From 4892081b9950d760bea325e6afc2047ca1913219 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Hugo=20H=C3=B6rnquist?= <hugo@lysator.liu.se>
Date: Mon, 25 Sep 2023 16:08:32 +0200
Subject: [PATCH] Start splitting output back into modules.
The difference between this split, and the old, is that everything is
now clearly under output.
---
doc/muppet.rst | 3 +-
muppet/{output.py => output/__init__.py} | 783 +++--------------------
muppet/output/docstring.py | 258 ++++++++
muppet/output/puppet_source.py | 418 ++++++++++++
muppet/output/util.py | 44 ++
5 files changed, 808 insertions(+), 698 deletions(-)
rename muppet/{output.py => output/__init__.py} (58%)
create mode 100644 muppet/output/docstring.py
create mode 100644 muppet/output/puppet_source.py
create mode 100644 muppet/output/util.py
diff --git a/doc/muppet.rst b/doc/muppet.rst
index 5a724a0..88d90ba 100644
--- a/doc/muppet.rst
+++ b/doc/muppet.rst
@@ -7,6 +7,7 @@ Subpackages
.. toctree::
:maxdepth: 4
+ muppet.output
muppet.puppet
muppet.syntax_highlight
@@ -21,11 +22,9 @@ Submodules
muppet.intersperse
muppet.lookup
muppet.markdown
- muppet.output
muppet.parser_combinator
muppet.symbols
muppet.tabs
- muppet.templates
muppet.util
Module contents
diff --git a/muppet/output.py b/muppet/output/__init__.py
similarity index 58%
rename from muppet/output.py
rename to muppet/output/__init__.py
index 2d3440a..3e1b59a 100644
--- a/muppet/output.py
+++ b/muppet/output/__init__.py
@@ -11,8 +11,8 @@ import os.path
import pathlib
import re
import json
+from glob import glob
from typing import (
- Any,
Optional,
Protocol,
Sequence,
@@ -28,18 +28,7 @@ from muppet.puppet.strings import (
DataTypeAlias,
DefinedType,
DocString,
- DocStringApiTag,
- DocStringAuthorTag,
- DocStringExampleTag,
- DocStringOptionTag,
- DocStringOverloadTag,
DocStringParamTag,
- DocStringRaiseTag,
- DocStringReturnTag,
- DocStringSeeTag,
- DocStringSinceTag,
- DocStringSummaryTag,
- DocStringTag,
Function,
PuppetClass,
PuppetStrings,
@@ -48,20 +37,16 @@ from muppet.puppet.strings import (
puppet_strings_cached,
)
from muppet.parser_combinator import (
- ParserCombinator,
ParseError,
- MatchCompound,
- MatchObject,
)
from muppet.markdown import markdown
from muppet.breadcrumbs import Breadcrumbs, breadcrumbs
from muppet.util import group_by, partition
from muppet.cache import AbstractCache
-from muppet.puppet.parser import puppet_parser
-from muppet.puppet.ast import build_ast
-from muppet.puppet.format.parser import ParserFormatter
+from .docstring import format_docstring
+from .puppet_source import hyperlink_puppet_source
jinja = Environment(
@@ -73,11 +58,32 @@ jinja = Environment(
logger = logging.getLogger(__name__)
-param_doc: dict[str, str] = {}
+class HtmlSerializable(Protocol):
+ """Classes which can be serialized as HTML."""
+
+ def to_html(self) -> str: # pragma: no cover
+ """Return HTML string."""
+ ...
+
+ def to_html_list(self) -> str: # pragma: no cover
+ """Return HTML suitable for a list."""
+ ...
class Templates:
- """Namespace for templates."""
+ """
+ Namespace for templates.
+
+ Almost all of these methods take these values:
+
+ :param path_base:
+ Prefix added to all links within output
+ TODO shouldn't this be bound to the object, since it should
+ NEVER change (during a run).
+
+ :path breadcrumbs:
+ Breadcrumb to current page.
+ """
def __init__(self) -> None:
self.jinja = Environment(
@@ -90,7 +96,15 @@ class Templates:
path_base: str,
breadcrumbs: Optional[Breadcrumbs] = None
) -> str: # pragma: no cover
- """Template for a page containing puppet code."""
+ """
+ Template for a page containing puppet code.
+
+ :param content:
+ Top level h1 tag of page.
+
+ :param content:
+ Free form string content, being the main body of the page.
+ """
template = self.jinja.get_template('code_page.html')
return template.render(
title=title,
@@ -103,7 +117,12 @@ class Templates:
path_base: str,
breadcrumbs: Optional[Breadcrumbs] = None
) -> str: # pragma: no cover
- """Template for a page with arbitrary content."""
+ """
+ Template for a page with arbitrary content.
+
+ :param content:
+ Arbitrary content
+ """
template = self.jinja.get_template('content.html')
return template.render(
content=content,
@@ -115,7 +134,12 @@ class Templates:
path_base: str,
breadcrumbs: Optional[Breadcrumbs] = None
) -> str: # pragma: no cover
- """Root index file."""
+ """
+ Environment index file.
+
+ :param modules:
+ List of PuppetModules to include in the environment index.
+ """
template = self.jinja.get_template('index.html')
return template.render(
path_base=path_base,
@@ -124,7 +148,7 @@ class Templates:
def module_index(self, *,
# content: list[], # something with to_html_list and to_html
- content: list[Any], # TODO something with to_html_list and to_html
+ content: Sequence[HtmlSerializable],
module_author: str,
module_name: str,
doc_files: list[tuple[str, str]],
@@ -133,7 +157,33 @@ class Templates:
left_sidebar: Optional[str] = None,
right_sidebar: Optional[str] = None,
) -> str: # pragma: no cover
- """Index for a single module."""
+ """
+ Index for a single module.
+
+ :param content:
+ Main content of the page.
+ :param module_author:
+ Author of the module, as it appears in the metadata.json file.
+ :param module_name:
+ Name of the module, without the author component.
+ :param doc_files:
+ The free-form documentation files bundled with the module.
+
+ Each element should be a pair of
+ - The idealized name of the file
+ - The relative path to the document inside the output
+ (the HTML generated version)
+
+ :param left_sidebar:
+ Free form content of the left sidebar.
+
+ This is assumed to be a list of modules in the environment.
+ :param right_sidebar:
+ Free form content of the right sidebar.
+
+ This is assumed to be a table of contents of the module,
+ really similar to the actual body contents.
+ """
template = self.jinja.get_template('module_index.html')
return template.render(
content=content,
@@ -149,18 +199,6 @@ class Templates:
templates = Templates()
-class HtmlSerializable(Protocol):
- """Classes which can be serialized as HTML."""
-
- def to_html(self) -> str: # pragma: no cover
- """Return HTML string."""
- ...
