Nicer dictionary filtering.

[wolnelektury.git] / apps / dictionary / models.py

diff --git a/apps/dictionary/models.py b/apps/dictionary/models.py
index f2a9ddf..d0d8b69 100644 (file)
--- a/apps/dictionary/models.py
+++ b/apps/dictionary/models.py
@@ -2,30 +2,90 @@
 # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
 #
-from django.db import models
-
-from librarian import html
+from django.db import models, transaction
+from celery.task import task
 from sortify import sortify
+from celery.utils.log import get_task_logger
+
+task_logger = get_task_logger(__name__)
 
 from catalogue.models import Book
 
 
+class Qualifier(models.Model):
+    qualifier = models.CharField(max_length=128, db_index=True, unique=True)
+    name = models.CharField(max_length=255)
+
+    class Meta:
+        ordering = ['qualifier']
+
+    def __unicode__(self):
+        return self.name or self.qualifier
+
+
 class Note(models.Model):
-    book = models.ForeignKey(Book)
-    anchor = models.CharField(max_length=64)
+    """Represents a single annotation from a book."""
     html = models.TextField()
     sort_key = models.CharField(max_length=128, db_index=True)
+    fn_type = models.CharField(max_length=10, db_index=True)
+    qualifiers = models.ManyToManyField(Qualifier, null=True)
+    language = models.CharField(max_length=10, db_index=True)
 
     class Meta:
         ordering = ['sort_key']
 
 
-def notes_from_book(sender, **kwargs):
-    Note.objects.filter(book=sender).delete()
-    if sender.html_file:
-        for anchor, text_str, html_str in html.extract_annotations(sender.html_file.path):
-            Note.objects.create(book=sender, anchor=anchor,
-                               html=html_str, sort_key=sortify(text_str)[:128])
+class NoteSource(models.Model):
+    """Represents a single annotation from a book."""
+    note = models.ForeignKey(Note)
+    book = models.ForeignKey(Book)
+    anchor = models.CharField(max_length=64)
+
+    class Meta:
+        ordering = ['book']
+
+
+@task(ignore_result=True)
+def build_notes(book):
+    task_logger.info(book.slug)
+    with transaction.atomic():
+        book.notesource_set.all().delete()
+        if book.html_file:
+            from librarian import html
+            from librarian.fn_qualifiers import FN_QUALIFIERS
+
+            for anchor, fn_type, qualifiers, text_str, html_str in \
+                    html.extract_annotations(book.html_file.path):
+                sort_key = sortify(text_str).strip()[:128]
+
+                language = book.language
+                note = None
+                notes = Note.objects.filter(sort_key=sort_key,
+                    fn_type=fn_type,
+                    language=language, html=html_str)
+                if notes:
+                    note = notes[0]
+                else:
+                    note = Note.objects.create(
+                        sort_key=sort_key,
+                        html=html_str,
+                        fn_type=fn_type,
+                        language=language
+                        )
+
+                qualifier_objects = []
+                for qualifier in qualifiers:
+                    obj, created = Qualifier.objects.get_or_create(
+                        qualifier=qualifier, defaults={
+                            'name': FN_QUALIFIERS.get(qualifier, '')
+                        })
+                    qualifier_objects.append(obj)
+                note.qualifiers = qualifier_objects
+                note.notesource_set.create(book=book, anchor=anchor)
+
+        Note.objects.filter(notesource=None).delete()
+
 
-# always re-extract notes after making a HTML in a Book
+def notes_from_book(sender, instance, **kwargs):
+    build_notes.delay(instance)
 Book.html_built.connect(notes_from_book)