Réorganisation en prévision du support des fichiers chordpro

Ref: #64

patacrep/content/__init__.py

@@ -173,8 +173,7 @@ def process_content(content, config=None):
     included in the .tex file.
     """
     contentlist = []
-    plugins = config.get('_content_plugins', {})
-
+    plugins = files.load_plugins(config, ["content"], "CONTENT_PLUGINS")
     keyword_re = re.compile(r'^ *(?P<keyword>\w*) *(\((?P<argument>.*)\))? *$')
     if not content:
         content = [["song"]]

patacrep/content/song.py

@@ -6,10 +6,35 @@ import os
 
 from patacrep.content import process_content, ContentError
 from patacrep import files, errors
-from patacrep.songs import Song
 
 LOGGER = logging.getLogger(__name__)
 
+class SongRenderer(Content):
+    """Render a song in as a tex code."""
+
+    def __init__(self, song):
+        super().__init__()
+        self.song = song
+
+    def begin_new_block(self, previous, __context):
+        """Return a boolean stating if a new block is to be created."""
+        return not isinstance(previous, SongRenderer)
+
+    def begin_block(self, context):
+        """Return the string to begin a block."""
+        indexes = context.resolve("indexes")
+        if isinstance(indexes, jinja2.runtime.Undefined):
+            indexes = ""
+        return r'\begin{songs}{%s}' % indexes
+
+    def end_block(self, __context):
+        """Return the string to end a block."""
+        return r'\end{songs}'
+
+    def render(self, context):
+        """Return the string that will render the song."""
+        return self.song.tex(output=context['filename'])
+
 #pylint: disable=unused-argument
 def parse(keyword, argument, contentlist, config):
     """Parse data associated with keyword 'song'.
@@ -23,6 +48,7 @@ def parse(keyword, argument, contentlist, config):
 
     Return a list of Song() instances.
     """
+    plugins = files.load_plugins(config, ["songs"], "SONG_PARSERS")
     if '_languages' not in config:
         config['_languages'] = set()
     songlist = []
@@ -30,8 +56,11 @@ def parse(keyword, argument, contentlist, config):
     for songdir in config['_songdir']:
         if contentlist:
             break
-        contentlist = files.recursive_find(songdir.fullpath, plugins.keys())
-
+        contentlist = [
+            filename
+            for filename
+            in files.recursive_find(songdir.fullpath, plugins.keys())
+            ]
     for elem in contentlist:
         before = len(songlist)
         for songdir in config['_songdir']:
@@ -39,18 +68,24 @@ def parse(keyword, argument, contentlist, config):
                 continue
             with files.chdir(songdir.datadir):
                 for filename in glob.iglob(os.path.join(songdir.subpath, elem)):
-                    LOGGER.debug('Parsing file "{}"…'.format(filename))
-                    try:
-                        renderer = plugins[filename.split('.')[-1]]
-                    except KeyError:
+                    extension = filename.split(".")[-1]
+                    if extension not in plugins:
                         LOGGER.warning((
-                            'I do not know how to parse file "{}". Ignored.'
-                            ).format(os.path.join(songdir.datadir, filename))
+                            'File "{}" does not end with one of {}. Ignored.'
+                            ).format(
+                                os.path.join(songdir.datadir, filename),
+                                ", ".join(["'.{}'".format(key) for key in plugins.keys()]),
+                                )
                             )
                         continue
-                    song = renderer(songdir.datadir, filename, config)
-                    songlist.append(song)
-                    config["_languages"].update(song.languages)
+                    LOGGER.debug('Parsing file "{}"…'.format(filename))
+                    renderer = SongRenderer(plugins[extension](
+                            songdir.datadir,
+                            filename,
+                            config,
+                            ))
+                    songlist.append(renderer)
+                    config["_languages"].update(renderer.song.languages)
             if len(songlist) > before:
                 break
         if len(songlist) == before:
@@ -69,8 +104,8 @@ CONTENT_PLUGINS = {'song': parse}
 class OnlySongsError(ContentError):
     "A list that should contain only songs also contain other type of content."
     def __init__(self, not_songs):
-        super(OnlySongsError, self).__init__()
         self.not_songs = not_songs
+        super().__init__('song', str(self))
 
     def __str__(self):
         return (

patacrep/content/sorted.py

@@ -40,8 +40,9 @@ def key_generator(sort):
         - sort: the list of keys used to sort.
     """
 
