initial commit

README.rst

@@ -0,0 +1,31 @@
+tldp - tools for publishing from TLDP sources
+=============================================
+
+TLDP = The Linux Documentation Project.
+These are a set of scripts that process committed documents in the
+TLDP document source repository to an output tree of choice.
+
+Installation
+------------
+
+You can install, upgrade, uninstall tldp tools with these commands::
+
+  $ pip install tldp
+  $ pip install --upgrade tldp
+  $ pip uninstall tldp
+
+There's also a package for Debian/Ubuntu, but it's not always the
+latest version.
+
+Example usages:
+---------------
+
+FIXME:  Missing examples.
+
+Links
+-----
+
+* `Output documentation tree (sample) <http://www.tldp.org/>`_
+
+* `Source tree on GitHub <https://github.com/tLDP/LDP>`_
+

setup.py

@@ -0,0 +1,33 @@
+# -*- coding: utf-8 -*-
+import os
+
+from setuptools import setup, find_packages
+
+
+with open(os.path.join(os.path.dirname(__file__), 'README.rst')) as r_file:
+    readme = r_file.read()
+
+
+setup(
+    name='tldp',
+    version='0.1',
+    license='GNU',
+    author='Martin A. Brown',
+    author_email='martin@linux-ip.net',
+    description='tools for processing all TLDP source documents',
+    long_description=readme,
+    packages=find_packages(),
+    test_suite='tests',
+    install_requires=[
+        'six>=1.7',
+    ],
+    classifiers=[
+        'Development Status :: 3 - Alpha',
+        'Intended Audience :: Developers',
+        'License :: OSI Approved :: MIT License',
+        'Operating System :: OS Independent',
+        'Programming Language :: Python',
+        'Programming Language :: Python :: 3',
+        'Topic :: Software Development :: Libraries :: Python Modules'
+    ],
+)

tldp/doctypes/__init__.py

@@ -0,0 +1,7 @@
+from .text import Text
+from .rst import RestructuredText
+from .markdown import Markdown
+from .linuxdoc import Linuxdoc
+from .docbooksgml import DocbookSGML
+from .docbook4xml import Docbook4XML
+from .docbook5xml import Docbook5XML

tldp/doctypes/common.py

@@ -0,0 +1,27 @@
+#! /usr/bin/python
+
+from __future__ import print_function
+
+import string
+from ..utils import logger
+
+
+class SignatureChecker(object):
+
+    @classmethod
+    def signatureLocation(cls, name, fin):
+        fin.seek(0)
+        buf = fin.read(1024)
+        for sig in cls.signatures:
+            try:
+                sindex = string.index(buf.lower(), sig.lower())
+                logger.debug("In file %s, signature %s found at %s; doctype %s found",
+                             name, sig, sindex, cls)
+                return sindex
+            except ValueError:
+                logger.debug("In file %s, signature %s not found for document type %s",
+                             name, sig, cls.__name__)
+        return None
+
+#
+# -- end of file

tldp/doctypes/docbook4xml.py

@@ -0,0 +1,17 @@
+#! /usr/bin/python
+
+from ..utils import logger
+
+from .common import SignatureChecker
+
+
+class Docbook4XML(SignatureChecker):
+    extensions = ['.xml']
+    signatures = ['-//OASIS//DTD DocBook XML V4.1.2//EN',
+                  '-//OASIS//DTD DocBook XML V4.2//EN',
+                  '-//OASIS//DTD DocBook XML V4.2//EN',
+                  '-//OASIS//DTD DocBook XML V4.4//EN',
+                  '-//OASIS//DTD DocBook XML V4.5//EN', ]
+
+#
+# -- end of file

tldp/doctypes/docbook5xml.py

@@ -0,0 +1,13 @@
+#! /usr/bin/python
+
+from ..utils import logger
+from .common import SignatureChecker
+
+
+class Docbook5XML(SignatureChecker):
+    extensions = ['.xml']
+    signatures = ['-//OASIS//DTD DocBook V5.0/EN',
+                  'http://docbook.org/ns/docbook', ]
+
+#
+# -- end of file

tldp/doctypes/docbooksgml.py

@@ -0,0 +1,15 @@
+#! /usr/bin/python
+
+from ..utils import logger
+from .common import SignatureChecker
+
+
+class DocbookSGML(SignatureChecker):
+    extensions = ['.sgml']
+    signatures = ['-//Davenport//DTD DocBook V3.0//EN',
+                  '-//OASIS//DTD DocBook V3.1//EN',
+                  '-//OASIS//DTD DocBook V4.1//EN',
+                  '-//OASIS//DTD DocBook V4.2//EN', ]
+
+#
+# -- end of file

