# -*- coding: utf-8 -*- import uuid import StringIO try: import Image except ImportError: from PIL import Image from django.core.files.base import ContentFile from ooxml.serialize import HeaderContext from booktype.utils import config from booktype.utils.misc import import_from_string def convert_image(image_type, content): img = Image.open(StringIO.StringIO(content.read())) out = StringIO.StringIO() img.save(out, format='PNG') data = out.getvalue() out.close() return ContentFile(data) def get_importer_class(): """ Dummy function to return the correct module to import DOCX files. If there is no custom class specified in client instance, it will use .docximporter.WordImporter class as it is in constants.py file """ DOCX_IMPORTER_CLASS = config.get_configuration('DOCX_IMPORTER_CLASS') ImporterClass = import_from_string(DOCX_IMPORTER_CLASS) return ImporterClass class DocHeaderContext(object, HeaderContext): def is_header(self, elem, font_size, node, style=None): """ Used for checking if specific element is a header or not. :Returns: True or False """ HEADING_STYLES = config.get_configuration('DOCX_HEADING_STYLES', []) # Check the defined list of styles if style: if style.style_id.lower() in HEADING_STYLES: return True if elem.rpr.get('style', None): for style_key in HEADING_STYLES: if elem.rpr.get('style').lower().replace('-', '').startswith(style_key): return True for e in elem.elements: if hasattr(e, 'rpr'): if e.rpr.get('style', None): for style_key in HEADING_STYLES: if e.rpr.get('style').lower().replace('-', '').startswith(style_key): return True return super(DocHeaderContext, self).is_header(elem, font_size, node, style) def get_header(self, elem, style, node): """ Returns HTML tag representing specific header for this element. :Returns: String representation of HTML tag. """ STYLES_TUPLE = config.get_configuration('DOCX_HEADING_STYLES_TUPLE') if style and not isinstance(style, int): for header_id, header_values in STYLES_TUPLE: if style.style_id.lower().replace('-', '') in header_values: return header_id if elem.rpr.get('style', None): for header_id, header_values in STYLES_TUPLE: for style_key in header_values: if elem.rpr.get('style').lower().replace('-', '').startswith(style_key): return header_id for e in elem.elements: if hasattr(e, 'rpr'): if e.rpr.get('style', None): for header_id, header_values in STYLES_TUPLE: for style_key in header_values: if e.rpr.get('style').lower().replace('-', '').startswith(style_key): return header_id return super(DocHeaderContext, self).get_header(elem, style, node) def serialize_empty(ctx, document, elem, root): return root def hook_p(ctx, document, el, elem): """ Hook for the paragraph elements. Add different classes used in the editor for different paragraph elements. """ P_STYLES = config.get_configuration('DOCX_PARAGRAPH_STYLES_MAP', {}) if hasattr(el, 'style_id'): class_name = P_STYLES.get(el.style_id, None) if class_name: elem.set('class', class_name) def hook_footnote(ctx, document, el, elem): """ Hook for the footnotes elements. Add custom footnote booktype class and remove link tag. """ data_id = str(uuid.uuid1()).replace('-', '') # set endnote class because we use it later to fix all notes elem.set('class', 'endnote') elem.set('data-id', data_id) # this will be used later to retrieve content of footnote elem.set('data-relationship', 'footnotes') elem.set('data-relation-id', el.rid) def hook_endnote(ctx, document, el, elem): """ Hook for the endnotes elements. Add custom endnote booktype class and remove link tag. """ data_id = str(uuid.uuid1()).replace('-', '') elem.set('class', 'endnote') elem.set('data-id', data_id) # this will be used later to retrieve content of endnote elem.set('data-relationship', 'endnotes') elem.set('data-relation-id', el.rid) def check_h_tags_hook(ctx, document, el, elem): """ This hook just checks if the are H tags with some wrong classes like H tags with paragraps classes which is wrong. """ P_STYLES = config.get_configuration('DOCX_PARAGRAPH_STYLES_MAP', {}) if hasattr(el, 'style_id'): class_name = P_STYLES.get(el.style_id, None) # if h tag was found with paragraps class, let's switch tag if class_name: elem.tag = 'p' elem.set('class', class_name) def hook_infobox_table(ctx, document, el, elem): """ To be a valid infobox, the table should only have two rows and one column. Besides that the first row should contain a paragraph with the style of infoboxcaption, otherwhise it should be taken as normal table Here is a table infobox example (attention: it is on html, not docx format)

Title or caption text here

Content of the box, the style/class here is optional This might also contain a cite.

""" if len(el.rows) != 2: return thead, tbody = elem try: tcaption = thead[0][0] except IndexError: return if 'infobox' not in tcaption.get('class', ''): return elem.tag = 'div' elem.set('class', 'bk-box') thead.tag = 'div' thead.set('class', 'box-caption') tbody.tag = 'div' tbody.set('class', 'box-content') def clean_infobox_content(tree): """ Clean infobox content from unneccesary tags or headers removes td orphan tags and tags of first level of children deep :Args: - tree (:class:`lxml.html`): HTML tree parsed from string """ # let's first get rid of all td orphans tags from table for td in tree.xpath('//div[contains(@class, "bk-box")]//td'): td.drop_tag() # then we should remove all unneccesary tags inside each bk-box for box in tree.xpath('//div[contains(@class, "bk-box")]'): # only first level of children deep for child in box.xpath('//div[contains(@class, "box-caption")]/*'): child.drop_tag() for child in box.xpath('//div[contains(@class, "box-content")]/*'): child.drop_tag() def fix_citations(tree): """ Sometimes citations are inline elements so the hook for paragraph is not able to match the styles to set the right class and tag. This is just for citations for now. We should improve this logic later :Args: - tree (:class:`lxml.html`): HTML tree parsed from string """ for cite in tree.xpath('//*[contains(@class, "citation")]'): cite.tag = 'p' cite.set('class', 'bk-cite')