-
- def to_html_list(self) -> str: # pragma: no cover
- """Return HTML suitable for a list."""
- ...
-
-
@dataclass
class ResourceTypeOutput:
"""Basic HTML implementation."""
@@ -381,8 +419,10 @@ class PuppetModule:
self.toc = self._build_module_toc()
self.output_prefix = output_prefix
- # TODO
- self.doc_files: list[str] = []
+ abspath = os.path.abspath(self.path)
+ self.doc_files: list[str] = \
+ glob(os.path.join(abspath, '*.md')) + \
+ glob(os.path.join(abspath, 'LICENSE'))
try:
with open(os.path.join(self.path, 'metadata.json')) as f:
@@ -390,7 +430,7 @@ class PuppetModule:
except FileNotFoundError:
self.metadata = {}
- def _build_module_toc(self) -> list[ResourceIndex | IndexCategory]:
+ def _build_module_toc(self) -> Sequence[ResourceIndex | IndexCategory]:
"""Build the TOC of the module."""
content: list[ResourceIndex | IndexCategory] = []
@@ -482,8 +522,8 @@ class PuppetModule:
breadcrumbs=crumbs,
content=toc,
path_base=self.output_prefix,
- # doc_files=list(doc_files.items())
# TODO
+ # doc_files=self.doc_files,
doc_files=[],
# left_sidebar=(),
right_sidebar=''.join([
@@ -872,158 +912,6 @@ def index_item(obj: PuppetClass | DefinedType) -> IndexItem:
return out
-def format_docstring(name: str, docstring: DocString) -> tuple[str, str]:
- """
- Format docstrings as they appear in some puppet types.
-
- Those types being:
-
- * puppet_classes,
- * puppet_type_aliases, and
- * defined_types
- """
- global param_doc
-
- # The api tag is ignored, since it instead is shown from context
-
- out = ''
-
- param_doc = {tag.name: tag.text or ''
- for tag in docstring.tags
- if isinstance(tag, DocStringParamTag)}
-
- grouped_tags = GroupedTags.from_taglist(docstring.tags)
-
- # --------------------------------------------------
-
- out += '<a href="#code">Jump to Code</a><br/>'
-
- if tags := grouped_tags.summary_:
- out += '<em class="summary">'
- for tag in tags:
- out += html.escape(tag.text)
- out += '</em>'
-
- out += '<div class="description">'
- # TODO "TODO" highlighting
- out += markdown(docstring.text)
- out += '</div>'
-
- # TODO proper handling of multiple @see tags
- if sees := grouped_tags.see_:
- out += '<b>See</b> '
- for see in sees:
- link: str
- m = re.match(r'((?P<url>https?://.*)|(?P<man>.*\([0-9]\))|(?P<other>.*))', see.name)
- assert m, "Regex always matched"
- if m['url']:
- link = f'<a href="{see.name}">{see.name}</a>'
- out += link
- elif m['man']:
- page = see.name[:-3]
- section = see.name[-2]
- # TODO man providers
- link = f"https://manned.org/man/{page}.{section}"
- out += link
- else:
- if '::' in m['other']:
- # TODO
- pass
- else:
- # TODO
- # link = see
- pass
- out += m['other']
- out += ' ' + see.text
-
- if authors := grouped_tags.author_:
- out += '<div class="author">'
- out += "<em>Written by </em>"
- if len(authors) == 1:
- out += parse_author(authors[0].text)
- else:
- out += '<ul>'
- for author in authors:
- out += f'<li>{parse_author(author.text)}</li>'
- out += '</ul>'
- out += '</div>'
-
- out += '<hr/>'
-
- t: DocStringTag
-
- for t in grouped_tags .example_:
- out += '<div class="code-example">'
-
- if name := t.name:
- # TODO markup for title
- out += f'<div class="code-example-header">{html.escape(name)}</div>\n'
- # TODO highlight?
- # Problem is that we don't know what language the example
- # is in. Pygemntize however seems to do a reasonable job
- # treating anything as puppet code
- text = html.escape(t.text)
- out += f'<pre><code class="puppet">{text}</code></pre>\n'
- out += '</div>'
-
- out += '<hr/>'
-
- out += '<dl>'
- for t in grouped_tags.param_:
- name = html.escape(t.name)
- out += f'<dt><span id="{name}" class="variable">{name}</span>'
- match t.types:
- case [x]:
- # TODO highlight type?
- out += f': <code>{html.escape(x)}</code>'
- case [_, *_]:
- raise ValueError("How did you get multiple types onto a parameter?")
-
- # TODO Fetch default values from puppet strings output
- # Then in javascript query Hiera to get the true "default"
- # values for a given machine (somewhere have a setting for
- # selecting machine).
- out += '</dt>'
-
- if text := t.text:
- text = re.sub(r'(NOTE|TODO)',
- r'<mark>\1</mark>',
- markdown(text))
-
- if options := grouped_tags.option_.get(t.name):
- text += '<dl>'
- for option in options:
- text += '<dt>'
- text += html.escape(option.opt_name)
- match option.opt_types:
- case [x]:
- text += f' [<code>{html.escape(x)}</code>]'
- case [_, *_]:
- raise ValueError("How did you get multiple types onto an option?")
- text += '</dt>'
- text += '<dd>'
- if option.opt_text:
- text += re.sub(r'(NOTE|TODO)',
- r'<mark>\1</mark>',
- markdown(option.opt_text))
- text += '</dd>'
- text += '</dl>'
-
- out += f"<dd>{text}</dd>"
- else:
- out += '<dd><em>Undocumented</em></dd>'
- out += '</dl>'
-
- # TODO remaining tags
- # "overload"
- # raise
- # return
- # since
- # _other
-
- return (name, out)
-
-
def format_class(d_type: DefinedType | PuppetClass) -> tuple[str, str]:
"""Format Puppet class."""
out = ''
@@ -1033,7 +921,7 @@ def format_class(d_type: DefinedType | PuppetClass) -> tuple[str, str]:
out += body
# ------ Old ---------------------------------------
- # t = parse_puppet(d_type.source)
+ # t = hyperlink_puppet_source(d_type.source)
# data = parse(t, 0, ['root'])
# renderer = HTMLRenderer(build_param_dict(d_type.docstring))
# out += render(renderer, data)
@@ -1047,7 +935,8 @@ def format_class(d_type: DefinedType | PuppetClass) -> tuple[str, str]:
in_parameters.append(cast(DocStringParamTag, tag).name)
try:
- result = parse_puppet(d_type.source, d_type.name, in_parameters)
+ # Calculation beforehand, for "atomic" formatting
+ result = hyperlink_puppet_source(d_type.source, d_type.name, in_parameters)
out += '<pre class="highlight-muppet"><code class="puppet">'
out += result
out += '</code></pre>'
@@ -1068,53 +957,6 @@ def format_class(d_type: DefinedType | PuppetClass) -> tuple[str, str]:
return name, out
-@dataclass
-class GroupedTags:
- """
- All tags from a class (or similar) docstring.