tldp/doctypes/linuxdoc.py

@@ -0,0 +1,24 @@
+#! /usr/bin/python
+
+from ..utils import logger
+from .common import SignatureChecker
+
+
+class Linuxdoc(SignatureChecker):
+    extensions = ['.sgml']
+    signatures = ['<!doctype linuxdoc system', ]
+
+    def create_txt(self):
+        pass
+
+    def create_pdf(self):
+        pass
+
+    def create_html(self):
+        pass
+
+    def create_htmls(self):
+        pass
+
+#
+# -- end of file

tldp/doctypes/markdown.py

@@ -0,0 +1,11 @@
+#! /usr/bin/python
+
+from ..utils import logger
+
+
+class Markdown(object):
+    extensions = ['.md']
+    signatures = []
+
+#
+# -- end of file

tldp/doctypes/rst.py

@@ -0,0 +1,11 @@
+#! /usr/bin/python
+
+from ..utils import logger
+
+
+class RestructuredText(object):
+    extensions = ['.rst']
+    signatures = []
+
+#
+# -- end of file

tldp/doctypes/text.py

@@ -0,0 +1,11 @@
+#! /usr/bin/python
+
+from ..utils import logger
+
+
+class Text(object):
+    extensions = ['.txt']
+    signatures = []
+
+#
+# -- end of file

tldp/guess.py

@@ -0,0 +1,86 @@
+#! /usr/bin/python
+
+from __future__ import print_function
+
+import os
+import inspect
+
+from . import doctypes
+from .utils import logger
+
+
+def makefh(thing):
+    if isinstance(thing, file):
+        f = thing
+    elif isinstance(thing, str) and os.path.isfile(thing):
+        f = open(thing)
+    else:
+        raise TypeError("Cannot make file from type %s of %r" %
+                        (type(thing), thing,))
+    return f
+
+
+def listDoctypes():
+    knowndoctypes = list()
+    for name, member in inspect.getmembers(doctypes, inspect.isclass):
+        logger.info("Located class %s (%r).", name, member)
+        knowndoctypes.append(member)
+    logger.info("Capable of handling %s document classes.", len(knowndoctypes))
+    return knowndoctypes
+
+
+def guess(thing):
+    try:
+        fin = makefh(thing)
+    except TypeError:
+        return None
+
+    _, ext = os.path.splitext(fin.name)
+    possible = [t for t in knowndoctypes if ext in t.extensions]
+    if not possible:
+        return None
+    if len(possible) == 1:
+        doctype = possible.pop()
+        return doctype
+
+    # -- for this extension, multiple document types, probably SGML, XML
+    #
+    logger.info("Extension is %s for %s; multiple possible document types.",
+                ext, fin.name)
+    for doctype in possible:
+        logger.debug("Extension is %s for %s; %s.", ext, fin.name, doctype)
+
+    guesses = list()
+    for doctype in possible:
+        sindex = doctype.signatureLocation(fin.name, fin)
+        if sindex is not None:
+            guesses.append((sindex, doctype))
+
+    if not guesses:
+        logger.info("Extension is %s for %s; but no matching signature found.",
+                    ext, fin.name)
+        return None
+    if len(guesses) == 1:
+        _, doctype = guesses.pop()
+        return doctype
+
+    # -- OK, this is unusual; we still found multiple document type
+    #    signatures.  Seems rare but unlikely, so we should choose the
+    #    first signature in the file as the more likely document type.
+    #
+    guesses.sort()
+    logger.info("Multiple guesses for file %s", fin.name)
+    for sindex, doctype in guesses:
+        logger.info("Could be %s (file position %s)", doctype, sindex)
+    logger.info("Going to guess that it is %s", doctype)
+    _, doctype = guesses.pop()
+    return doctype
+
+
+knowndoctypes = listDoctypes()
+knownextensions = set()
+for x in knowndoctypes:
+    knownextensions.update(x.extensions)
+
+#
+# -- end of file

tldp/outputs.py

@@ -0,0 +1,20 @@
+#! /usr/bin/python
+
+import os
+import sys
+import logging
+
+def getLogger(opts):
+    logging.basicConfig(stream=sys.stderr, level=logging.INFO)
+    logger = logging.getLogger()
+    return logger
+
+class SourceDocument(object):
+
+    def __init__(self, filename):
+        self.filename = fname
+        self.stem = 
+
+    def 
+
+# -- end of file