-    def ordered_song_keys(song):
+    def ordered_song_keys(songrenderer):
         """Return the list of values used to sort the song."""
+        song = songrenderer.song
         songkey = []
         for key in sort:
             if key == "@title":

patacrep/files.py

@@ -1,7 +1,7 @@
 """File system utilities."""
 
 from contextlib import contextmanager
-import glob
+import fnmatch
 import importlib
 import logging
 import os
@@ -11,14 +11,10 @@ import sys
 
 LOGGER = logging.getLogger(__name__)
 
-def recursive_find(root_directory, extensions):
-    """Recursively find files with some extension, from a root_directory.
+def recursive_find(root_directory, patterns):
+    """Recursively find files matching one of the patterns, from a root_directory.
 
-    Return a list of files matching those conditions.
-
-    Arguments:
-    - `extensions`: list of accepted extensions.
-    - `root_directory`: root directory of the search.
+    Return a list of files matching one of the patterns.
     """
     if not os.path.isdir(root_directory):
         return []
@@ -27,8 +23,11 @@ def recursive_find(root_directory, extensions):
     pattern = re.compile(r'.*\.({})$'.format('|'.join(extensions)))
     with chdir(root_directory):
         for root, __ignored, filenames in os.walk(os.curdir):
-            for filename in filenames:
-                if pattern.match(filename):
+            for pattern in patterns:
+                for filename in fnmatch.filter(
+                        filenames,
+                        "*.{}".format(pattern),
+                    ):
                     matches.append(os.path.join(root, filename))
     return matches
 
@@ -71,29 +70,31 @@ def chdir(path):
     else:
         yield
 
-def load_plugins(datadirs, subdir, variable, error):
-    """Load all content plugins, and return a dictionary of those plugins.
-
-    A plugin is a .py file, submodule of `subdir`, located in one of the
-    directories of `datadirs`. It contains a dictionary `variable`. The return
-    value is the union of the dictionaries of the loaded plugins.
+def load_plugins(config, root_modules, keyword):
+    """Load all plugins, and return a dictionary of those plugins.
 
     Arguments:
-    - datadirs: list of directories (as strings) in which files has to be
-      searched.
-    - subdir: modules (as a list of strings) files has to be submodules of
-      (e.g. if `subdir` is `['first', 'second']`, search files are of the form
-      `first/second/*.py`.
-    - variable: Name of the variable holding the dictionary.
-    - error: Error message raised if a key appears several times.
+    - config: the configuration dictionary of the songbook
+    - root_modules: the submodule in which plugins are to be searched, as a
+      list of modules (e.g. ["some", "deep", "module"] for
+      "some.deep.module").
+    - keyword: attribute containing plugin information.
+
+    Return value: a dictionary where:
+    - keys are the keywords ;
+    - values are functions triggered when this keyword is met.
     """
+    # pylint: disable=star-args
     plugins = {}
     directory_list = (
             [
-                os.path.join(datadir, "python", *subdir) #pylint: disable=star-args
-                for datadir in datadirs
+                os.path.join(datadir, "python", *root_modules)
+                for datadir in config.get('datadir', [])
             ]
-            + [os.path.dirname(__file__)]
+            + [os.path.join(
+                os.path.dirname(__file__),
+                *root_modules
+                )]
             )
     for directory in directory_list:
         if not os.path.exists(directory):
@@ -102,31 +103,26 @@ def load_plugins(datadirs, subdir, variable, error):
                     directory
                     )
             continue
-        sys.path.append(directory)
-        for name in glob.glob(os.path.join(directory, *(subdir + ['*.py']))):
-            if name.endswith(".py") and os.path.basename(name) != "__init__.py":
-                if directory == os.path.dirname(__file__):
-                    plugin = importlib.import_module(
-                            'patacrep.{}.{}'.format(
-                                ".".join(subdir),
-                                os.path.basename(name[:-len('.py')])
-                                )
-                            )
+        for (dirpath, __ignored, filenames) in os.walk(directory):
+            modules = ["patacrep"] + root_modules
+            if os.path.relpath(dirpath, directory) != ".":
+                modules.extend(os.path.relpath(dirpath, directory).split("/"))
+            for name in filenames:
+                if name == "__init__.py":
+                    modulename = []
+                elif name.endswith(".py"):
+                    modulename = [name[:-len('.py')]]
                 else:
-                    plugin = importlib.import_module(
-                                os.path.basename(name[:-len('.py')])
+                    continue
+                plugin = importlib.import_module(".".join(modules + modulename))
+                if hasattr(plugin, keyword):
+                    for (key, value) in getattr(plugin, keyword).items():
+                        if key in plugins:
+                            LOGGER.warning(
+                                "File %s: Keyword '%s' is already used. Ignored.",
+                                relpath(os.path.join(dirpath, name)),
+                                key,
                                 )
-                for (key, value) in getattr(plugin, variable, {}).items():
-                    if key in plugins:
-                        LOGGER.warning(
-                                error.format(
-                                filename=relpath(name),
-                                key=key,
-                                )
-                                )
-                        continue
-                    plugins[key] = value
-        del sys.path[-1]
+                            continue
+                        plugins[key] = value
     return plugins
-
-