-
- Most fields are simply lists of tags. The reason for trailing
- underscores on each entry is since some tag names collide with
- python keywords (e.g. ``raise``).
- """
-
- param_: list[DocStringParamTag] = field(default_factory=list) # noqa: E221
- example_: list[DocStringExampleTag] = field(default_factory=list) # noqa: E221
- overload_: list[DocStringOverloadTag] = field(default_factory=list) # noqa: E221
- option_: dict[str, list[DocStringOptionTag]] = field(default_factory=dict) # noqa: E221
- """
- Options document Hash parameters valid values.
-
- Each key is the corresponding parameter, and the value is the list
- of registered options for that hash.
- """
-
- author_: list[DocStringAuthorTag] = field(default_factory=list) # noqa: E221
- api_: list[DocStringApiTag] = field(default_factory=list) # noqa: E221
- raise_: list[DocStringRaiseTag] = field(default_factory=list) # noqa: E221
- return_: list[DocStringReturnTag] = field(default_factory=list) # noqa: E221
- since_: list[DocStringSinceTag] = field(default_factory=list) # noqa: E221
- summary_: list[DocStringSummaryTag] = field(default_factory=list) # noqa: E221
- see_: list[DocStringSeeTag] = field(default_factory=list) # noqa: E221
- other_: list[DocStringTag] = field(default_factory=list) # noqa: E221
- """All tags of unknown type."""
-
- @classmethod
- def from_taglist(cls, tags: list[DocStringTag]) -> 'GroupedTags':
- """Group a list of tags."""
- grouped_tags = cls()
- for tag in tags:
- if tag.tag_name == 'option':
- tag = cast(DocStringOptionTag, tag)
- grouped_tags.option_.setdefault(tag.parent, []).append(tag)
- elif tag.tag_name in {'param', 'example', 'overload', 'author', 'api',
- 'raise', 'return', 'since', 'summary', 'see'}:
- getattr(grouped_tags, tag.tag_name + '_').append(tag)
- else:
- grouped_tags.other_.append(tag)
- return grouped_tags
-
-
def build_param_dict(docstring: DocString) -> dict[str, str]:
"""
Extract all parameter documentation from a docstring dict.
@@ -1155,7 +997,7 @@ def format_type_alias(d_type: DataTypeAlias, file: str) -> tuple[str, str]:
out += '\n'
out += '<pre class="highlight-muppet"><code class="puppet">'
try:
- out += parse_puppet(d_type.alias_of, file, [])
+ out += hyperlink_puppet_source(d_type.alias_of, file, [])
except ParseError as e:
logger.error("Parsing %(name)s failed: %(err)s",
{'name': d_type.alias_of, 'err': e})
@@ -1175,7 +1017,7 @@ def format_defined_type(d_type: DefinedType, file: str) -> tuple[str, str]:
out += '<pre class="highlight-muppet"><code class="puppet">'
try:
- out += parse_puppet(d_type.source, file, [])
+ out += hyperlink_puppet_source(d_type.source, file, [])
except ParseError as e:
logger.error("Parsing %(name)s failed: %(err)s",
{'name': d_type.source, 'err': e})
@@ -1240,7 +1082,7 @@ def format_puppet_function(function: Function, file: str) -> str:
elif t == 'puppet':
out += '<pre class="highlight-muppet"><code class="puppet">'
try:
- out += parse_puppet(function.source, file, [])
+ out += hyperlink_puppet_source(function.source, file, [])
except ParseError as e:
logger.error("Parsing %(name)s failed: %(err)s",
{'name': function.source, 'err': e})
@@ -1261,454 +1103,3 @@ def format_puppet_task() -> str:
def format_puppet_plan() -> str:
"""Format Puppet plan."""
return 'TODO format_puppet_plan not implemented'
-
-
-def inner_text(obj: MatchObject | list[MatchObject]) -> str:
- """
- Extract the text content from a set of MatchObjects.
-
- This is really similar to HTML's inner_text.
-
- Empty whitespace tags are expanded into nothing, non-empty
- whitespace tags becomes a single space (note that this discards
- commets).
-
- This only works properly if no function was mapped over the parser
- return values in tree, see :func:`muppet.parser_combinator.fmap`.
-
- :param obj:
- Match Objects to search.
- """
- match obj:
- case str(s):
- return s
- case MatchCompound(type='ws', matched=[]):
- return ''
- case MatchCompound(type='ws'):
- return ' '
- case MatchCompound(matched=xs):
- return ''.join(inner_text(x) for x in xs)
- case [*xs]:
- return ''.join(inner_text(x) for x in xs)
- case _:
- raise ValueError('How did we get here')
-
-
-def name_to_url(name: str) -> tuple[str | None, str]:
- """
- Resolve a class or resource name into an url.
-
- :param name:
- The name of a class or resource, surch as "example::resource".
- :return:
- A tuple consisting of
-
- - One of
- - An internal link to the definition of that type
- - A link to the official puppet documentation
- - ``None``, if `name` is "class"
- - A string indicating extra HTML classes for this url.
- This is mostly so external references can be marked properly.
- """
- if name in built_in_types:
- return (f'https://www.puppet.com/docs/puppet/7/types/{name}.html', 'puppet-doc')
- elif name == 'class':
- return (None, '')
- else:
- # TODO special cases for puppet's built in types.
- # https://www.puppet.com/docs/puppet/7/cheatsheet_core_types.html
- module, *items = name.lstrip(':').split('::')
- # TODO get prefix from the command line/config file
- return ('/code/muppet-strings/output/'
- + '/'.join([module, 'manifests', *(items if items else ['init'])]),
- '')
-
-
-puppet_doc_base = 'https://www.puppet.com/docs/puppet/7'
-lang_facts_builtin_variables = (f'{puppet_doc_base}/lang_facts_builtin_variables'
- '#lang_facts_builtin_variables')
-server_variables = f'{lang_facts_builtin_variables}-server-variables'
-compiler_variables = f'{lang_facts_builtin_variables}-compiler-variables'
-trusted_facts = f'{lang_facts_builtin_variables}-trusted-facts'
-server_facts = f'{lang_facts_builtin_variables}-server-facts'
-
-built_in_variables = {
- 'facts': 'https://google.com',
- # clientcert, clientversion, puppetversion, clientnoop,
- # agent_specified_environment:
- # https://www.puppet.com/docs/puppet/7/lang_facts_builtin_variables#lang_facts_builtin_variables-agent-facts
- 'trusted': trusted_facts,
- 'server_facts': server_facts,
- 'environment': server_variables,
- 'servername': server_variables,
- 'serverip': server_variables,
- 'serverversion': server_variables,
- 'module_name': compiler_variables,
- 'caller_module_name': compiler_variables,
-
- # Also note the special variable $title and $name
- # https://www.puppet.com/docs/puppet/7/lang_defined_types#lang_defined_types-title-and-name
-}
-
-
-def parse_author(author: str) -> str:
- """
- Format author tags' content.