tldp/processors.py

@@ -0,0 +1,20 @@
+#! /usr/bin/python
+
+import os
+import sys
+import logging
+
+def getLogger(opts):
+    logging.basicConfig(stream=sys.stderr, level=logging.INFO)
+    logger = logging.getLogger()
+    return logger
+
+class SourceDocument(object):
+
+    def __init__(self, filename):
+        self.filename = fname
+        self.stem = 
+
+    def 
+
+# -- end of file

tldp/sources.py

@@ -0,0 +1,118 @@
+#! /usr/bin/python
+
+from __future__ import print_function
+
+import os
+import sys
+import logging
+
+from .utils import logger
+from .guess import knownextensions
+
+# class SourceTree(object):
+#
+#     def __init__(self, root, config):
+#         pass
+
+
+class SourceDir(object):
+
+    def __repr__(self):
+        return self.dirname
+
+    def __init__(self, dirname):
+        self.dirname = os.path.abspath(dirname)
+        self.docs = list()
+        if not os.path.exists(dirname):
+            raise OSError("[Errno 2] No such file or directory: " + dirname)
+        logger.info("Time to go for an enumeration stroll in %s", dirname)
+        self.enumerateDocuments()
+
+    def enumerateDocuments(self):
+        for fname in os.listdir(self.dirname):
+            possible = os.path.join(self.dirname, fname)
+            if os.path.isfile(possible):
+                self.docs.append(SourceDocument(possible))
+            elif os.path.isdir(fname):
+                stem = os.path.basename(fname)
+                for ext in knownextensions:
+                    possible = os.path.join(self.dirname, fname, stem + ext)
+                    if os.path.isfile(possible):
+                        self.docs.append(SourceDocument(possible))
+        logger.info("Discovered %s documents in %s",
+                    len(self.docs), self.dirname)
+
+
+class SourceDocument(object):
+
+    def __repr__(self):
+        return self.filename
+
+    def __init__(self, filename):
+        # -- canonicalize the pathname we are given.
+        self.filename = os.path.abspath(filename)
+        if not os.path.exists(self.filename):
+            raise OSError("Missing source document: " + self.filename)
+
+        self.dirname, self.basename = os.path.split(self.filename)
+        self.stem, self.ext = os.path.splitext(self.basename)
+        self.stat = os.stat(self.filename)
+
+        self.doctype = None
+        self.resources = False  # -- assume no ./images/, ./resources/
+        self.singlefile = True  # -- assume only one file
+        parentdir = os.path.basename(self.dirname)
+        if parentdir == self.stem:
+            self.singlefile = False
+            for rdir in ('resources', 'images'):
+                if os.path.exists(os.path.join(self.dirname, rdir)):
+                    self.resources = True
+
+    def doctype(self):
+        if self.doctype is None:
+            self.doctype = guess(self.filename)
+        return self.doctype
+
+
+class OutputDocument(object):
+
+    formats = {'pdf':  '.pdf',
+               'txt':  '.txt',
+               'html':  '.html',
+               'htmls':  '-single.html', }
+
+    def __init__(self, filename):
+        pass
+
+    @property
+    def txt(self):
+        return os.path.join(self.dirname, self.stem, '.txt')
+
+
+class OutputDir(object):
+
+    def __init__(self, dirname):
+        self.dirname = os.path.abspath(dirname)
+        self.parent = os.path.dirname(dirname)
+        self.stem = os.path.basename(dirname)
+        self.members = list()
+
+    def mkdir(self):
+        if not os.path.exists(self.parent):
+            raise OSError("Missing parent directory: " + self.parent)
+        os.mkdir(self.dirname)
+        return os.path.exists(self.dirname)
+
+    @property
+    def members(self):
+        return os.path.exists(self.dirname)
+
+    @property
+    def exists(self):
+        return os.path.exists(self.dirname)
+
+    @property
+    def isComplete(self):
+        return all(self.pdf, self.html, self.htmls, self.txt)
+
+# -- end of file

tldp/utils.py

@@ -0,0 +1,16 @@
+#! /usr/bin/python
+
+import sys
+import logging
+
+
+def getLogger(**opts):
+    level = opts.get('level', logging.INFO)
+    logging.basicConfig(stream=sys.stderr, level=level)
+    logger = logging.getLogger()
+    return logger
+
+logger = getLogger()
+
+#
+# -- end of file