用 python-reportlab 将 rst 转换为中文 PDF

Transcription

1 用 python-reportlab 将 rst 转换为中文 PDF Author Keywords License Requires Jiahua Huang RST, PDF LGPL docutils, reportlab python 里用 Reportlab 应该比用 tex docbook fop 合适, 不过 python-reportlab 默认没中文字体, 也没中文换行, 分析 实际上新的 python-reportlab 已经有能力处理中文字体, 也有了一个基本的 CJK Warp (breaklinescjk), 只是通常使用 python-reportlab 的程序都没有去注意 现有 reportlab 程序主要问题就是 1. 使用默认英文字体, 而没考虑中文字 2. 使用默认英文折行, 而没考虑中文长行 解决 那么, 针对现有问题, 可以考虑如下方向解决 问题没有中文字体默认不是中文字体默认不是 CJK 换行 解决注册中文字体文泉驿正黑覆盖选字处理, 一律用文泉驿正黑覆盖折行处理, 一律尝试 breaklinescjk 动手吧 拿把斧头直接砍 python-reportlab 当然省事, 不过系统的库修改了以后维护还真是麻烦, 能不能不修改库本身呢? 既然 Python 是动态语言, 可以运行中动态修改几乎一切东西, 那么运行时修改 reportlab 的类吧, 只要在最终调用 reportlab 生成 PDF 前动态修改到自己要的就行 注册字体 因为主要是自己用, 就硬编码使用自己喜欢的文泉驿正黑字体吧, 用 reportlab.pdfbase.pdfmetrics.registerfont 注册文泉驿 import reportlab.rl_config reportlab.rl_config.warnonmissingfontglyphs = 0 import reportlab.pdfbase.pdfmetrics import reportlab.pdfbase.ttfonts reportlab.pdfbase.pdfmetrics.registerfont(reportlab.pdfbase.ttfonts.ttfont('song', '/usr/share/fonts/truetype/wqy/wqy-zenhei.ttf')) 改默认字体 先好歹让中文能显示吧, 不如暂时让所有字体都用文泉驿 ( 刚才把文泉驿注册为 "song", 所以这里用 "song" 字体 ) 修改 reportlab.lib.fonts.ps2tt 和 reportlab.lib.fonts.tt2ps import reportlab.lib.fonts reportlab.lib.fonts.ps2tt = lambda psfn ('song', 0, 0) reportlab.lib.fonts.tt2ps = lambda fn,b,i 'song'

2 中文换行 为了省事, 就不考虑哪些要 CJK 折行哪些不要了, 直接 try CJK 换行, except 再英文换行 修改 reportlab.platypus.paragraph.wrap import reportlab.platypus def wrap(self, availwidth, availheight) # work out widths array for breaking self.width = availwidth leftindent = self.style.leftindent first_line_width = availwidth - (leftindent+self.style.firstlineindent) - self.style.rightindent later_widths = availwidth - leftindent - self.style.rightindent try self.blpara = self.breaklinescjk([first_line_width, later_widths]) except self.blpara = self.breaklines([first_line_width, later_widths]) self.height = len(self.blpara.lines) * self.style.leading return (self.width, self.height) reportlab.platypus.paragraph.wrap = wrap 完整的一段 就是将上边 3 个合起来 import reportlab.rl_config reportlab.rl_config.warnonmissingfontglyphs = 0 import reportlab.pdfbase.pdfmetrics import reportlab.pdfbase.ttfonts reportlab.pdfbase.pdfmetrics.registerfont(reportlab.pdfbase.ttfonts.ttfont('song', '/usr/share/fonts/truetype/wqy/wqy-zenhei.ttf')) import reportlab.lib.fonts reportlab.lib.fonts.ps2tt = lambda psfn ('song', 0, 0) reportlab.lib.fonts.tt2ps = lambda fn,b,i 'song' ## for CJK Wrap import reportlab.platypus def wrap(self, availwidth, availheight) # work out widths array for breaking self.width = availwidth leftindent = self.style.leftindent first_line_width = availwidth - (leftindent+self.style.firstlineindent) - self.style.rightindent later_widths = availwidth - leftindent - self.style.rightindent try self.blpara = self.breaklinescjk([first_line_width, later_widths]) except self.blpara = self.breaklines([first_line_width, later_widths]) self.height = len(self.blpara.lines) * self.style.leading return (self.width, self.height) reportlab.platypus.paragraph.wrap = wrap 使用 具体的说了那么多, 那实际怎么做呢? 前面说了, Python 是动态语言, 你可以将上边的代码随便插入到你的任何程序或模块里, 只要在实际调用 reportlab 生成 PDF 前就可以了 1. 直接点, 在你的程序 import reportlab 前插入上边的代码 2. 如果想让你原先的文件不要插入这突兀的语句, 那可以将上边的代码写入到如 cjkreportlab.py, 在你需要用 reportlab 做中文 pdf 的代码里加个 import cjkreportlab 例子 这里给个修改的 rst2pdf.py #!/usr/bin/env python # -*- coding utf-8 -*- import docutils.core,docutils.nodes,sys,re import sys import pprint from types import StringType from docutils import version, version_details, SettingsSpec from docutils import frontend, io, utils, readers, writers from docutils.frontend import OptionParser from docutils.transforms import Transformer

