urls: Decouple regexes, variable consistency (#463)

Author: tony

CHANGES

@@ -15,6 +15,16 @@ $ pip install --user --upgrade --pre libvcs
 
 <!-- Maintainers, insert changes / features for the next release here -->
 
+### Breaking changes
+
+#### urls: Variable changes (#463)
+
+- `RE_PIP_REV` moved from `libvcs.url.git` to `libvcs.url.constants`.
+- Regex pattern for user (e.g. `git@`) decoupled to `RE_USER`.
+- `RE_PATH` and `SCP_REGEX` (now `RE_SCP`) no longer include user regex pattern
+  - Existing patterns now use `RE_USER` explicitly.
+- `REGEX_SCP` renamed to `RE_SCP` for consistency.
+
 ### Documentation
 
 - Automatically linkify links that were previously only text.
@@ -26,6 +36,7 @@ $ pip install --user --upgrade --pre libvcs
 - poetry: 1.8.1 -> 1.8.2
 
   See also: https://github.com/python-poetry/poetry/blob/1.8.2/CHANGELOG.md
+
 - Code quality: Use f-strings in more places (#460)
 
   via [ruff 0.4.2](https://github.com/astral-sh/ruff/blob/v0.4.2/CHANGELOG.md).

docs/url/constants.md

@@ -0,0 +1,14 @@
+---
+myst:
+  html_meta:
+    "property=og:locale": "en_US"
+---
+(url-parser-constants)=
+
+# Constants - `libvcs.url.constants`
+
+```{eval-rst}
+.. automodule:: libvcs.url.constants
+   :members:
+   :undoc-members:
+```