-
- :param author:
- The contents of the author tag. If the string is on the
- regular "author" format of ``"Firstname Lastname
- <first.last@example.com>"`` then the email will be formatted
- and hyperlinked. Otherwise the string is returned verbatim.
- :return:
- An HTML safe string, possibly including tags.
- """
- m = re.match(r'(?P<author>.*) (<(?P<email>.*)>)|(?P<any>.*)', author)
- assert m, "The above regex can't fail"
- if m['author'] and m['email']:
- author = html.escape(m['author'])
- email = html.escape(m['email'])
- return f'{author} <a class="email" href="mailto:{email}"><{email}></a>;'
- else:
- return html.escape(m['any'])
-
-
-# https://www.puppet.com/docs/puppet/7/cheatsheet_core_types.html
-# https://www.puppet.com/docs/puppet/7/types/file.html
-# ...
-built_in_types = {
- 'package',
- 'file',
- 'service',
- 'notify',
- 'exec',
- 'user',
- 'group',
-}
-
-# https://www.puppet.com/docs/puppet/7/function.html#{}
-built_in_functions = {
- 'abs',
- 'alert',
- 'all',
- 'annotate',
- 'any',
- 'assert_type',
- 'binary_file',
- 'break',
- 'call',
- 'camelcase',
- 'capitalize',
- 'ceiling',
- 'chomp',
- 'chop',
- 'compare',
- 'contain',
- 'convert_to',
- 'create_resources',
- 'crit',
- 'debug',
- 'defined',
- 'dig',
- 'digest',
- 'downcase',
- 'each',
- 'emerg',
- 'empty',
- 'epp',
- 'err',
- 'eyaml_lookup_key',
- 'fail',
- 'file',
- 'filter',
- 'find_file',
- 'find_template',
- 'flatten',
- 'floor',
- 'fqdn_rand',
- 'generate',
- 'get',
- 'getvar',
- 'group_by',
- 'hiera',
- 'hiera_array',
- 'hiera_hash',
- 'hiera_include',
- 'hocon_data',
- 'import',
- 'include',
- 'index',
- 'info',
- 'inline_epp',
- 'inline_template',
- 'join',
- 'json_data',
- 'keys',
- 'length',
- 'lest',
- 'lookup',
- 'lstrip',
- 'map',
- 'match',
- 'max',
- 'md5',
- 'min',
- 'module_directory',
- 'new',
- 'next',
- 'notice',
- 'partition',
- 'realize',
- 'reduce',
- 'regsubst',
- 'require',
- 'return',
- 'reverse_each',
- 'round',
- 'rstrip',
- 'scanf',
- 'sha1',
- 'sha256',
- 'shellquote',
- 'size',
- 'slice',
- 'sort',
- 'split',
- 'sprintf',
- 'step',
- 'strftime',
- 'strip',
- 'tag',
- 'tagged',
- 'template',
- 'then',
- 'tree_each',
- 'type',
- 'unique',
- 'unwrap',
- 'upcase',
- 'values',
- 'versioncmp',
- 'warning',
- 'with',
- 'yaml_data',
-}
-
-
-def find_declarations(objs: list[MatchObject]) -> list[str]:
- """
- Find all local variable declarations.
-
- Searches the code for all local variable declarations, returing a
- list of variable names.
-
- Note that the same variable might appear multiple times, for example:
-
- .. code-block:: puppet
- :caption: The same variable being declared twice
-
- if $something {
- $x = 10
- } else {
- $x = 20
- }
- """
- declarations = []
- for obj in objs:
- match obj:
- case MatchCompound(type='declaration', matched=xs):
- for x in xs:
- match x:
- case MatchCompound(type='var', matched=ys):
- declarations.append(inner_text(ys))
- return declarations
-
-
-class Reserializer:
- """
- Context for reserializing parsed data back into code.
-
- :param local_vars:
- Variables declared within this file. Used when resolving
- hyperlinks.
- """
-
- def __init__(self, local_vars: list[str]):
- self.local_vars: list[str] = local_vars
-
- def reserialize(self,
- obj: MatchObject | Sequence[MatchObject]) -> str:
- """
- Reconstruct puppet code after parsing it.
-
- After building the parser, and parsing the puppet code into a tree
- of MatchObjects; this procedure returns it into puppet code.
- Difference being that we now have metadata, meaning that syntax
- highlighting and variable hyperlinks can be inserted.
-
- :param obj:
- Should be assumed to be a list of MatchObject's, or something similar.
-
- MatchCompound objects are serialized as
-
- .. code-block:: html
-
- <span class="{type}">{body}</span>
-
- strings as themselves, and lists have reserialize mapped over them.
-
- """
- out: list[str] = []
- # logger.info("obj = %a", obj)
-
- # TODO hyperlink functions.
- # The problem is that a function can either be implemented in
- # Puppet, or in Ruby. And Ruby functions' names aren't bound
- # by the directory layout.
- match obj:
- case str(s):
- out.append(html.escape(s))
-
- case MatchCompound(type='resource-name', matched=xs):
- name = inner_text(xs)
- url, cls = name_to_url(name)
- if url:
- out.append(f'<a href="{url}" class="resource-name {cls}">{name}</a>')
- else:
- # TODO this is class, but the class name should
- # also be hyperlinked
- out.append(f'<span class="resource-name {cls}">{name}</span>')
-
- case MatchCompound(type='invoke', matched=xs):
- function = None
- for x in xs:
- match x:
- case MatchCompound(type='qn', matched=ys):
- if function is None:
- function = inner_text(ys)
- if function in built_in_functions:
- # class="qn"
- url = f"https://www.puppet.com/docs/puppet/7/function.html#{function}" # noqa: E501
- tag = f'<a href="{url}" class="puppet-doc">{self.reserialize(ys)}</a>' # noqa: E501
- out.append(tag)
- else:
- # TODO function to url
- out.append(f'<span class="qn">{self.reserialize(ys)}</span>')
- else:
- if function == 'include':
- url, cls = name_to_url(inner_text(ys))
- # class="qn"
- tag = f'<a href="{url}" class="{cls}">{self.reserialize(ys)}</a>' # noqa: E501
- out.append(tag)
- else:
- out.append(self.reserialize(ys))
- case _:
- out.append(self.reserialize(x))
-
- case MatchCompound(type='declaration', matched=xs):
- for x in xs:
- match x:
- case MatchCompound(type='var', matched=ys):
- inner = ''.join(self.reserialize(y) for y in ys)
- out.append(f'<span id="{inner_text(ys)}">{inner}</span>')
- case _:
- out.append(self.reserialize(x))
-
- case MatchCompound(type='var', matched=xs):
- out.append(self.var_to_url(inner_text(xs)))
-
- case MatchCompound(type=type, matched=xs):
- body = ''.join(self.reserialize(x) for x in xs)
- out.append(f'<span class="{type}">{body}</span>')
-
- case [*xs]:
- out.extend(self.reserialize(x) for x in xs)
-
- case rest:
- logger.error("Unknown type: %a", rest)
-
- return ''.join(out)
-
- def var_to_url(self, var: str) -> str:
- """
- Format variable, adding hyperlink to its definition.