3 import docutils.readers.doctree from urlparse import * from reportlab.platypus import * #from reportlab.platypus.para import Paragraph,FastPara,Para from reportlab.pdfbase.pdfmetrics import stringwidth from reportlab.lib.enums import * import reportlab.lib.colors as colors from reportlab.lib.units import * from reportlab.lib.pagesizes import * from copy import copy from cgi import escape try from wordaxe.rl.paragraph import Paragraph from wordaxe.rl.styles import ParagraphStyle, getsamplestylesheet except print "No support for hyphenation, install wordaxe" ## 就是在这里添加前面说的代码啦 ############################ import reportlab.rl_config reportlab.rl_config.warnonmissingfontglyphs = 0 import reportlab.pdfbase.pdfmetrics import reportlab.pdfbase.ttfonts reportlab.pdfbase.pdfmetrics.registerfont(reportlab.pdfbase.ttfonts.ttfont('song', '/usr/share/fonts/truetype/wqy/wqy-zenhei.ttf')) import reportlab.lib.fonts reportlab.lib.fonts.ps2tt = lambda psfn ('song', 0, 0) reportlab.lib.fonts.tt2ps = lambda fn,b,i 'song' ## for CJK Wrap import reportlab.platypus def wrap(self, availwidth, availheight) # work out widths array for breaking self.width = availwidth leftindent = self.style.leftindent first_line_width = availwidth - (leftindent+self.style.firstlineindent) - self.style.rightindent later_widths = availwidth - leftindent - self.style.rightindent try self.blpara = self.breaklinescjk([first_line_width, later_widths]) except self.blpara = self.breaklines([first_line_width, later_widths]) self.height = len(self.blpara.lines) * self.style.leading return (self.width, self.height) reportlab.platypus.paragraph.wrap = wrap import sys reload(sys) sys.setdefaultencoding('utf8') ######################################################## ## style ########################################################################################## # Demo stylesheet module.for rst2pdf from reportlab.platypus import * import reportlab.lib.colors as colors from reportlab.lib.units import * from reportlab.pdfbase.ttfonts import TTFont from reportlab.lib.fonts import addmapping from reportlab.lib.styles import * from reportlab.lib.enums import * from reportlab.pdfbase import pdfmetrics try from wordaxe.rl.paragraph import Paragraph from wordaxe.rl.styles import ParagraphStyle,getSampleStyleSheet except print "No hyphenation support install wordaxe" import os # Set these, and then **maybe** think of setting the stylesheets below # if you want finer control stdfont = 'song' stdbold = 'song' stditalic = 'song' stdbolditalic = 'song' stdmono = 'song' # You can embed your own fonts and use them later if you want if os.path.isfile('wrabbit/whitrabt.ttf')

