seeentry)
                 continue
+
+            # Add See also entries pointing to main entry
+            if 1 < len(w):
+                for t in w[1:]:
+                    t = t.strip().lstrip()
+                    if t not in redirects:
+                    #print("info: Adding see also entry for %s" % t)
+                        seeentry = E.glossentry()
+                        seeentry.append(E.glossterm(t))
+                        seeentry.append(E.glosssee(otherterm=id))
+                        glossary.append(seeentry)
+                        redirects[t] = id
+            elif ldesc not in e:
+                print("warning: term %s missing primary language %s description" % (e[lang], lang))
             entry = E.glossentry(id=id)
             if list_topic and 'topic' in e:
                 entry.append(E.glossterm('%s [%s]' % (e[lang], e['topic'])))
@@ -110,23 +160,16 @@ def make_glossary_docbook(lang, desccodes, langcodes, output='glossary.xml'):
                     indexit(entry, e[l], l)
             if "" != lstr:
                 entry.append(E.glossdef(E.para(lstr)))
+            else:
+                # only single word witout translations, skip it
+                continue
             for desccode in desccodes:
                 codestr = 'desc-%s' % desccode
                 if codestr in e:
-                    entry.append(E.glossdef(E.para("%s: %s" % (desccode,
+                    entry.append(E.glossdef(E.para("(%s): %s" % (desccode,
                                                                e[codestr]))))
             glossary.append(entry)
 
-            # Add See also entries pointing to main entry
-            if 1 < len(w):
-                for t in w[1:]:
-                    t = t.strip().lstrip()
-                    entry = E.glossentry()
-                    entry.append(E.glossterm(t))
-                    # FIXME
-                    entry.append(E.glosssee(otherterm=id))
-                glossary.append(entry)
-
     def glosstermlocale(x):
         # Look up glossterm (FIXME figure out more robust way)
         t = x.getchildren()[0].text
@@ -136,6 +179,9 @@ def make_glossary_docbook(lang, desccodes, langcodes, output='glossary.xml'):
             return ""
     # Sort list to mix seealso entries into their correct location.
     glossary[:] = sorted(glossary, key=glosstermlocale)
+
+    l = len(glossary)
+    print("info: dictionary contain %d entries" % l)
     
     content = lxml.etree.tostring(glossary,
                                   pretty_print=True,
@@ -157,19 +203,19 @@ if 'nb' == args.langcode:
     print("Norsk/bokmål")
     print()
     make_glossary_docbook(lang='nb', desccodes=('nb',),
-                          langcodes=('en', 'sme', 'sv', 'da', 'fi', 'is',),
+                          langcodes=('en', 'se', 'sv', 'da', 'fi', 'is',),
                           output=args.output)
-elif 'sme' == args.langcode:
+elif 'se' == args.langcode:
     print("Nordsamisk")
     print()
-    make_glossary_docbook(lang='sme', desccodes=('sme', 'nb'),
+    make_glossary_docbook(lang='se', desccodes=('se', 'nb'),
                           langcodes=('nb', 'en', 'sv', 'da', 'fi', 'is',),
                           output=args.output)
 elif 'en' == args.langcode:
     print("Engelsk")
     print()
-    make_glossary_docbook(lang='en', desccodes=('en', 'nb'),
-                          langcodes=('en', 'nb', 'sme', 'sv', 'da', 'fi', 'is',),
+    make_glossary_docbook(lang='en', desccodes=('en'),
+                          langcodes=('en', 'nb', 'se', 'sv', 'da', 'fi', 'is',),
                           output=args.output)
 else:
     print("error: Unknown language code %s" % args.langcode)