-
- TODO these can refer to both defined types (`manifests/*.pp`),
- as well as resource types (`lib/puppet/provider/*/*.rb` /
- `lib/tpuppet/type/*.rb`)
-
- Same goes for functions (`functions/*.pp`),
- (`lib/puppet/functions.rb`).
-
- :param var:
- Name of the variable.
-
- :return:
- An HTML anchor element.
- """
- match var.split('::'):
- case [name]:
- # Either a local or global variable
- # https://www.puppet.com/docs/puppet/7/lang_facts_and_builtin_vars.html
-
- href = None
- cls = ''
- if name in self.local_vars:
- href = f'#{html.escape(var)}'
- elif name in built_in_variables:
- href = html.escape(built_in_variables[name])
- cls = 'puppet-doc'
-
- if href:
- return f'<a class="var {cls}" href="{href}">{var}</a>'
- else:
- # `name` refers to a global fact.
- return f'<span class="var">{var}</span>'
-
- case ['', name]:
- # A global variable
- if name in built_in_variables:
- href = html.escape(built_in_variables[name])
- img = '<img src="/code/muppet-strings/output/static/favicon.ico" />'
- return f'<a class="var" href="{href}">{var}{img}</a>'
- else:
- return f'<span class="var">{var}</span>'
-
- # Note the "special module" 'settings',
- # https://www.puppet.com/docs/puppet/7/lang_facts_builtin_variables#lang_facts_builtin_variables-server-variables
- case ['', module, *items, name]:
- s = '/code/muppet-strings/output/' \
- + '/'.join([module, 'manifests', *(items if items else ['init'])])
- s += f'#{name}'
- return f'<a class="var" href="{s}">{var}</a>'
- case [module, *items, name]:
- s = '/code/muppet-strings/output/' \
- + '/'.join([module, 'manifests', *(items if items else ['init'])])
- s += f'#{name}'
- return f'<a class="var" href="{s}">{var}</a>'
- case _:
- raise ValueError()
-
-
-def parse_puppet(source: str, file: str, in_parameters: list[str]) -> str:
- """
- Parse and syntax highlight the given puppet source.
-
- :returns: An HTML string
- """
- # Run the upstream puppet parser,
- # then masage the tree into a usable form.
- ast = build_ast(puppet_parser(source))
-
- # From the ast, build a parser combinator parser.
- # This parser will attach sufficient metadata to allow syntax
- # highlighting and hyperlinking
- parser = ParserFormatter().serialize(ast)
-
- # Run the generated parser, giving us a list of match objects.
- match_objects = ParserCombinator(source, file).get(parser)
-
- # Reserialize the matched data back into puppet code, realizing
- # the syntax highlighting and hyperlinks.
- return Reserializer(find_declarations(match_objects) + (in_parameters)) \
- .reserialize(match_objects)
diff --git a/muppet/output/docstring.py b/muppet/output/docstring.py
new file mode 100644
index 0000000..b85676c
--- /dev/null
+++ b/muppet/output/docstring.py
@@ -0,0 +1,258 @@
+"""
+Generate output for Puppet Docstrings.
+
+Docstrings are the functions preceeding any top level puppet
+declaration (such as classes, rosource definitions, ...). These have a
+number of "magic" tags for attaching metadata, along with usually
+being Markdown formatted. This module assumes that they all are
+Markdown formatted, which unfortunately leads to some (minor) errors.
+
+(The final output also contains the original source, allowing these
+errors to be overlooked).
+"""
+
+from dataclasses import dataclass, field
+import html
+import re
+from typing import cast
+from muppet.markdown import markdown
+from muppet.puppet.strings import (
+ DocString,
+ DocStringApiTag,
+ DocStringAuthorTag,
+ DocStringExampleTag,
+ DocStringOptionTag,
+ DocStringOverloadTag,
+ DocStringParamTag,
+ DocStringRaiseTag,
+ DocStringReturnTag,
+ DocStringSeeTag,
+ DocStringSinceTag,
+ DocStringSummaryTag,
+ DocStringTag,
+)
+
+
+# TODO what even is this for?
+param_doc: dict[str, str] = {}
+
+
+@dataclass
+class GroupedTags:
+ """
+ All tags from a class (or similar) docstring.
+
+ Most fields are simply lists of tags. The reason for trailing
+ underscores on each entry is since some tag names collide with
+ python keywords (e.g. ``raise``).
+ """
+
+ param_: list[DocStringParamTag] = field(default_factory=list) # noqa: E221
+ example_: list[DocStringExampleTag] = field(default_factory=list) # noqa: E221
+ overload_: list[DocStringOverloadTag] = field(default_factory=list) # noqa: E221
+ option_: dict[str, list[DocStringOptionTag]] = field(default_factory=dict) # noqa: E221
+ """
+ Options document Hash parameters valid values.
+
+ Each key is the corresponding parameter, and the value is the list
+ of registered options for that hash.
+ """
+
+ author_: list[DocStringAuthorTag] = field(default_factory=list) # noqa: E221
+ api_: list[DocStringApiTag] = field(default_factory=list) # noqa: E221
+ raise_: list[DocStringRaiseTag] = field(default_factory=list) # noqa: E221
+ return_: list[DocStringReturnTag] = field(default_factory=list) # noqa: E221
+ since_: list[DocStringSinceTag] = field(default_factory=list) # noqa: E221
+ summary_: list[DocStringSummaryTag] = field(default_factory=list) # noqa: E221
+ see_: list[DocStringSeeTag] = field(default_factory=list) # noqa: E221
+ other_: list[DocStringTag] = field(default_factory=list) # noqa: E221
+ """All tags of unknown type."""
+
+ @classmethod
+ def from_taglist(cls, tags: list[DocStringTag]) -> 'GroupedTags':
+ """Group a list of tags."""