4 pdfmetrics.registerfont(ttfont('whiterabbit', 'wrabbit/whitrabt.ttf')) addmapping('whiterabbit', 0, 0, 'WhiteRabbit') #normal addmapping('whiterabbit', 0, 1, 'WhiteRabbit') #italic addmapping('whiterabbit', 1, 0, 'WhiteRabbit') #bold addmapping('whiterabbit', 1, 1, 'WhiteRabbit') #italic and bold stdmono = 'WhiteRabbit' def getstylesheet() """Returns a stylesheet object""" stylesheet = StyleSheet1() stylesheet.add(paragraphstyle(name='normal', fontname=stdfont, bulletfontname=stdfont, fontsize=10, bulletfontsize=10, leading=12, language='en', hyphenation=true )) stylesheet.add(paragraphstyle(name='bodytext', spacebefore=6, language='en', hyphenation=true )) stylesheet.add(paragraphstyle(name='footer', backcolor='#efefef', alignment=ta_center) ) stylesheet.add(paragraphstyle(name='attribution', parent=stylesheet['bodytext'], alignment=ta_right) ) stylesheet.add(paragraphstyle(name='fieldname', parent=stylesheet['bodytext'], alignment=ta_right, fontname=stdbold, bulletfontname=stdbold, ) ) stylesheet.add(paragraphstyle(name='rubric', parent=stylesheet['bodytext'], textcolor=colors.darkred, alignment=ta_center) ) stylesheet.add(paragraphstyle(name='italic', parent=stylesheet['bodytext'], fontname = stditalic, bulletfontname=stditalic) ) stylesheet.add(paragraphstyle(name='title', fontname = stdbold, bulletfontname=stdbold, fontsize=18, bulletfontsize=18, leading=22, alignment=ta_center, spaceafter=6), alias='title') stylesheet.add(paragraphstyle(name='subtitle', parent=stylesheet['title'], fontsize=14, bulletfontsize=14), alias='subtitle') stylesheet.add(paragraphstyle(name='heading1', fontname = stdbold, bulletfontname=stdbold, fontsize=18, bulletfontsize=18, leading=22, keepwithnext=true, spaceafter=6), alias='h1')

5 stylesheet.add(paragraphstyle(name='heading2', fontname = stdbold, bulletfontname=stdbold, fontsize=14, bulletfontsize=14, leading=18, spacebefore=12, keepwithnext=true, spaceafter=6), alias='h2') stylesheet.add(paragraphstyle(name='heading3', fontname = stdbolditalic, bulletfontname=stdbolditalic, fontsize=12, bulletfontsize=12, leading=14, spacebefore=12, keepwithnext=true, spaceafter=6), alias='h3') stylesheet.add(paragraphstyle(name='heading4', fontname = stdbolditalic, bulletfontname=stdbolditalic, fontsize=12, bulletfontsize=12, leading=14, spacebefore=12, keepwithnext=true, spaceafter=6), alias='h4') stylesheet.add(paragraphstyle(name='bullet', firstlineindent=0, spacebefore=3), alias='bu') stylesheet.add(paragraphstyle(name='definition', firstlineindent=0, leftindent=36, bulletindent=0, spacebefore=6, fontname=stdbold, bulletfontname=stdbold), alias='df') stylesheet.add(paragraphstyle(name='code', fontname=stdmono, bulletfontname=stdmono, fontsize=8, bulletfontsize=8, leading=8.8, firstlineindent=0, leftindent=12, spacebefore=4)) return stylesheet # Some table styles used for pieces of the document tstyles={} # Used for regular tables tstylenorm = [ ('VALIGN',(0,0),(-1,-1),'TOP'), ('INNERGRID', (0,0), (-1,-1), 0.25, colors.black), ('BOX', (0,0), (-1,-1), 0.25, colors.black), ] # Header row in tables tstylehead = ('BACKGROUND',(0,0),(-1,0),colors.yellow) tstyles['normal']=tablestyle(tstylenorm) # Used for field lists