docs/url/index.md

@@ -269,4 +269,5 @@ svn
 hg
 base
 registry
+constants
 ```

src/libvcs/url/constants.py

@@ -0,0 +1,28 @@
+"""Constants shared across ``libvcs.url``."""
+
+RE_USER = r"""
+    ((?P<user>[^/:@]+)@)?
+"""
+"""Optional user, e.g. 'git@'"""
+
+# Credit, pip (license: MIT):
+# https://github.com/pypa/pip/blob/22.1.2/src/pip/_internal/vcs/git.py#L39-L52
+# We modified it to have groupings
+RE_SCP = r"""
+    # Server, e.g. 'github.com'.
+    (?P<hostname>([^/:]+))
+    (?P<separator>:)
+    # The server-side path. e.g. 'user/project.git'. Must start with an
+    # alphanumeric character so as not to be confusable with a Windows paths
+    # like 'C:/foo/bar' or 'C:\foo\bar'.
+    (?P<path>(\w[^:.]+))
+"""
+"""Regular expression for scp-style of git URLs."""
+
+#
+# Third-party URLs, e.g. npm, pip, etc.
+#
+RE_PIP_REV = r"""
+    (@(?P<rev>.*))
+"""
+"""Pip-style revision for branch or revision."""

src/libvcs/url/git.py

@@ -23,24 +23,9 @@
 from libvcs._internal.dataclasses import SkipDefaultFieldsReprMixin
 
 from .base import Rule, RuleMap, URLProtocol
-
-# Credit, pip (license: MIT):
-# https://github.com/pypa/pip/blob/22.1.2/src/pip/_internal/vcs/git.py#L39-L52
-# We modified it to have groupings
-SCP_REGEX = r"""
-    # Optional user, e.g. 'git@'
-    ((?P<user>\w+)@)?
-    # Server, e.g. 'github.com'.
-    (?P<hostname>([^/:]+))
-    (?P<separator>:)
-    # The server-side path. e.g. 'user/project.git'. Must start with an
-    # alphanumeric character so as not to be confusable with a Windows paths
-    # like 'C:/foo/bar' or 'C:\foo\bar'.
-    (?P<path>(\w[^:.]+))
-    """
+from .constants import RE_PIP_REV, RE_SCP, RE_USER
 
 RE_PATH = r"""
-    ((?P<user>\w+)@)?
     (?P<hostname>([^/:]+))
     (:(?P<port>\d{1,5}))?
     (?P<separator>[:,/])?
@@ -69,6 +54,7 @@
             rf"""
         ^{RE_SCHEME}
         ://
+        {RE_USER}
         {RE_PATH}
         {RE_SUFFIX}?
         """,
@@ -83,7 +69,8 @@
         pattern=re.compile(
             rf"""
         ^(?P<scheme>ssh)?
-        {SCP_REGEX}
+        {RE_USER}
+        {RE_SCP}
         {RE_SUFFIX}?
         """,
             re.VERBOSE,
@@ -121,11 +108,6 @@
     )
 """
 
-RE_PIP_REV = r"""
-    (@(?P<rev>.*))
-"""
-
-
 PIP_DEFAULT_RULES: list[Rule] = [
     Rule(
         label="pip-url",
@@ -134,6 +116,7 @@
             rf"""
         {RE_PIP_SCHEME}
         ://
+        {RE_USER}
         {RE_PATH}
         {RE_SUFFIX}?
         {RE_PIP_REV}?
@@ -148,7 +131,8 @@
         pattern=re.compile(
             rf"""
         {RE_PIP_SCP_SCHEME}
-        {SCP_REGEX}?
+        {RE_USER}
+        {RE_SCP}?
         {RE_SUFFIX}?
         {RE_PIP_REV}?
         """,

src/libvcs/url/hg.py

@@ -22,12 +22,12 @@
 from typing import Optional
 
 from libvcs._internal.dataclasses import SkipDefaultFieldsReprMixin
-from libvcs.url.git import RE_PIP_REV, RE_SUFFIX, SCP_REGEX
+from libvcs.url.git import RE_SUFFIX
 
 from .base import Rule, RuleMap, URLProtocol
+from .constants import RE_PIP_REV, RE_SCP, RE_USER
 
 RE_PATH = r"""
-    ((?P<user>\w+)@)?
     (?P<hostname>([^/:]+))
     (:(?P<port>\d{1,5}))?
     (?P<separator>[:,/])?
@@ -53,6 +53,7 @@
             rf"""
         ^{RE_SCHEME}
         ://
+        {RE_USER}
         {RE_PATH}
         {RE_SUFFIX}?
         {RE_PIP_REV}?
@@ -66,7 +67,8 @@
         pattern=re.compile(
             rf"""
         ^(?P<scheme>ssh)?
-        {SCP_REGEX}
+        {RE_USER}
+        {RE_SCP}
         {RE_SUFFIX}?
         """,
             re.VERBOSE,
@@ -100,6 +102,7 @@
             rf"""
         ^{RE_PIP_SCHEME}
         ://
+        {RE_USER}
         {RE_PATH}
         {RE_SUFFIX}?
         {RE_PIP_REV}?

src/libvcs/url/svn.py

@@ -23,12 +23,11 @@
 from typing import Optional
 
 from libvcs._internal.dataclasses import SkipDefaultFieldsReprMixin
-from libvcs.url.git import RE_PIP_REV, SCP_REGEX
 
 from .base import Rule, RuleMap, URLProtocol
+from .constants import RE_PIP_REV, RE_SCP, RE_USER
 
 RE_PATH = r"""
-    ((?P<user>[^/:@]+)@)?
     (?P<hostname>([^/:@]+))
     (:(?P<port>\d{1,5}))?
     (?P<separator>[:,/])?
@@ -56,6 +55,7 @@
             rf"""
         ^{RE_SCHEME}
         ://
+        {RE_USER}
         {RE_PATH}
         {RE_PIP_REV}?
         """,
@@ -68,7 +68,8 @@
         pattern=re.compile(
             rf"""
         ^(?P<scheme>ssh)?
-        {SCP_REGEX}
+        {RE_USER}
+        {RE_SCP}
         {RE_PIP_REV}?
         """,
             re.VERBOSE,
@@ -101,6 +102,7 @@
             rf"""
         ^{RE_PIP_SCHEME}
         ://
+        {RE_USER}
         {RE_PATH}
         {RE_PIP_REV}?
         """,