patacrep/latex/__init__.py

@@ -1,19 +1,3 @@
-"""Very simple LaTeX parser
+"""Dumb and very very incomplete LaTeX parser."""
 
-This module uses an LALR parser to try to parse LaTeX code. LaTeX language
-*cannot* be parsed by an LALR parser, so this is a very simple attemps, which
-will work on simple cases, but not on complex ones.
-"""
-
-from patacrep.latex.syntax import tex2plain
-from patacrep.latex.syntax import parsesong as syntax_parsesong
-from patacrep import encoding
-
-def parsesong(path, fileencoding=None):
-    """Return a dictonary of data read from the latex file `path`.
-
-    """
-    with encoding.open_read(path, encoding=fileencoding) as songfile:
-        data = syntax_parsesong(songfile.read(), path)
-    data['@path'] = path
-    return data
+from patacrep.latex.syntax import tex2plain, parse_song

patacrep/latex/syntax.py

@@ -245,12 +245,21 @@ def tex2plain(string):
                 )
         )
 
-def parsesong(string, filename=None):
-    """Parse song and return its metadata."""
-    return detex(
-        silent_yacc(module=Parser(filename)).parse(
-            string,
-            lexer=SongLexer().lexer,
-            ).metadata
-        )
+def parse_song(content, filename=None):
+    """Parse some LaTeX code, expected to be a song.
 
+    Arguments:
+    - content: the code to parse.
+    - filename: the name of file where content was read from. Used only to
+      display error messages.
+    """
+    return detex(
+            yacc.yacc(
+                    module=Parser(filename),
+                    write_tables=0,
+                    debug=0,
+                ).parse(
+                    content,
+                    lexer=SongLexer().lexer,
+                    ).metadata
+            )

patacrep/songs/__init__.py

@@ -9,7 +9,7 @@ import pickle
 import re
 
 from patacrep.authors import processauthors
-from patacrep.content import Content
+from patacrep import files, encoding
 
 LOGGER = logging.getLogger(__name__)
 
@@ -95,6 +95,12 @@ class Song(Content):
             "_version",
             ]
 
+    # Default data
+    DEFAULT_DATA = {
+            '@titles': [],
+            '@languages': [],
+            }
+
     def __init__(self, datadir, subpath, config):
         self.fullpath = os.path.join(datadir, subpath)
         self.datadir = datadir
@@ -123,14 +129,14 @@ class Song(Content):
                         self.fullpath
                         ))
 