6 tstyles['field']=tablestyle([ ('VALIGN',(0,0),(-1,-1),'TOP'), ('ALIGNMENT',(0,0),(1,-1),'RIGHT'), ]) fieldlist_lwidth=3*cm # Used for endnotes tstyles['endnote']=tablestyle([ ('VALIGN',(0,0),(-1,-1),'TOP'), ('ALIGNMENT',(0,0),(1,-1),'RIGHT'), ]) # Left column of the endnote. The content of the note takes the rest of # the available space endnote_lwidth=2*cm # Used for sidebars tstyles['sidebar']=tablestyle([ ('VALIGN',(0,0),(-1,-1),'TOP'), ('BACKGROUND',(0,0),(-1,-1),colors.lightyellow), ]) ################################################################################################### styles=getstylesheet() try import wordaxe from wordaxe.pyhnjhyphenator import PyHnjHyphenator try wordaxe.hyphregistry['en'] = PyHnjHyphenator('en_US',5) except wordaxe.hyphregistry['en'] = PyHnjHyphenator('en_US',5,purePython=True) except print "No support for hyphenation, install wordaxe" # This is A4 paper, change it as you wish pw,ph=ps=a4 # Margins tm=2*cm bm=2*cm lm=3*cm rm=1.5*cm # tw is the text width. # We need it to calculate header-footer height # and compress literal blocks. tw=pw-lm-rm page_break_level=0 marks="#=-_*^>%& " lowerroman=['i','ii','iii','iv','v','vi','vii','viii','ix','x','xi'] loweralpha="abcdefghijklmnopqrstuvwxyz" class MyIndenter(Indenter) # Bugs in reportlab? def draw(self) pass width=0 height=0 def depth (node) if node.parent==none return 0 return 1+depth(node.parent) decoration = {'header'none, 'footer'none, 'endnotes'[]} def gather_pdftext (node, depth, in_line_block=false,replaceent=true) return ''.join([gen_pdftext(n,depth,in_line_block,replaceent) for n in node.children ]) def gen_pdftext(node, depth, in_line_block=false,replaceent=true) pre="" post="" if isinstance (node, docutils.nodes.paragraph) \ or isinstance (node, docutils.nodes.title) \ or isinstance (node, docutils.nodes.subtitle) \ +"\n" elif isinstance (node, docutils.nodes.text) node.pdftext=node.astext()

7 elif isinstance (node, docutils.nodes.strong) pre="<b>" post="</b>" elif isinstance (node, docutils.nodes.emphasis) pre="<i>" post="</i>" elif isinstance (node, docutils.nodes.literal) pre='<font face="%s">'%styles['code'].fontname post="</font>" elif isinstance (node, docutils.nodes.superscript) pre='<super>' post="</super>" elif isinstance (node, docutils.nodes.subscript) pre='_{' post="}" elif isinstance (node, docutils.nodes.title_reference) # FIXME needs to work as a link pre='<font color="blue">' post="</font>" elif isinstance (node, docutils.nodes.reference) pre='<font color="blue">' post="</font>" uri=node.get('refuri') if uri if urlparse(uri)[0] pre+=u'<a href="%s">'%uri post='</a>'+post uri=node.get('refid') if uri pre+=u'<a href="#%s">'%uri post='</a>'+post node.pdftext=node.astext() elif isinstance (node, docutils.nodes.option_string) \ or isinstance (node, docutils.nodes.option_argument) \ node.pdftext=node.astext() elif isinstance (node, docutils.nodes.header) \ or isinstance (node, docutils.nodes.footer) \ elif isinstance (node, docutils.nodes.system_message) \

