From e95696ff530d1ef5cc103a4c6f148f983f6a998d Mon Sep 17 00:00:00 2001
From: Anne Redulla <aredulla@google.com>
Date: Wed, 16 Aug 2023 06:48:43 +0000
Subject: [PATCH] [ssci] Defined URL metadata field

Bug: b:277147404
Change-Id: I4e1fb86afb991fd3b6bf59a2a96d620fc7eea469
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/4776732
Reviewed-by: Rachael Newitt <renewitt@google.com>
Commit-Queue: Anne Redulla <aredulla@google.com>
---
 metadata/fields/custom/url.py | 59 +++++++++++++++++++++++++++++++++++
 metadata/fields/known.py      |  3 ++
 metadata/tests/fields_test.py | 19 +++++++++++
 3 files changed, 81 insertions(+)
 create mode 100644 metadata/fields/custom/url.py

diff --git a/metadata/fields/custom/url.py b/metadata/fields/custom/url.py
new file mode 100644
index 0000000000..90d7ea307e
--- /dev/null
+++ b/metadata/fields/custom/url.py
@@ -0,0 +1,59 @@
+#!/usr/bin/env python3
+# Copyright 2023 The Chromium Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+import os
+import re
+import sys
+from typing import Union
+
+_THIS_DIR = os.path.abspath(os.path.dirname(__file__))
+# The repo's root directory.
+_ROOT_DIR = os.path.abspath(os.path.join(_THIS_DIR, "..", "..", ".."))
+
+# Add the repo's root directory for clearer imports.
+sys.path.insert(0, _ROOT_DIR)
+
+import metadata.fields.types as field_types
+import metadata.fields.util as util
+import metadata.validation_result as vr
+
+# The delimiter used to separate multiple URLs.
+_VALUE_DELIMITER = ","
+
+_PATTERN_URL_ALLOWED = re.compile(r"^(https?|ftp|git):\/\/\S+$")
+_PATTERN_URL_CANONICAL_REPO = re.compile(
+    r"^This is the canonical (public )?repo(sitory)?\.?$", re.IGNORECASE)
+
+
+class URLField(field_types.MetadataField):
+  """Custom field for the package URL(s)."""
+  def __init__(self):
+    super().__init__(name="URL", one_liner=False)
+
+  def validate(self, value: str) -> Union[vr.ValidationResult, None]:
+    """Checks the given value has acceptable URL values only.
+
+    Note: this field supports multiple values.
+    """
+    if util.matches(_PATTERN_URL_CANONICAL_REPO, value):
+      return None
+
+    invalid_values = []
+    for url in value.split(_VALUE_DELIMITER):
+      url = url.strip()
+      if not util.matches(_PATTERN_URL_ALLOWED, url):
+        invalid_values.append(url)
+
+    if invalid_values:
+      template = ("{field_name} has invalid values. URLs must use a protocol "
+                  "scheme in [http, https, ftp, git]. If there are multiple "
+                  "URLs, separate them with a '{delim}'. Invalid values: "
+                  "{values}.")
+      message = template.format(field_name=self._name,
+                                delim=_VALUE_DELIMITER,
+                                values=util.quoted(invalid_values))
+      return vr.ValidationError(message)
+
+    return None
diff --git a/metadata/fields/known.py b/metadata/fields/known.py
index 2ab88036f2..edb9460876 100644
--- a/metadata/fields/known.py
+++ b/metadata/fields/known.py
@@ -16,6 +16,7 @@ sys.path.insert(0, _ROOT_DIR)
 
 import metadata.fields.custom.cpe_prefix
 import metadata.fields.custom.date
+import metadata.fields.custom.url
 import metadata.fields.custom.version
 import metadata.fields.types as field_types
 
@@ -36,11 +37,13 @@ LICENSE_ANDROID_COMPATIBLE = field_types.YesNoField(
 # Custom fields.
 CPE_PREFIX = metadata.fields.custom.cpe_prefix.CPEPrefixField()
 DATE = metadata.fields.custom.date.DateField()
+URL = metadata.fields.custom.url.URLField()
 VERSION = metadata.fields.custom.version.VersionField()
 
 ALL_FIELDS = (
     NAME,
     SHORT_NAME,
+    URL,
     VERSION,
     DATE,
     REVISION,
diff --git a/metadata/tests/fields_test.py b/metadata/tests/fields_test.py
index 343dc1141f..409ce53fb6 100644
--- a/metadata/tests/fields_test.py
+++ b/metadata/tests/fields_test.py
@@ -82,6 +82,25 @@ class FieldValidationTest(unittest.TestCase):
         error_values=["", "\n", "April 3, 2012", "2012/03/04"],
     )
 
+  def test_url_validation(self):
+    self._run_field_validation(
+        field=known_fields.URL,
+        valid_values=[
+            "https://www.example.com/a",
+            "http://www.example.com/b",
+            "ftp://www.example.com/c,git://www.example.com/d",
+            "This is the canonical public repository",
+        ],
+        error_values=[
+            "",
+            "\n",
+            "ghttps://www.example.com/e",
+            "https://www.example.com/ f",
+            "Https://www.example.com/g",
+            "This is an unrecognized message for the URL",
+        ],
+    )
+
   def test_version_validation(self):
     self._run_field_validation(
         field=known_fields.VERSION,