-        # Default values
-        self.data = {}
-        self.titles = []
-        self.languages = []
-        self.authors = []
-
-        # Parsing and data processing
-        self.parse()
+        # Data extraction from the latex song
+        self.data = self.DEFAULT_DATA
+        self.data['@path'] = self.fullpath
+        self.data.update(self.parse(
+                encoding.open_read(self.fullpath).read()
+                ))
+        self.titles = self.data['@titles']
+        self.languages = self.data['@languages']
         self.datadir = datadir
         self.unprefixed_titles = [
                 unprefixed_title(
@@ -169,49 +175,17 @@ class Song(Content):
     def __repr__(self):
         return repr((self.titles, self.data, self.fullpath))
 
-    def begin_new_block(self, previous, __context):
-        """Return a boolean stating if a new block is to be created."""
-        return not isinstance(previous, Song)
-
-    def begin_block(self, context):
-        """Return the string to begin a block."""
-        indexes = context.resolve("indexes")
-        if isinstance(indexes, jinja2.runtime.Undefined):
-            indexes = ""
-        return r'\begin{songs}{%s}' % indexes
-
-    def end_block(self, __context):
-        """Return the string to end a block."""
-        return r'\end{songs}'
-
-    def render(self, __context):
-        """Returns the TeX code rendering the song.
+    def tex(self, output): # pylint: disable=no-self-use, unused-argument
+        """Return the LaTeX code rendering this song.
 
-        This function is to be defined by subclasses.
+        Arguments:
+        - output: Name of the output file.
         """
-        return ''
+        return NotImplementedError()
 
-    def parse(self):
-        """Parse file `self.fullpath`.
-
-        This function is to be defined by subclasses.
-
-        It set the following attributes:
-
-        - titles: the list of (raw) titles. This list will be processed to
-          remove prefixes.
-        - languages: the list of languages used in the song, as languages
-          recognized by the LaTeX babel package.
-        - authors: the list of (raw) authors. This list will be processed to
-          'clean' it (see function :func:`patacrep.authors.processauthors`).
-        - data: song metadata. Used (among others) to sort the songs.
-        - cached: additional data that will be cached. Thus, data stored in
-          this attribute must be picklable.
-        """
-        self.data = {}
-        self.titles = []
-        self.languages = []
-        self.authors = []
+    def parse(self, content): # pylint: disable=no-self-use, unused-argument
+        """Parse song, and return a dictionary of its data."""
+        return NotImplementedError()
 
 def unprefixed_title(title, prefixes):
     """Remove the first prefix of the list in the beginning of title (if any).
@@ -221,4 +195,3 @@ def unprefixed_title(title, prefixes):
         if match:
             return match.group(2)
     return title
-

patacrep/songs/chordpro/__init__.py

@@ -0,0 +1,9 @@
+
+from patacrep.songs import Song
+
+class ChordproSong(Song):
+    pass
+
+SONG_PARSERS = {
+    'sgc': ChordproSong,
+    }

patacrep/chordpro/parser.py → patacrep/songs/chordpro/parser.py

@@ -4,8 +4,8 @@
 import logging
 import ply.yacc as yacc
 
-from patacrep.chordpro.lexer import tokens, ChordProLexer
-from patacrep.chordpro import ast
+from patacrep.songs.chordpro.lexer import tokens, ChordProLexer
+from patacrep.songs.chordpro import ast
 from patacrep.errors import SongbookError
 
 LOGGER = logging.getLogger()

patacrep/songs/latex/__init__.py

@@ -0,0 +1,32 @@
+"""Very simple LaTeX parser
+
+This module uses an LALR parser to try to parse LaTeX code. LaTeX language
+*cannot* be parsed by an LALR parser, so this is a very simple attemps, which
+will work on simple cases, but not on complex ones.
+"""
+
+import os
+
+from patacrep import files
+from patacrep.latex import parse_song
+from patacrep.songs import Song
+
+class LatexSong(Song):
+    """LaTeX song parser."""
+
+    def parse(self, content):
+        """Parse content, and return the dictinory of song data."""
+        return parse_song(content, self.fullpath)
+
+    def tex(self, output):
+        """Return the LaTeX code rendering the song."""
+        return r'\input{{{}}}'.format(files.path2posix(
+                                    files.relpath(
+                                        self.fullpath,
+                                        os.path.dirname(output)
+                                    )))
+
+SONG_PARSERS = {
+    'is': LatexSong,
+    'sg': LatexSong,
+    }