8 or isinstance (node, docutils.nodes.problematic) \ sys.stderr.write (node.astext()+"\n") sys.stderr.flush() pre='<font color="red">' post="</font>" elif isinstance (node, docutils.nodes.generated) elif isinstance (node, docutils.nodes.image) node.pdftext='<img src="%s" />'%node.get('uri') elif isinstance (node, docutils.nodes.footnote_reference) # Fixme link to the right place node.pdftext=u'<super><font color="blue">%s</font></super>'%node.astext() elif isinstance (node, docutils.nodes.citation_reference) # Fixme link to the right place node.pdftext=u'<font color="blue">[%s]</font>'%node.astext() # FIXME nodes we are ignoring for the moment elif isinstance (node, docutils.nodes.target) # FIXME make it work as a target for links print "Unkn. node (gen_pdftext) ", node. class print node #print node.transform sys.exit(1) return node.pdftext def PreformattedFit(text,style) """Preformatted section that gets horizontally compressed if needed.""" w=max(map(lambda linestringwidth(line,style.fontname,style.fontsize),text.splitlines())) mw=tw-style.leftindent-style.rightindent if w>mw style=copy(style) f=max((0.375,mw/w)) style.fontsize*=f style.leading*=f return Preformatted(text,style) def gen_elements(node, depth, in_line_block=false, style=styles['bodytext']) global decoration if isinstance (node, docutils.nodes.document) node.elements=gather_elements(node,depth,style=style) ####################### ## Tables ####################### elif isinstance (node, docutils.nodes.table) node.elements=gather_elements(node,depth) elif isinstance (node, docutils.nodes.tgroup) rows=[] hashead=false for n in node.children if isinstance (n,docutils.nodes.thead) hashead=true for row in n.children r=[] for cell in row.children r.append(cell) rows.append(r) elif isinstance (n,docutils.nodes.tbody) for row in n.children r=[] for cell in row.children

9 r.append(cell) rows.append(r) spans=filltable (rows) data=[] for row in rows r=[] for cell in row if isinstance(cell,str) r.append("") r.append(gather_elements(cell,depth)) data.append(r) st=spans+tstylenorm if hashead st+=[tstylehead] node.elements=[table(data,style=tablestyle(st))] elif isinstance (node, docutils.nodes.title) # Special cases (Not sure this is right ;-) if isinstance (node.parent, docutils.nodes.document) # FIXME maybe make it a coverpage? node.elements=[paragraph(gen_pdftext(node,depth), styles['title'])] elif isinstance (node.parent, docutils.nodes.topic) # FIXME style correctly node.elements=[paragraph(gen_pdftext(node,depth), styles['heading3'])] elif isinstance (node.parent, docutils.nodes.admonition) or \ isinstance (node.parent, docutils.nodes.sidebar) node.elements=[paragraph(gen_pdftext(node,depth), styles['heading3'])] node.elements=[paragraph(gen_pdftext(node,depth), styles['heading%d'%min(depth,3)])] elif isinstance (node, docutils.nodes.subtitle) if isinstance (node.parent,docutils.nodes.sidebar) node.elements=[paragraph(gen_pdftext(node,depth), styles['heading4'])] elif isinstance (node.parent,docutils.nodes.document) node.elements=[paragraph(gen_pdftext(node,depth), styles['subtitle'])] elif isinstance (node, docutils.nodes.paragraph) node.elements=[paragraph(gen_pdftext(node,depth), style)] elif isinstance (node, docutils.nodes.docinfo) # A docinfo usually contains several fields. # We'll render it as a series of elements, one field each. node.elements=gather_elements(node,depth,style=style) elif isinstance (node, docutils.nodes.field) # A field has two child elements, a field_name and a field_body. # We render as a two-column table, left-column is right-aligned, # bold, and much smaller fn=paragraph(gather_pdftext(node.children[0],depth)+"",style=styles['fieldname']) fb=gen_elements(node.children[1],depth) node.elements=[table([[fn,fb]],style=tstyles['field'],colwidths=[fieldlist_lwidth,none])] elif isinstance (node, docutils.nodes.decoration) # This is a tricky one. We need to switch our document's # page templates based on this. If decoration contains a # header and/or a footer, we need to use those # right now, we avoid trouble. # FIXME Implement node.elements=gather_elements(node,depth,style=style) elif isinstance (node, docutils.nodes.header) decoration['header']=paragraph(gather_pdftext(node,depth),style=styles['footer']) node.elements=[] elif isinstance (node, docutils.nodes.footer) decoration['footer']=paragraph(gather_pdftext(node,depth),style=styles['footer']) node.elements=[] elif isinstance (node, docutils.nodes.author) if isinstance (node.parent,docutils.nodes.authors) # Is only one of multiple authors. Return a paragraph node.elements=[paragraph(gather_pdftext(node,depth), style=style)] # A single author works like a field node.elements=[table([[paragraph("author",style=styles['fieldname']),

