Peter's pure Python PicoPDF: Create a PDF in 100 lines of Python

peterspurepythonpicopdf.py

from datetime import datetime
from zlib import compress

def serialize(x):
    if isinstance(x, dict):
        return '<<' + '\n'.join(serialize(k) + ' ' + serialize(v) for k, v in x.items()) + '>>'
    if isinstance(x, list):
        return '[' + ' '.join(serialize(it) for it in x) + ']'
    return str(x)

def reference(objectNumber, generation=0):
    return '{0} {1} R'.format(objectNumber, generation)

def stream(streamDict, streamBytes):
    return '\n'.join([serialize(streamDict), 'stream', streamBytes, 'endstream'])

class PetersPurePythonPicoPDF(object):
    def __init__(self):
        self.compress = True
        self.pageWidth = 595.28     # A4
        self.pageHeight = 841.89
        self.pages = [[]]
        self.fonts = {}
        for fontName in [ "Times-Roman", "Helvetica", "Courier", "Symbol", "Times-Bold", "Helvetica-Bold", "Courier-Bold", "ZapfDingbats", "Times-Italic", "Helvetica-Oblique", "Courier-Oblique", "Times-BoldItalic", "Helvetica-BoldOblique", "Courier-BoldOblique" ]:
            self.addFont(fontName, "StandardEncoding")

    def addFont(self, baseFont, encoding):
        id = 'F{0}'.format(len(self.fonts) + 1)
        self.fonts[id] = { 'id': id, 'baseFont' : baseFont, 'encoding': encoding }
        return self.fonts[id]

    def addText(self, lines, x, y, font='F1', size=16, leading=1, color='0 g'):
        lines = 'T* '.join('({line}) Tj\n'.format(line=line) for line in lines)
        self.pages[-1].append('BT\n/{font} {size} Tf\n{leading} TL\n{color}\n'
                 '{x} {y} Td\n{lines}'
                 'ET'.format(font=font, size=size, leading=size*leading, color=color,
                             x=x, y=self.pageHeight - y, lines=lines))

    def save(self, filename):
        indirectObjectOffsets = [0] # Current offset in [-1]
        output = []

        def out(*args):
            v = '\n'.join(serialize(x) for x in args)
            indirectObjectOffsets[-1] += len(v) + 1
            output.append(v)

        def lastObjectNumber():
            return len(indirectObjectOffsets) - 1

        def indirectObject(v):
            indirectObjectOffsets.append(indirectObjectOffsets[-1])
            return '{0} 0 obj\n{1}\nendobj'.format(lastObjectNumber(), serialize(v))

        out('%PDF-1.3') # Header
        pagesRootNumber = lastObjectNumber() + 2 * len(self.pages) + 1
        resourcesNumber = pagesRootNumber + len(self.fonts) + 1
        pageNumbers = []
        for pageContent in self.pages:
            contentNumber = lastObjectNumber() + 2
            out(indirectObject({ '/Type': '/Page',
                                 '/Parent': reference(pagesRootNumber),
                                 '/Resources': reference(resourcesNumber),
                                 '/Contents': reference(contentNumber) }))
            pageNumbers.append(lastObjectNumber())
            pageBytes = '\n'.join(pageContent)
            pageStreamDict = {}
            if self.compress:
                pageStreamDict['/Filter'] = ['/FlateDecode']
                pageBytes = compress(pageBytes)
            pageStreamDict['/Length'] = len(pageBytes)
            out(indirectObject(stream(pageStreamDict, pageBytes)))
        out(indirectObject({ '/Type': '/Pages',
                             '/Kids': [reference(id) for id in pageNumbers],
                             '/Count': len(self.pages),
                             '/MediaBox': [0, 0, self.pageWidth, self.pageHeight] }))
        for font in self.fonts.values():
            out(indirectObject({ '/Type': '/Font',
                                 '/Subtype': '/Type1',
                                 '/BaseFont': '/' + font['baseFont'],
                                 '/Encoding': '/' + font['encoding'] }))
            font['objectNumber'] = lastObjectNumber()
        out(indirectObject({ '/ProcSet':  ['/PDF', '/Text' '/ImageB', '/ImageC', '/ImageI'],
                             '/Font': { '/' + key: reference(font['objectNumber']) for key, font in self.fonts.items() }}))
        out(indirectObject({ '/Type': '/Catalog',
                             '/Pages': reference(pagesRootNumber),
                             '/PageLayout': '/OneColumn' }))
        catalogNumber = lastObjectNumber()
        totalObjects = catalogNumber + 1
        cross_ref_offset = indirectObjectOffsets[-1]
        out('xref')
        out('0 ' + str(totalObjects))
        out('0000000000 65535 f ') # First object 0 shall be free
        for offset in indirectObjectOffsets[:-1]:
            out('{0:010} 00000 n '.format(offset))
        out('trailer', { '/Size': totalObjects,
                         '/Root': reference(catalogNumber) },
            'startxref', cross_ref_offset, '%%EOF')
        with open(filename, 'wb') as f:
            f.write('\n'.join(output))

p = PetersPurePythonPicoPDF()
p.addText(["Please Peter Parker", "Princess Python possesses proding powers"], 123, 456)
p.save("pppp.pdf")