X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/fe763827605d48b2ad27d313cbcd84b588066070..8df0bad21a18b56933084288320fa35e2a3fce6f:/src/librarian/epub.py

diff --git a/src/librarian/epub.py b/src/librarian/epub.py
index fc5ee16..e2cdae7 100644
--- a/src/librarian/epub.py
+++ b/src/librarian/epub.py
@@ -15,6 +15,7 @@ from mimetypes import guess_type
 
 from ebooklib import epub
 from lxml import etree
+from PIL import Image
 from tempfile import mkdtemp, NamedTemporaryFile
 from shutil import rmtree
 
@@ -378,7 +379,7 @@ def remove_empty_lists_from_toc(toc):
 
 def transform(wldoc, verbose=False, style=None,
               sample=None, cover=None, flags=None, hyphenate=False,
-              ilustr_path='', output_type='epub'):
+              base_url='file://./', output_type='epub'):
     """ produces a EPUB file
 
     sample=n: generate sample e-book (with at least n paragraphs)
@@ -447,6 +448,14 @@ def transform(wldoc, verbose=False, style=None,
             output.add_item(item)
             spine.append(item)
 
+            toc[-1][1].append(
+                epub.Link(
+                    "part1.xhtml",
+                    "Początek utworu",
+                    "part1"
+                )
+            )
+
         elif wldoc.book_info.parts:
             # write title page for every parent
             if sample is not None and sample <= 0:
@@ -540,16 +549,18 @@ def transform(wldoc, verbose=False, style=None,
     output.set_identifier(six.text_type(document.book_info.url))
     output.set_language(functions.lang_code_3to2(document.book_info.language))
     output.set_title(document.book_info.title)
-    for author in document.book_info.authors:
+    for i, author in enumerate(document.book_info.authors):
         output.add_author(
             author.readable(),
-            file_as=six.text_type(author)
+            file_as=six.text_type(author),
+            uid='creator{}'.format(i)
         )
     for translator in document.book_info.translators:
         output.add_author(
             translator.readable(),
             file_as=six.text_type(translator),
-            role='translator'
+            role='trl',
+            uid='translator{}'.format(i)
         )
     for publisher in document.book_info.publisher:
         output.add_metadata("DC", "publisher", publisher)
@@ -567,23 +578,40 @@ def transform(wldoc, verbose=False, style=None,
 
     functions.reg_mathml_epub(output)
 
-    if os.path.isdir(ilustr_path):
-        ilustr_elements = set(ilustr.get('src')
-                              for ilustr in document.edoc.findall('//ilustr'))
-        for i, filename in enumerate(os.listdir(ilustr_path)):
-            if filename not in ilustr_elements:
-                continue
-            file_path = os.path.join(ilustr_path, filename)
-            with open(file_path, 'rb') as f:
-                output.add_item(
-                    epub.EpubItem(
-                        uid='image%s' % i,
-                        file_name=filename,
-                        media_type=guess_type(file_path)[0],
-                        content=f.read()
-                    )
-                )
+    # FIXME
+    for i, ilustr in enumerate(document.edoc.findall('//ilustr')):
+        url = six.moves.urllib.parse.urljoin(
+            base_url,
+            ilustr.get('src')
+        )
+        with six.moves.urllib.request.urlopen(url) as imgfile:
+            img = Image.open(imgfile)
+
+        th_format, ext, media_type = {
+            'GIF': ('GIF', 'gif', 'image/gif'),
+            'PNG': ('PNG', 'png', 'image/png'),
+        }.get(img.format, ('JPEG', 'jpg', 'image/jpeg'))
 
+        width = 1200
+        if img.size[0] < width:
+            th = img
+        else:
+            th = img.resize((width, round(width * img.size[1] / img.size[0])))
+
+        buffer = six.BytesIO()
+        th.save(buffer, format=th_format)
+
+        file_name = 'image%d.%s' % (i, ext)
+        ilustr.set('src', file_name)
+        output.add_item(
+            epub.EpubItem(
+                uid='image%s' % i,
+                file_name=file_name,
+                media_type=media_type,
+                content=buffer.getvalue()
+            )
+        )
+            
     # write static elements
 
     with open(get_resource('res/wl-logo-small.png'), 'rb') as f:
@@ -652,15 +680,6 @@ def transform(wldoc, verbose=False, style=None,
     toc, chunk_counter, chars, sample = transform_file(document, sample=sample)
     output.toc = toc[0][1]
 
-    if len(toc) < 2:
-        output.toc.append(
-            epub.Link(
-                "part1.xhtml",
-                "Początek utworu",
-                "part1"
-            )
-        )
-
     # Last modifications in container files and EPUB creation
     if len(annotations) > 0:
         output.toc.append(