10 ] elif isinstance (node, docutils.nodes.authors) # Multiple authors. Create a two-column table. Author references on the right. td=[[paragraph("authors",style=styles['fieldname']),gather_elements(node,depth,style=style)]] node.elements=[table(td,style=tstyles['field'],colwidths=[fieldlist_lwidth,none])] elif isinstance (node, docutils.nodes.organization) t=table([[paragraph("organization",style=styles['fieldname']), elif isinstance (node, docutils.nodes.contact) t=table([[ Paragraph("Contact",style=styles['FieldName']), elif isinstance (node, docutils.nodes.address) t=table([[ Paragraph("Address",style=styles['FieldName']), elif isinstance (node, docutils.nodes.version) t=table([[ Paragraph("Version",style=styles['FieldName']), elif isinstance (node, docutils.nodes.revision) t=table([[ Paragraph("Revision",style=styles['FieldName']), elif isinstance (node, docutils.nodes.status) t=table([[ Paragraph("Version",style=styles['FieldName']), elif isinstance (node, docutils.nodes.date) t=table([[ Paragraph("Date",style=styles['FieldName']), elif isinstance (node, docutils.nodes.copyright) t=table([[paragraph("copyright",style=styles['fieldname']), elif isinstance (node, docutils.nodes.topic) \ or isinstance (node, docutils.nodes.field_body) \ node.elements=gather_elements(node,depth,style=style) elif isinstance (node, docutils.nodes.section) if depth<page_break_level node.elements=[pagebreak()]+gather_elements(node,depth+1) node.elements=gather_elements(node,depth+1) elif isinstance (node, docutils.nodes.bullet_list) \ or isinstance (node, docutils.nodes.enumerated_list) \ or isinstance (node, docutils.nodes.definition_list) \ or isinstance (node, docutils.nodes.option_list) \ or isinstance (node, docutils.nodes.field_list) \ or isinstance (node, docutils.nodes.definition) \ node.elements=gather_elements(node,depth,style=style) elif isinstance (node, docutils.nodes.option_list_item) og = gather_elements(node.children[0],depth,style) desc = gather_elements(node.children[1],depth,style) node.elements=[table([[og,desc]],style=tstyles['field'])] elif isinstance (node, docutils.nodes.definition_list_item) # I need to catch the classifiers here tt=[] dt=[] for n in node.children if isinstance(n,docutils.nodes.term) or \ isinstance(n,docutils.nodes.classifier) tt.append(gather_pdftext(n,depth,style))

11 dt=dt+gen_elements(n,depth,style) node.elements=[paragraph(''.join(tt),style),myindenter(left=10)]+dt+[myindenter(left=-10)] elif isinstance (node, docutils.nodes.list_item) # A list_item is a table of two columns. # The left one is the bullet itself, the right is the # item content. This way we can nest them. el=gather_elements(node,depth,style=style) b="" if node.parent.get('bullet') or isinstance(node.parent,docutils.nodes.bullet_list) # FIXME use correct bullet symbols, check inter-paragraph spacing b=str(node.parent.get('bullet')) if b=="none" b="" elif node.parent.get ('enumtype')=='arabic' b=str(node.parent.children.index(node)+1)+'.' elif node.parent.get ('enumtype')=='lowerroman' b=str(lowerroman[node.parent.children.index(node)])+'.' elif node.parent.get ('enumtype')=='upperroman' b=str(lowerroman[node.parent.children.index(node)].upper())+'.' elif node.parent.get ('enumtype')=='loweralpha' b=str(loweralpha[node.parent.children.index(node)])+'.' elif node.parent.get ('enumtype')=='upperalpha' b=str(loweralpha[node.parent.children.index(node)].upper())+'.' print "Unknown kind of list_item" print node.parent sys.exit(1) el[0].bullettext = b node.elements=el elif isinstance (node, docutils.nodes.transition) node.elements=[separation()] elif isinstance (node, docutils.nodes.system_message) \ or isinstance (node, docutils.nodes.problematic) \ # FIXME show the error in the document, red, whatever sys.stderr.write (node.astext()+"\n") sys.stderr.flush() node.elements=[] elif isinstance (node, docutils.nodes.block_quote) node.elements=[myindenter(left=20)]+gather_elements(node,depth,style)+[myindenter(left=-20)] elif isinstance (node, docutils.nodes.attribution) node.elements=[paragraph(gather_pdftext(node,depth),styles['attribution'])] elif isinstance (node, docutils.nodes.comment) # Class that generates no output node.elements=[] elif isinstance (node, docutils.nodes.line_block) # Obsolete? Let's do something anyway. # FIXME indent or not? qstyle=copy(style) qstyle.leftindent+=30 node.elements=gather_elements(node,depth,style=qstyle) elif isinstance (node, docutils.nodes.line) # All elements in one line node.elements=[paragraph(gather_pdftext(node,depth),style=style)] elif isinstance (node, docutils.nodes.literal_block) \ or isinstance (node, docutils.nodes.doctest_block) \ or isinstance (node, docutils.nodes.option) \ node.elements=[preformattedfit(gather_pdftext(node,depth,replaceent=false),styles['code'])] elif isinstance (node, docutils.nodes.attention) \ or isinstance (node, docutils.nodes.caution) \ or isinstance (node, docutils.nodes.danger) \ or isinstance (node, docutils.nodes.error) \ or isinstance (node, docutils.nodes.hint) \ or isinstance (node, docutils.nodes.important) \ or isinstance (node, docutils.nodes.note) \ or isinstance (node, docutils.nodes.tip) \ or isinstance (node, docutils.nodes.warning) \

