Merge pull request #1611 from woodruffw-forks/ww/pep-753

webknjaz · web-flow · commit 696e13631f83 · 2024-10-30T00:35:32.000Z
guides, specifications: update for PEP 753
diff --git a/source/guides/writing-pyproject-toml.rst b/source/guides/writing-pyproject-toml.rst
@@ -399,13 +399,21 @@ To prevent a package from being uploaded to PyPI, use the special ``Private ::
 Do Not Upload`` classifier. PyPI will always reject packages with classifiers
 beginning with ``Private ::``.
 
+.. _writing-pyproject-toml-urls:
 
 ``urls``
 --------
 
 A list of URLs associated with your project, displayed on the left
 sidebar of your PyPI project page.
 
+.. note::
+
+    See :ref:`well-known-labels` for a listing
+    of labels that PyPI and other packaging tools are specifically aware of,
+    and `PyPI's project metadata docs <https://docs.pypi.org/project_metadata/#project-urls>`_
+    for PyPI-specific URL processing.
+
 .. code-block:: toml
 
    [project.urls]
@@ -415,11 +423,34 @@ sidebar of your PyPI project page.
    Issues = "https://github.com/me/spam/issues"
    Changelog = "https://github.com/me/spam/blob/master/CHANGELOG.md"
 
-Note that if the key contains spaces, it needs to be quoted, e.g.,
+Note that if the label contains spaces, it needs to be quoted, e.g.,
 ``Website = "https://example.com"`` but
 ``"Official Website" = "https://example.com"``.
 
+Users are advised to use :ref:`well-known-labels` for their project URLs
+where appropriate, since consumers of metadata (like package indices) can
+specialize their presentation.
+
+For example in the following metadata, neither ``MyHomepage`` nor
+``"Download Link"`` is a well-known label, so they will be rendered verbatim:
 
+.. code-block:: toml
+
+   [project.urls]
+   MyHomepage = "https://example.com"
+   "Download Link" = "https://example.com/abc.tar.gz"
+
+
+Whereas in this metadata ``HomePage`` and ``DOWNLOAD`` both have
+well-known equivalents (``homepage`` and ``download``), and can be presented
+with those semantics in mind (the project's home page and its external
+download location, respectively).
+
+.. code-block:: toml
+
+   [project.urls]
+   HomePage = "https://example.com"
+   DOWNLOAD = "https://example.com/abc.tar.gz"
 
 Advanced plugins
 ================
diff --git a/source/specifications/core-metadata.rst b/source/specifications/core-metadata.rst
@@ -341,32 +341,6 @@ Example::
    These tools have been very widely used for many years, so it was
    easier to update the specification to match the de facto standard.
 
-.. _home-page-optional:
-.. _core-metadata-home-page:
-
-Home-page
-=========
-
-.. versionadded:: 1.0
-
-A string containing the URL for the distribution's home page.
-
-Example::
-
-    Home-page: http://www.example.com/~cschultz/bvote/
-
-.. _core-metadata-download-url:
-
-Download-URL
-============
-
-.. versionadded:: 1.1
-
-A string containing the URL from which this version of the distribution
-can be downloaded.  (This means that the URL can't be something like
-".../BeagleVote-latest.tgz", but instead must be ".../BeagleVote-0.45.tgz".)
-
-
 .. _author-optional:
 .. _core-metadata-author:
 
@@ -669,6 +643,10 @@ Example::
 
 The label is free text limited to 32 characters.
 
+Starting with :pep:`753`, project metadata consumers (such as the Python
+Package Index) can use a standard normalization process to discover "well-known"
+labels, which can then be given special presentations when being rendered
+for human consumption. See :ref:`well-known-project-urls`.
 
 .. _metadata_provides_extra:
 .. _core-metadata-provides-extra:
@@ -819,6 +797,40 @@ Examples::
 Deprecated Fields
 =================
 
+.. _home-page-optional:
+.. _core-metadata-home-page:
+
+Home-page
+---------
+
+.. versionadded:: 1.0
+
+.. deprecated:: 1.2
+
+    Per :pep:`753`, use :ref:`core-metadata-project-url` instead.
+
+A string containing the URL for the distribution's home page.
+
+Example::
+
+    Home-page: http://www.example.com/~cschultz/bvote/
+
+.. _core-metadata-download-url:
+
+Download-URL
+------------
+
+.. versionadded:: 1.1
+
+.. deprecated:: 1.2
+
+    Per :pep:`753`, use :ref:`core-metadata-project-url` instead.
+
+A string containing the URL from which this version of the distribution
+can be downloaded.  (This means that the URL can't be something like
+"``.../BeagleVote-latest.tgz``", but instead must be
+"``.../BeagleVote-0.45.tgz``".)
+
 Requires
 --------
 
diff --git a/source/specifications/pyproject-toml.rst b/source/specifications/pyproject-toml.rst
@@ -318,7 +318,8 @@ Trove classifiers which apply to the project.
   :ref:`Project-URL <core-metadata-project-url>`
 
 A table of URLs where the key is the URL label and the value is the
-URL itself.
+URL itself. See :ref:`well-known-project-urls` for normalization rules
+and well-known rules when processing metadata for presentation.
 
