src/sources/models.py

   1 import os
   2 import subprocess
   3 import uuid
   4 from django.apps import apps
   5 from django.conf import settings
   6 from django.db import models
   7 from django.urls import reverse
   8 from django.utils.timezone import now
   9 from django.utils.translation import gettext_lazy as _
  10 from . import conversion
  11 from . import document
  12 from . import utils
  13
  14
  15 class Source(models.Model):
  16     name = models.CharField(_('name'), max_length=1024)
  17     notes = models.TextField(blank=True, help_text=_('private'))
  18     wikisource = models.CharField(max_length=1024, blank=True)
  19     modified_at = models.DateTimeField(null=True, blank=True)
  20     processed_at = models.DateTimeField(null=True, blank=True)
  21
  22     def __str__(self):
  23         return self.name
  24
  25     def get_absolute_url(self):
  26         return reverse('source', args=[self.pk])
  27
  28     def touch(self):
  29         self.modified_at = now()
  30         self.save(update_fields=['modified_at'])
  31
  32     def get_upload_directory(self):
  33         return f'sources/upload/{self.pk}/'
  34
  35     def get_view_directory(self):
  36         return f'sources/view/{self.pk}/'
  37
  38     def get_ocr_directory(self):
  39         return f'sources/ocr/{self.pk}/'
  40
  41     def has_upload_files(self):
  42         d = os.path.join(settings.MEDIA_ROOT, self.get_upload_directory())
  43         return os.path.isdir(d) and os.listdir(d)
  44
  45     def get_view_files(self):
  46         d = self.get_view_directory()
  47         return [
  48             d + name
  49             for name in sorted(os.listdir(
  50                     os.path.join(settings.MEDIA_ROOT, d)
  51             ))
  52         ]
  53
  54     def has_view_files(self):
  55         d = os.path.join(settings.MEDIA_ROOT, self.get_view_directory())
  56         return os.path.isdir(d) and os.listdir(d)
  57
  58     def get_ocr_files(self):
  59         d = os.path.join(settings.MEDIA_ROOT, self.get_ocr_directory())
  60         return [
  61             d + name
  62             for name in sorted(os.listdir(d))
  63         ]
  64
  65     def has_ocr_files(self):
  66         d = os.path.join(settings.MEDIA_ROOT, self.get_ocr_directory())
  67         return os.path.isdir(d) and os.listdir(d)
  68
  69     def process(self):
  70         updir = os.path.join(
  71             settings.MEDIA_ROOT,
  72             self.get_upload_directory()
  73         )
  74         view_dir = os.path.join(
  75             settings.MEDIA_ROOT,
  76             self.get_view_directory()
  77         )
  78         ocr_dir = os.path.join(
  79             settings.MEDIA_ROOT,
  80             self.get_ocr_directory()
  81         )
  82         with utils.replace_dir(view_dir) as d:
  83             self.build_view_directory(updir, d)
  84         with utils.replace_dir(ocr_dir) as d:
  85             self.build_ocr_directory(updir, d)
  86         self.processed_at = now()
  87         self.save(update_fields=['processed_at'])
  88
  89     def build_view_directory(self, srcpath, targetpath):
  90         for source_file_name in os.listdir(srcpath):
  91             print(source_file_name)
  92             src = os.path.join(srcpath, source_file_name)
  93             ext = source_file_name.rsplit('.', 1)[-1].lower()
  94             if ext in ('png', 'jpg', 'jpeg'):
  95                 conversion.resize_image(src, targetpath)
  96                 # cp?
  97                 # maybe resize
  98             elif ext in ('tiff', 'tif'):
  99                 conversion.convert_image(src, targetpath)
 100             elif ext == 'pdf':
 101                 conversion.convert_pdf(src, targetpath)
 102             elif ext == 'djvu':
 103                 conversion.convert_djvu(src, targetpath)
 104             else:
 105                 pass
 106
 107     def build_ocr_directory(self, srcpath, targetpath):
 108         for source_file_name in os.listdir(srcpath):
 109             print(source_file_name)
 110             subprocess.run([
 111                 'tesseract',
 112                 os.path.join(srcpath, source_file_name),
 113                 os.path.join(targetpath, source_file_name),
 114                 '-l', 'pol'
 115             ])
 116
 117
 118 class BookSource(models.Model):
 119     book = models.ForeignKey('catalogue.Book', models.CASCADE)
 120     source = models.ForeignKey(Source, models.CASCADE)
 121     page_start = models.IntegerField(null=True, blank=True)
 122     page_end = models.IntegerField(null=True, blank=True)
 123
 124     class Meta:
 125         ordering = ('page_start',)
 126
 127     def __str__(self):
 128         return f'{self.source} -> {self.book}'
 129
 130     def get_absolute_url(self):
 131         return reverse('source_book_prepare', args=[self.pk])
 132
 133     def get_view_files(self):
 134         # TODO: won't work for PDFs.
 135         files = self.source.get_view_files()
 136         if self.page_end:
 137             files = files[:self.page_end]
 138         if self.page_start:
 139             files = files[self.page_start - 1:]
 140         return files
 141
 142     def get_ocr_files(self):
 143         # TODO: won't work for PDFs.
 144         files = self.source.get_ocr_files()
 145         if self.page_end:
 146             files = files[:self.page_end]
 147         if self.page_start:
 148             files = files[self.page_start - 1:]
 149         return files
 150
 151     def get_document(self):
 152         return self.book.document_books.first()
 153
 154     def prepare_document(self, user=None):
 155         DBook = apps.get_model('documents', 'Book')
 156         texts = document.build_document_texts(self)
 157
 158         dbook = self.get_document()
 159         if dbook is None:
 160             dbook = DBook.create(
 161                 user, texts[0],
 162                 title=self.book.title,
 163                 slug=str(uuid.uuid4()),
 164             )
 165         else:
 166             dbook[0].commit(text=texts[0], description='OCR', author=user)
 167         for text in texts[1:]:
 168             dbook[0].commit(text=text, description='OCR', author=user)
 169
 170         dbook[0].head.set_publishable(True)
 171         return dbook
 172