12 or isinstance (node, docutils.nodes.admonition) \ node.elements=[paragraph(node.tagname.title(),style=styles['heading3'])]+gather_elements(node,depth,style=style) elif isinstance (node, docutils.nodes.image) # FIXME handle all the other attributes i=image(filename=str(node.get("uri"))) if node.get('align') i.halign=node.get('align').upper() node.elements=[i] elif isinstance (node, docutils.nodes.figure) # The sub-elements are the figure and the caption, and't ugly if # they separate node.elements=[keeptogether(gather_elements(node,depth,style=style))] elif isinstance (node, docutils.nodes.caption) node.elements=[paragraph('<i>'+gather_pdftext(node,depth)+'</i>',style=style)] elif isinstance (node, docutils.nodes.legend) node.elements=gather_elements(node,depth,style=style) elif isinstance (node, docutils.nodes.sidebar) node.elements=[table([[ gather_elements(node,depth,style=style)]],style=tstyles['sidebar'])] elif isinstance (node, docutils.nodes.rubric) node.elements=[paragraph(gather_pdftext(node,depth),styles['rubric'])] elif isinstance (node, docutils.nodes.compound) # FIXME think if this is even implementable node.elements=gather_elements(node,depth,style) elif isinstance (node, docutils.nodes.container) # FIXME think if this is even implementable node.elements=gather_elements(node,depth,style) elif isinstance (node, docutils.nodes.substitution_definition) node.elements=[] elif isinstance (node, docutils.nodes.tbody) rows=[gen_elements(n,depth) for n in node.children] t=[] for r in rows if not r continue t.append(r) node.elements=[table(t,style=tstyles['normal'])] elif isinstance (node, docutils.nodes.footnote) # It seems a footnote contains a label and a series of elements label=paragraph(gather_pdftext(node.children[0],depth),style) contents=gather_elements(node,depth,style)[1] decoration['endnotes'].append([label,contents]) node.elements=[] elif isinstance (node, docutils.nodes.label) node.elements=[paragraph(gather_pdftext(node,depth),style)] elif isinstance (node, docutils.nodes.text) node.elements=[paragraph(gather_pdftext(node,depth),style)] elif isinstance (node, docutils.nodes.entry) node.elements=gather_elements(node,depth,style) # FIXME nodes we are ignoring for the moment elif isinstance (node, docutils.nodes.target) \ or isinstance (node, docutils.nodes.footnote) \ or isinstance (node, docutils.nodes.citation) \ or isinstance (node, docutils.nodes.reference) \ or isinstance (node, docutils.nodes.raw) \ node.elements=[] print "Unkn. node (gen_elements) ", node. class print node sys.exit(1) # set anchors for internal references for id in node['ids'] node.elements.insert( node.elements and isinstance(node.elements[0], PageBreak) and 1 or 0, Paragraph('<a name="%s"/>'%id,style)) return node.elements def gather_elements (node, depth, in_line_block=false,style=styles['bodytext']) r=[]