 
 Entry points
diff --git a/source/specifications/section-distribution-metadata.rst b/source/specifications/section-distribution-metadata.rst
@@ -12,3 +12,4 @@ Package Distribution Metadata
    pyproject-toml
    inline-script-metadata
    platform-compatibility-tags
+   well-known-project-urls
diff --git a/source/specifications/well-known-project-urls.rst b/source/specifications/well-known-project-urls.rst
@@ -0,0 +1,176 @@
+.. _`well-known-project-urls`:
+
+===================================
+Well-known Project URLs in Metadata
+===================================
+
+.. important::
+
+    This document is primarily of interest to metadata *consumers*,
+    who should use the normalization rules and well-known list below
+    to make their presentation of project URLs consistent across the
+    Python ecosystem.
+
+    Metadata *producers* (such as build tools and individual package
+    maintainers) may continue to use any labels they please, within the
+    overall ``Project-URL`` length restrictions. However, when possible, users are
+    *encouraged* to pick meaningful labels that normalize to well-known
+    labels.
+
+.. note::
+
+    See :ref:`Writing your pyproject.toml - urls <writing-pyproject-toml-urls>`
+    for user-oriented guidance on choosing project URL labels in your package's
+    metadata.
+
+.. note:: This specification was originally defined in :pep:`753`.
+
+:pep:`753` deprecates the :ref:`core-metadata-home-page` and
+:ref:`core-metadata-download-url` metadata fields in favor of
+:ref:`core-metadata-project-url`, and defines a normalization and
+lookup procedure for determining whether a ``Project-URL`` is
+"well-known," i.e. has the semantics assigned to ``Home-page``,
+``Download-URL``, or other common project URLs.
+
+This allows indices (such as the Python Package Index) and other downstream
+metadata consumers to present project URLs in a
+consistent manner.
+
+.. _project-url-label-normalization:
+
+Label normalization
+===================
+
+.. note::
+
+    Label normalization is performed by metadata *consumers*, not metadata
+    producers.
+
+To determine whether a ``Project-URL`` label is "well-known," metadata
+consumers should normalize the label before comparing it to the
+:ref:`list of well-known labels <well-known-labels>`.
+
+The normalization procedure for ``Project-URL`` labels is defined
+by the following Python function:
+
+.. code-block:: python
+
+    import string
+
+    def normalize_label(label: str) -> str:
+        chars_to_remove = string.punctuation + string.whitespace
+        removal_map = str.maketrans("", "", chars_to_remove)
+        return label.translate(removal_map).lower()
+
+In plain language: a label is *normalized* by deleting all ASCII punctuation
+and whitespace, and then converting the result to lowercase.
+
+The following table shows examples of labels before (raw) and after
+normalization:
+
+.. list-table::
+    :header-rows: 1
+
+    * - Raw
+      - Normalized
+    * - ``Homepage``
+      - ``homepage``
+    * - ``Home-page``
+      - ``homepage``
+    * - ``Home page``
+      - ``homepage``
+    * - ``Change_Log``
+      - ``changelog``
+    * - ``What's New?``
+      - ``whatsnew``
+    * - ``github``
+      - ``github``
+
+.. _well-known-labels:
+
+Well-known labels
+=================
+
+.. note::
+
+    The list of well-known labels is a living standard, maintained as part of
+    this document.
+
+The following table lists labels that are well-known for the purpose of
+specializing the presentation of ``Project-URL`` metadata:
+
+.. list-table::
+   :header-rows: 1
+
+   * - Label (Human-readable equivalent)
+     - Description
+     - Aliases
+   * - ``homepage`` (Homepage)
+     - The project's home page
+     - *(none)*
+   * - ``source`` (Source Code)
+     - The project's hosted source code or repository
+     - ``repository``, ``sourcecode``, ``github``
+   * - ``download`` (Download)
+     - A download URL for the current distribution, equivalent to ``Download-URL``
+     - *(none)*
+   * - ``changelog`` (Changelog)
+     - The project's comprehensive changelog
+     - ``changes``, ``whatsnew``, ``history``
+   * - ``releasenotes`` (Release Notes)
+     - The project's curated release notes
+     - *(none)*
+   * - ``documentation`` (Documentation)
+     - The project's online documentation
+     - ``docs``
+   * - ``issues`` (Issue Tracker)
+     - The project's bug tracker
+     - ``bugs``, ``issue``, ``tracker``, ``issuetracker``, ``bugtracker``
+   * - ``funding`` (Funding)
+     - Funding Information
+     - ``sponsor``, ``donate``, ``donation``
+
+Package metadata consumers may choose to render aliased labels the same as
+their "parent" well known label, or further specialize them.
+
+Example behavior
+================
+
+The following shows the flow of project URL metadata from
+``pyproject.toml`` to core metadata to a potential index presentation:
+
+.. code-block:: toml
+    :caption: Example project URLs in standard configuration
+
+    [project.urls]
+    "Home Page" = "https://example.com"
+    DOCUMENTATION = "https://readthedocs.org"
+    Repository = "https://upstream.example.com/me/spam.git"
+    GitHub = "https://github.com/example/spam"
+
+.. code-block:: email
+    :caption: Core metadata representation
+
+    Project-URL: Home page, https://example.com
+    Project-URL: DOCUMENTATION, https://readthedocs.org
+    Project-URL: Repository, https://upstream.example.com/me/spam.git
+    Project-URL: GitHub, https://github.com/example/spam
+
+.. code-block:: text
+    :caption: Potential rendering
+
+    Homepage: https://example.com
+    Documentation: https://readthedocs.org
+    Source Code: https://upstream.example.com/me/spam.git
+    Source Code (GitHub): https://github.com/example/spam
+
+Observe that the core metadata appears in the form provided by the user
+(since metadata *producers* do not perform normalization), but the
+metadata *consumer* normalizes and identifies appropriate
+human-readable equivalents based on the normalized form:
+
+* ``Home page`` becomes ``homepage``, which is rendered as ``Homepage``
+* ``DOCUMENTATION`` becomes ``documentation``, which is rendered as ``Documentation``
+* ``Repository`` becomes ``repository``, which is rendered as ``Source Code``
+* ``GitHub`` becomes ``github``, which is rendered as ``Source Code (GitHub)``
+  (as a specialization of ``Source Code``)