+ grouped_tags = cls()
+ for tag in tags:
+ if tag.tag_name == 'option':
+ tag = cast(DocStringOptionTag, tag)
+ grouped_tags.option_.setdefault(tag.parent, []).append(tag)
+ elif tag.tag_name in {'param', 'example', 'overload', 'author', 'api',
+ 'raise', 'return', 'since', 'summary', 'see'}:
+ getattr(grouped_tags, tag.tag_name + '_').append(tag)
+ else:
+ grouped_tags.other_.append(tag)
+ return grouped_tags
+
+
+def parse_author(author: str) -> str:
+ """
+ Format author tags' content.
+
+ :param author:
+ The contents of the author tag. If the string is on the
+ regular "author" format of ``"Firstname Lastname
+ <first.last@example.com>"`` then the email will be formatted
+ and hyperlinked. Otherwise the string is returned verbatim.
+ :return:
+ An HTML safe string, possibly including tags.
+ """
+ m = re.match(r'(?P<author>.*) (<(?P<email>.*)>)|(?P<any>.*)', author)
+ assert m, "The above regex can't fail"
+ if m['author'] and m['email']:
+ author = html.escape(m['author'])
+ email = html.escape(m['email'])
+ return f'{author} <a class="email" href="mailto:{email}"><{email}></a>;'
+ else:
+ return html.escape(m['any'])
+
+
+def format_docstring(name: str, docstring: DocString) -> tuple[str, str]:
+ """
+ Format docstrings as they appear in some puppet types.
+
+ Those types being:
+
+ * puppet_classes,
+ * puppet_type_aliases, and
+ * defined_types
+ """
+ global param_doc
+
+ # The api tag is ignored, since it instead is shown from context
+
+ out = ''
+
+ param_doc = {tag.name: tag.text or ''
+ for tag in docstring.tags
+ if isinstance(tag, DocStringParamTag)}
+
+ grouped_tags = GroupedTags.from_taglist(docstring.tags)
+
+ # --------------------------------------------------
+
+ out += '<a href="#code">Jump to Code</a><br/>'
+
+ if tags := grouped_tags.summary_:
+ out += '<em class="summary">'
+ for tag in tags:
+ out += html.escape(tag.text)
+ out += '</em>'
+
+ out += '<div class="description">'
+ # TODO "TODO" highlighting
+ out += markdown(docstring.text)
+ out += '</div>'
+
+ # TODO proper handling of multiple @see tags
+ if sees := grouped_tags.see_:
+ out += '<b>See</b> '
+ for see in sees:
+ link: str
+ m = re.match(r'((?P<url>https?://.*)|(?P<man>.*\([0-9]\))|(?P<other>.*))', see.name)
+ assert m, "Regex always matched"
+ if m['url']:
+ link = f'<a href="{see.name}">{see.name}</a>'
+ out += link
+ elif m['man']:
+ page = see.name[:-3]
+ section = see.name[-2]
+ # TODO man providers
+ link = f"https://manned.org/man/{page}.{section}"
+ out += link
+ else:
+ if '::' in m['other']:
+ # TODO
+ pass
+ else:
+ # TODO
+ # link = see
+ pass
+ out += m['other']
+ out += ' ' + see.text
+
+ if authors := grouped_tags.author_:
+ out += '<div class="author">'
+ out += "<em>Written by </em>"
+ if len(authors) == 1:
+ out += parse_author(authors[0].text)
+ else:
+ out += '<ul>'
+ for author in authors:
+ out += f'<li>{parse_author(author.text)}</li>'
+ out += '</ul>'
+ out += '</div>'
+
+ out += '<hr/>'
+
+ t: DocStringTag
+
+ for t in grouped_tags .example_:
+ out += '<div class="code-example">'
+
+ if name := t.name:
+ # TODO markup for title
+ out += f'<div class="code-example-header">{html.escape(name)}</div>\n'
+ # TODO highlight?
+ # Problem is that we don't know what language the example
+ # is in. Pygemntize however seems to do a reasonable job
+ # treating anything as puppet code
+ text = html.escape(t.text)
+ out += f'<pre><code class="puppet">{text}</code></pre>\n'
+ out += '</div>'
+
+ out += '<hr/>'
+
+ out += '<dl>'
+ for t in grouped_tags.param_:
+ name = html.escape(t.name)
+ out += f'<dt><span id="{name}" class="variable">{name}</span>'
+ match t.types:
+ case [x]:
+ # TODO highlight type?
+ out += f': <code>{html.escape(x)}</code>'
+ case [_, *_]:
+ raise ValueError("How did you get multiple types onto a parameter?")
+
+ # TODO Fetch default values from puppet strings output
+ # Then in javascript query Hiera to get the true "default"
+ # values for a given machine (somewhere have a setting for
+ # selecting machine).
+ out += '</dt>'
+
+ if text := t.text:
+ text = re.sub(r'(NOTE|TODO)',
+ r'<mark>\1</mark>',
+ markdown(text))
+
+ if options := grouped_tags.option_.get(t.name):
+ text += '<dl>'
+ for option in options:
+ text += '<dt>'
+ text += html.escape(option.opt_name)
+ match option.opt_types:
+ case [x]:
+ text += f' [<code>{html.escape(x)}</code>]'
+ case [_, *_]:
+ raise ValueError("How did you get multiple types onto an option?")
+ text += '</dt>'
+ text += '<dd>'
+ if option.opt_text:
+ text += re.sub(r'(NOTE|TODO)',
+ r'<mark>\1</mark>',
+ markdown(option.opt_text))
+ text += '</dd>'
+ text += '</dl>'
+
+ out += f"<dd>{text}</dd>"
+ else:
+ out += '<dd><em>Undocumented</em></dd>'
+ out += '</dl>'
+
+ # TODO remaining tags
+ # "overload"
+ # raise
+ # return
+ # since
+ # _other
+
+ return (name, out)
diff --git a/muppet/output/puppet_source.py b/muppet/output/puppet_source.py
new file mode 100644
index 0000000..d3b9f4d
--- /dev/null
+++ b/muppet/output/puppet_source.py
@@ -0,0 +1,418 @@
+"""Generate output for Puppet Source code."""