13 for n in node.children r=r+(gen_elements(n,depth,in_line_block,style=style)) return r class Separation(Flowable) " A simple <hr>-like thingie" def wrap(self,w,h) self.w=w return (w,1*cm) def draw(self) self.canv.line(0,0.5*cm,self.w,0.5*cm) class FancyPage(PageTemplate) def init (self,_id,pw,ph,tm,bm,lm,rm,hh,fh,head,foot) tw=pw-lm-rm #textframe=frame(lm,tm+hh,tw,ph-tm-bm-hh-fh) textframe=frame(lm,tm+hh,tw,ph-tm-bm-hh-fh,toppadding=hh,bottompadding=fh) self.head=head self.hx=lm self.hy=ph-tm self.foot=foot self.fx=lm self.fy=bm PageTemplate. init (self,_id,[textframe]) def beforedrawpage(self,canv,doc) if self.head self.head.drawon(canv,self.hx,self.hy) if self.foot self.foot.drawon(canv,self.fx,self.fy) def filltable (rows) # If there is a multicol cell, we need to insert Continuation Cells # to make all rows the same length for y in range(0,len( rows)) for x in range (0,len(rows[y])) cell=rows[y][x] if isinstance (cell,str) continue if cell.get("morecols") for i in range(0,cell.get("morecols")) rows[y].insert(x+1,"") for y in range(0,len( rows)) for x in range (0,len(rows[y])) cell=rows[y][x] if isinstance (cell,str) continue if cell.get("morerows") for i in range(0,cell.get("morerows")) rows[y+i+1].insert(x,"") # Create spans list for reportlab's table style spans=[] for y in range(0,len( rows)) for x in range (0,len(rows[y])) cell=rows[y][x] if isinstance (cell,str) continue if cell.get("morecols") mc=cell.get("morecols") mc=0 if cell.get("morerows") mr=cell.get("morerows") mr=0 if mc or mr spans.append(('span',(x,y),(x+mc,y+mr))) return spans if name == " main " input=open(sys.argv[1]).read() import docutils.core doc=docutils.core.publish_doctree(input) elements=gen_elements(doc,0)

14 endnotes = decoration['endnotes'] if endnotes elements.append(spacer(1,2*cm)) elements.append(separation()) for n in decoration['endnotes'] elements.append(table([[n[0],n[1]]],style=tstyles['endnote'],colwidths=[endnote_lwidth,none])) head=decoration['header'] foot=decoration['footer'] if head hh=head.wrap(tw,ph)[1] hh=0 if foot fh=foot.wrap(tw,ph)[1] fh=0 # So, now, create the FancyPage with the right sizes and elements FP=FancyPage("fancypage",pw,ph,tm,bm,lm,rm,hh,fh,head,foot) # A basic document for us to write to 'rl_hello_platypus.pdf' pdfdoc = BaseDocTemplate(sys.argv[1]+'.pdf',pageTemplates=[FP],showBoundary=0,pagesize=ps) pdfdoc.build(elements) 唔, 您可以拿本 RST 文件来测试 rst2pdf 中文, 注意编码得是 utf8 命令./rst2pdf.py XXX.rst 其他问题 现存问题 1. 硬编码的文泉驿正黑字体路径不能灵活适应不同系统 2. 文泉驿正黑中文没有粗体和斜体 3. 由于没有选字, 而是一律文泉驿字体, 使得英文也没有了粗斜体 您说, 本文做的 pdf 没书签没目录? 唔, 这是这个 rst2pdf 程序原先的问题, 本文只解决 python-reportlab 中文显示和换行 还有其他问题么? 请洽 python-cn 的老大 Zoomq 大妈 (Zoom. Quiet <zoom.quiet@gmail.com>) ~ 奥运进行时, 请尽情搔扰 ~