[views,pdf] minimal support for xhtml->pdf conversion 3.5
authorAurelien Campeas <aurelien.campeas@logilab.fr>
Thu, 20 Aug 2009 14:48:35 +0200
branch3.5
changeset 2935 e06b3eadef31
parent 2934 cf4d39416fca
child 2936 f35b64718d02
[views,pdf] minimal support for xhtml->pdf conversion
utils.py
web/views/basetemplates.py
web/xhtml2fo.py
--- a/utils.py	Thu Aug 20 14:30:45 2009 +0200
+++ b/utils.py	Thu Aug 20 14:48:35 2009 +0200
@@ -320,3 +320,27 @@
                                                  self.body.getvalue())
 
 
+def can_do_pdf_conversion(__answer=[None]):
+    """pdf conversion depends on
+    * pyxmltrf (python package)
+    * fop 0.9x
+    """
+    if __answer[0] is not None:
+        return __answer[0]
+    try:
+        import pyxmltrf
+    except ImportError:
+        __answer[0] = False
+        return False
+    from subprocess import Popen, STDOUT
+    import os
+    try:
+        Popen(['/usr/bin/fop', '-q'],
+              stdout=open(os.devnull, 'w'),
+              stderr=STDOUT)
+    except OSError, e:
+        print e
+        __answer[0] = False
+        return False
+    __answer[0] = True
+    return True
--- a/web/views/basetemplates.py	Thu Aug 20 14:30:45 2009 +0200
+++ b/web/views/basetemplates.py	Thu Aug 20 14:48:35 2009 +0200
@@ -13,7 +13,7 @@
 from cubicweb.appobject import objectify_selector
 from cubicweb.selectors import match_kwargs
 from cubicweb.view import View, MainTemplate, NOINDEX, NOFOLLOW
-from cubicweb.utils import make_uid, UStringIO
+from cubicweb.utils import make_uid, UStringIO, can_do_pdf_conversion
 
 
 # main templates ##############################################################
@@ -266,6 +266,44 @@
             self.w(u'</td>\n')
             self.w(u'</tr></table>\n')
 
+if can_do_pdf_conversion():
+    from xml.etree.cElementTree import ElementTree
+    from subprocess import Popen as sub
+    from StringIO import StringIO
+    from tempfile import NamedTemporaryFile
+    from cubicweb.web.xhtml2fo import ReportTransformer
+
+    class PdfMainTemplate(TheMainTemplate):
+        id = 'pdf-main-template'
+
+        def call(self, view):
+            """build the standard view, then when it's all done, convert xhtml to pdf
+            """
+            super(PdfMainTemplate, self).call(view)
+            pdf = self.to_pdf(self._stream)
+            self.req.set_content_type('application/pdf', filename='report.pdf')
+            self.binary = True
+            self.w = None
+            self.set_stream()
+            # pylint needs help
+            self.w(pdf)
+
+        def to_pdf(self, stream, section='contentmain'):
+            # XXX see ticket/345282
+            stream = stream.getvalue().replace('&nbsp;', '&#160;').encode('utf-8')
+            xmltree = ElementTree()
+            xmltree.parse(StringIO(stream))
+            foptree = ReportTransformer(section).transform(xmltree)
+            foptmp = NamedTemporaryFile()
+            pdftmp = NamedTemporaryFile()
+            foptree.write(foptmp)
+            foptmp.flush()
+            fopproc = sub(['/usr/bin/fop', foptmp.name, pdftmp.name])
+            fopproc.wait()
+            pdftmp.seek(0)
+            pdf = pdftmp.read()
+            return pdf
+
 # page parts templates ########################################################
 