+
+import html
+import logging
+from typing import Sequence
+
+from muppet.parser_combinator import (
+ ParserCombinator,
+ MatchCompound,
+ MatchObject,
+)
+from muppet.puppet.ast import build_ast
+from muppet.puppet.parser import puppet_parser
+from muppet.puppet.format.parser import ParserFormatter
+
+from .util import inner_text
+
+
+logger = logging.getLogger(__name__)
+
+
+_puppet_doc_base = 'https://www.puppet.com/docs/puppet/7'
+_lang_facts_builtin_variables = (f'{_puppet_doc_base}/lang_facts_builtin_variables'
+ '#lang_facts_builtin_variables')
+_server_variables = f'{_lang_facts_builtin_variables}-server-variables'
+_compiler_variables = f'{_lang_facts_builtin_variables}-compiler-variables'
+_trusted_facts = f'{_lang_facts_builtin_variables}-trusted-facts'
+_server_facts = f'{_lang_facts_builtin_variables}-server-facts'
+
+_built_in_variables = {
+ 'facts': 'https://google.com',
+ # clientcert, clientversion, puppetversion, clientnoop,
+ # agent_specified_environment:
+ # https://www.puppet.com/docs/puppet/7/lang_facts_builtin_variables#lang_facts_builtin_variables-agent-facts
+ 'trusted': _trusted_facts,
+ 'server_facts': _server_facts,
+ 'environment': _server_variables,
+ 'servername': _server_variables,
+ 'serverip': _server_variables,
+ 'serverversion': _server_variables,
+ 'module_name': _compiler_variables,
+ 'caller_module_name': _compiler_variables,
+
+ # Also note the special variable $title and $name
+ # https://www.puppet.com/docs/puppet/7/lang_defined_types#lang_defined_types-title-and-name
+}
+
+
+# https://www.puppet.com/docs/puppet/7/cheatsheet_core_types.html
+# https://www.puppet.com/docs/puppet/7/types/file.html
+# ...
+_built_in_types = {
+ 'package',
+ 'file',
+ 'service',
+ 'notify',
+ 'exec',
+ 'user',
+ 'group',
+}
+
+# https://www.puppet.com/docs/puppet/7/function.html#{}
+_built_in_functions = {
+ 'abs',
+ 'alert',
+ 'all',
+ 'annotate',
+ 'any',
+ 'assert_type',
+ 'binary_file',
+ 'break',
+ 'call',
+ 'camelcase',
+ 'capitalize',
+ 'ceiling',
+ 'chomp',
+ 'chop',
+ 'compare',
+ 'contain',
+ 'convert_to',
+ 'create_resources',
+ 'crit',
+ 'debug',
+ 'defined',
+ 'dig',
+ 'digest',
+ 'downcase',
+ 'each',
+ 'emerg',
+ 'empty',
+ 'epp',
+ 'err',
+ 'eyaml_lookup_key',
+ 'fail',
+ 'file',
+ 'filter',
+ 'find_file',
+ 'find_template',
+ 'flatten',
+ 'floor',
+ 'fqdn_rand',
+ 'generate',
+ 'get',
+ 'getvar',
+ 'group_by',
+ 'hiera',
+ 'hiera_array',
+ 'hiera_hash',
+ 'hiera_include',
+ 'hocon_data',
+ 'import',
+ 'include',
+ 'index',
+ 'info',
+ 'inline_epp',
+ 'inline_template',
+ 'join',
+ 'json_data',
+ 'keys',
+ 'length',
+ 'lest',
+ 'lookup',
+ 'lstrip',
+ 'map',
+ 'match',
+ 'max',
+ 'md5',
+ 'min',
+ 'module_directory',
+ 'new',
+ 'next',
+ 'notice',
+ 'partition',
+ 'realize',
+ 'reduce',
+ 'regsubst',
+ 'require',
+ 'return',
+ 'reverse_each',
+ 'round',
+ 'rstrip',
+ 'scanf',
+ 'sha1',
+ 'sha256',
+ 'shellquote',
+ 'size',
+ 'slice',
+ 'sort',
+ 'split',
+ 'sprintf',
+ 'step',
+ 'strftime',
+ 'strip',
+ 'tag',
+ 'tagged',
+ 'template',
+ 'then',
+ 'tree_each',
+ 'type',
+ 'unique',
+ 'unwrap',
+ 'upcase',
+ 'values',
+ 'versioncmp',
+ 'warning',
+ 'with',
+ 'yaml_data',
+}
+
+
+def _find_declarations(objs: list[MatchObject]) -> list[str]:
+ """
+ Find all local variable declarations.
+
+ Searches the code for all local variable declarations, returing a
+ list of variable names.
+
+ Note that the same variable might appear multiple times, for example:
+
+ .. code-block:: puppet
+ :caption: The same variable being declared twice
+
+ if $something {
+ $x = 10
+ } else {
+ $x = 20
+ }
+ """
+ declarations = []
+ for obj in objs:
+ match obj:
+ case MatchCompound(type='declaration', matched=xs):
+ for x in xs:
+ match x:
+ case MatchCompound(type='var', matched=ys):
+ declarations.append(inner_text(ys))
+ return declarations
+
+
+class _PuppetReserializer:
+ """
+ Reserializes parsed puppet code back into puppet code.
+
+ This allows syntax highlighting, and hyperlinking to be added to the code.
+
+ :param local_vars:
+ Variables declared within this file. Used when resolving
+ hyperlinks.
+ """
+
+ def __init__(self, local_vars: list[str]):
+ self.local_vars: list[str] = local_vars
+
+ def reserialize(self, obj: MatchObject | Sequence[MatchObject]) -> str:
+ """
+ Reconstruct puppet code after parsing it.
+
+ After building the parser, and parsing the puppet code into a tree
+ of MatchObjects; this procedure returns it into puppet code.
+ Difference being that we now have metadata, meaning that syntax
+ highlighting and variable hyperlinks can be inserted.
+
+ :param obj:
+ Should be assumed to be a list of MatchObject's, or something similar.
+
+ MatchCompound objects are serialized as
+
+ .. code-block:: html
+
+ <span class="{type}">{body}</span>
+
+ esrings as themselves, and lists have reserialize mapped over them.
+
+ """
+ out: list[str] = []
+ # logger.info("obj = %a", obj)
+
+ # TODO hyperlink functions.
+ # The problem is that a function can either be implemented in
+ # Puppet, or in Ruby. And Ruby functions' names aren't bound
+ # by the directory layout.