 class HTMLHeader(View):
@@ -489,4 +527,6 @@
 
 ## vregistry registration callback ############################################
 def registration_callback(vreg):
-    vreg.register_all(globals().values(), modname=__name__)
+    vreg.register_all(globals().values(), __name__, (PdfMainTemplate,) )
+    if can_do_pdf_conversion():
+        vreg.register(PdfMainTemplate)
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/web/xhtml2fo.py	Thu Aug 20 14:48:35 2009 +0200
@@ -0,0 +1,142 @@
+from cubicweb.utils import can_do_pdf_conversion
+assert can_do_pdf_conversion()
+
+from xml.etree.ElementTree import QName, fromstring
+from pyxmltrf.standard.xhtml_xslfo.transformer import XHTML2FOTransformer
+from pyxmltrf.utils.xslfo.standard import cm
+from pyxmltrf.utils.xslfo import SimplePageMaster
+from pyxmltrf.standard.xhtml_xslfo.default_styling import default_styles
+from pyxmltrf.standard.xhtml_xslfo import XHTML_NS
+
+
+class ReportTransformer(XHTML2FOTransformer):
+    """
+    Class transforming an XHTML input tree into a FO document
+    displaying reports (one report for each <div class="contentmain">
+    element in the input tree.
+    """
+
+    def __init__(self, section,
+                 page_width=21.0, page_height=29.7,
+                 margin_top=1.0, margin_bottom=1.0,
+                 margin_left=1.0, margin_right=1.0,
+                 header_footer_height=0.75,
+                 standard_font_size=11.0, default_lang=u"fr" ):
+        """
+        Initializes a transformer turning an XHTML input tree
+        containing <div class="contentmain"> elements representing
+        main content sections into a FO output tree displaying the
+        reports.
+
+        page_width: float - width of the page (in cm)
+        page_height: float - height of the page (in cm)
+        margin_top: float - top margin of the page (in cm)
+        margin_bottom: float - bottom margin of the page (in cm)
+        margin_left: float - left margin of the page (in cm)
+        margin_right: float - right margin of the page (in cm)
+        header_footer_height: float - height of the header or the footer of the
+                              page that the page number (if any) will be
+                              inserted in.
+        standard_font_size: float - standard size of the font (in pt)
+        default_lang: u"" - default language (used for hyphenation)
+        """
+        self.section = section
+        self.page_width = page_width
+        self.page_height = page_height
+
+        self.page_tmargin = margin_top
+        self.page_bmargin = margin_bottom
+        self.page_lmargin = margin_left
+        self.page_rmargin = margin_right
+
+        self.hf_height = header_footer_height
+
+        self.font_size = standard_font_size
+        self.lang = default_lang
+
+        XHTML2FOTransformer.__init__(self)
+
+
+    def define_pagemasters(self):
+        """
+        Defines the page masters for the FO output document.
+        """
+        pm = SimplePageMaster(u"page-report")
+        pm.set_page_dims( self.page_width*cm, self.page_height*cm )
+        pm.set_page_margins({u'top'   : self.page_tmargin*cm,
+                             u'bottom': self.page_bmargin*cm,
+                             u'left'  : self.page_lmargin*cm,
+                             u'right' : self.page_rmargin*cm })
+        pm.add_peripheral_region(u"end",self.hf_height)
+        dims = {}
+        dims[u"bottom"] = self.hf_height + 0.25
+        pm.set_main_region_margins(dims)
+        return [pm]
+
+    def _visit_report(self, in_elt, _out_elt, params):
+        """
+        Specific visit function for the input <div> elements whose class is
+        "report". The _root_visit method of this class selects these input
+        elements and asks the process of these elements with this specific
+        visit function.
+        """
+
+        ps = self.create_pagesequence(u"page-report")
+        props = { u"force-page-count": u"no-force",
+                  u"initial-page-number": u"1",
+                  u"format": u"1", }
+        self._output_properties(ps,props)
+
+        sc = self.create_staticcontent(ps, u"end")
+        sc_bl = self.create_block(sc)
+        attrs = { u"hyphenate": u"false", }
+        attrs[u"font-size"] = u"%.1fpt" %(self.font_size*0.7)
+        attrs[u"language"] = self.lang
+        attrs[u"text-align"] = u"center"
+        self._output_properties(sc_bl,attrs)
+        sc_bl.text = u"Page" + u" " # ### Should be localised!
+        pn = self.create_pagenumber(sc_bl)
+        pn.tail = u"/"
+        lpn = self.create_pagenumbercitation( sc_bl,
+                                              u"last-block-of-report-%d" % params[u"context_pos"]
+                                              )
+
+
+        fl = self.create_flow(ps,u"body")
+        bl = self.create_block(fl)
+
+        # Sets on the highest block element the properties of the XHTML body
+        # element. These properties (at the least the inheritable ones) will
+        # be inherited by all the future FO elements.
+        bodies = list(self.in_tree.getiterator(QName(XHTML_NS,u"body")))
+        if len(bodies) > 0:
+            attrs = self._extract_properties([bodies[0]])
+        else:
+            attrs = default_styles[u"body"].copy()
+        attrs[u"font-size"] = u"%.1fpt" %self.font_size
+        attrs[u"language"] = self.lang
+        self._output_properties(bl,attrs)
+
+        # Processes the report content
+        self._copy_text(in_elt,bl)
+        self._process_nodes(in_elt.getchildren(),bl)
+
+        # Inserts an empty block at the end of the report in order to be able
+        # to compute the last page number of this report.
+        last_bl = self.create_block(bl)
+        props = { u"keep-with-previous": u"always", }
+        props[u"id"] = u"last-block-of-report-%d" % params[u"context_pos"]
+        self._output_properties(last_bl,props)
+
+
+    def _root_visit(self):
+        """
+        Visit function called when starting the process of the input tree.
+        """
+        content = [ d for d in self.in_tree.getiterator(QName(XHTML_NS,u"div"))
+                    if d.get(u"id") == self.section ]
+        # Asks the process of the report elements with a specific visit
+        # function
+        self._process_nodes(content, self.fo_root,
+                            with_function=self._visit_report)
+