+ match obj:
+ case str(s):
+ out.append(html.escape(s))
+
+ case MatchCompound(type='resource-name', matched=xs):
+ name = inner_text(xs)
+ url, cls = name_to_url(name)
+ if url:
+ out.append(f'<a href="{url}" class="resource-name {cls}">{name}</a>')
+ else:
+ # TODO this is class, but the class name should
+ # also be hyperlinked
+ out.append(f'<span class="resource-name {cls}">{name}</span>')
+
+ case MatchCompound(type='invoke', matched=xs):
+ function = None
+ for x in xs:
+ match x:
+ case MatchCompound(type='qn', matched=ys):
+ if function is None:
+ function = inner_text(ys)
+ if function in _built_in_functions:
+ # class="qn"
+ url = f"https://www.puppet.com/docs/puppet/7/function.html#{function}" # noqa: E501
+ tag = f'<a href="{url}" class="puppet-doc">{self.reserialize(ys)}</a>' # noqa: E501
+ out.append(tag)
+ else:
+ # TODO function to url
+ out.append(f'<span class="qn">{self.reserialize(ys)}</span>')
+ else:
+ if function == 'include':
+ url, cls = name_to_url(inner_text(ys))
+ # class="qn"
+ tag = f'<a href="{url}" class="{cls}">{self.reserialize(ys)}</a>' # noqa: E501
+ out.append(tag)
+ else:
+ out.append(self.reserialize(ys))
+ case _:
+ out.append(self.reserialize(x))
+
+ case MatchCompound(type='declaration', matched=xs):
+ for x in xs:
+ match x:
+ case MatchCompound(type='var', matched=ys):
+ inner = ''.join(self.reserialize(y) for y in ys)
+ out.append(f'<span id="{inner_text(ys)}">{inner}</span>')
+ case _:
+ out.append(self.reserialize(x))
+
+ case MatchCompound(type='var', matched=xs):
+ out.append(self.var_to_url(inner_text(xs)))
+
+ case MatchCompound(type=type, matched=xs):
+ body = ''.join(self.reserialize(x) for x in xs)
+ out.append(f'<span class="{type}">{body}</span>')
+
+ case [*xs]:
+ out.extend(self.reserialize(x) for x in xs)
+
+ case rest:
+ logger.error("Unknown type: %a", rest)
+
+ return ''.join(out)
+
+ def var_to_url(self, var: str) -> str:
+ """
+ Format variable, adding hyperlink to its definition.
+
+ TODO these can refer to both defined types (`manifests/*.pp`),
+ as well as resource types (`lib/puppet/provider/*/*.rb` /
+ `lib/tpuppet/type/*.rb`)
+
+ Same goes for functions (`functions/*.pp`),
+ (`lib/puppet/functions.rb`).
+
+ :param var:
+ Name of the variable.
+
+ :return:
+ An HTML anchor element.
+ """
+ match var.split('::'):
+ case [name]:
+ # Either a local or global variable
+ # https://www.puppet.com/docs/puppet/7/lang_facts_and_builtin_vars.html
+
+ href = None
+ cls = ''
+ if name in self.local_vars:
+ href = f'#{html.escape(var)}'
+ elif name in _built_in_variables:
+ href = html.escape(_built_in_variables[name])
+ cls = 'puppet-doc'
+
+ if href:
+ return f'<a class="var {cls}" href="{href}">{var}</a>'
+ else:
+ # `name` refers to a global fact.
+ return f'<span class="var">{var}</span>'
+
+ case ['', name]:
+ # A global variable
+ if name in _built_in_variables:
+ href = html.escape(_built_in_variables[name])
+ img = '<img src="/code/muppet-strings/output/static/favicon.ico" />'
+ return f'<a class="var" href="{href}">{var}{img}</a>'
+ else:
+ return f'<span class="var">{var}</span>'
+
+ # Note the "special module" 'settings',
+ # https://www.puppet.com/docs/puppet/7/lang_facts_builtin_variables#lang_facts_builtin_variables-server-variables
+ case ['', module, *items, name]:
+ s = '/code/muppet-strings/output/' \
+ + '/'.join([module, 'manifests', *(items if items else ['init'])])
+ s += f'#{name}'
+ return f'<a class="var" href="{s}">{var}</a>'
+ case [module, *items, name]:
+ s = '/code/muppet-strings/output/' \
+ + '/'.join([module, 'manifests', *(items if items else ['init'])])
+ s += f'#{name}'
+ return f'<a class="var" href="{s}">{var}</a>'
+ case _:
+ raise ValueError()
+
+
+def hyperlink_puppet_source(source: str, file: str, in_parameters: list[str]) -> str:
+ """
+ Parse and syntax highlight the given puppet source.
+
+ :returns: An HTML string
+ """
+ # Run the upstream puppet parser,
+ # then masage the tree into a usable form.
+ ast = build_ast(puppet_parser(source))
+
+ # From the ast, build a parser combinator parser.
+ # This parser will attach sufficient metadata to allow syntax
+ # highlighting and hyperlinking
+ parser = ParserFormatter().serialize(ast)
+
+ # Run the generated parser, giving us a list of match objects.
+ match_objects = ParserCombinator(source, file).get(parser)
+
+ # Reserialize the matched data back into puppet code, realizing
+ # the syntax highlighting and hyperlinks.
+ return _PuppetReserializer(_find_declarations(match_objects) + (in_parameters)) \
+ .reserialize(match_objects)
+
+
+def name_to_url(name: str) -> tuple[str | None, str]:
+ """
+ Resolve a class or resource name into an url.
+
+ :param name:
+ The name of a class or resource, surch as "example::resource".
+ :return:
+ A tuple consisting of
+
+ - One of
+ - An internal link to the definition of that type
+ - A link to the official puppet documentation
+ - ``None``, if `name` is "class"
+ - A string indicating extra HTML classes for this url.
+ This is mostly so external references can be marked properly.
+ """
+ if name in _built_in_types:
+ return (f'https://www.puppet.com/docs/puppet/7/types/{name}.html', 'puppet-doc')
+ elif name == 'class':
+ return (None, '')
+ else:
+ # TODO special cases for puppet's built in types.
+ # https://www.puppet.com/docs/puppet/7/cheatsheet_core_types.html
+ module, *items = name.lstrip(':').split('::')
+ # TODO get prefix from the command line/config file
+ return ('/code/muppet-strings/output/'
+ + '/'.join([module, 'manifests', *(items if items else ['init'])]),
+ '')
diff --git a/muppet/output/util.py b/muppet/output/util.py
new file mode 100644
index 0000000..b1d69f5
--- /dev/null
+++ b/muppet/output/util.py
@@ -0,0 +1,44 @@
+"""
+Misc utilities for the final output.
+
+These don't really belong to any sub-system, even though some are more
+useful than other.
+
+The aim is to only have pure functions here.
+"""
+
+from muppet.parser_combinator import (
+ MatchCompound,
+ MatchObject,
+)
+
+
+def inner_text(obj: MatchObject | list[MatchObject]) -> str:
+ """
+ Extract the text content from a set of MatchObjects.
+
+ This is really similar to HTML's inner_text.
+
+ Empty whitespace tags are expanded into nothing, non-empty
+ whitespace tags becomes a single space (note that this discards
+ commets).
+
+ This only works properly if no function was mapped over the parser
+ return values in tree, see :func:`muppet.parser_combinator.fmap`.
+
+ :param obj:
+ Match Objects to search.
+ """
+ match obj:
+ case str(s):
+ return s
+ case MatchCompound(type='ws', matched=[]):
+ return ''
+ case MatchCompound(type='ws'):
+ return ' '
+ case MatchCompound(matched=xs):
+ return ''.join(inner_text(x) for x in xs)
+ case [*xs]:
+ return ''.join(inner_text(x) for x in xs)
+ case _:
+ raise ValueError('How did we get here')
--
GitLab