Merge pull request #1 from nicholasmhughes/add-yaml-value-regex

Add yaml value regex hook

.pre-commit-hooks.yaml

@@ -16,3 +16,11 @@
   always_run: True
   types:
     - yaml
+- id: valueregex
+  name: valueregex
+  description: Disallows the use of certain YAML values based on a regex.
+  entry: valueregex
+  language: python
+  always_run: False
+  types:
+    - yaml

README.md

@@ -11,10 +11,12 @@ Add this to your `.pre-commit-config.yaml`
 ```yaml
 repos:
 - repo: https://github.com/eitrtechnologies/pre-commit-yamlpolicy
-  rev: v1.1.0  # Use the ref you want to point to
+  rev: v1.2.0  # Use the ref you want to point to
   hooks:
     - id: bannedk8skinds
     - id: disallowunquoted
+    - id: valueregex
+      args: [--jmespath, '*.matchers[].match', --regex, '\([^ ]|[^ ]\)']
 ```
 
 ### Hooks Available
@@ -33,3 +35,14 @@ Deny commits where certain YAML values are found but not quoted.
     Defaults to `on,off,yes,no,y,n`.
   - `--case-sensitive` - Flag to turn off case insensitivity when searching for
     values. Operation defaults to ignore case.
+
+#### `valueregex`
+Deny commits where certain YAML values are found and match a given regex. A
+JMESPath query is used in conjunction with a regular expression to match string
+values in YAML.
+  - `--jmespath` - [JMESPath expression](https://jmespath.org/) which returns
+    the values to run a regex against. *REQUIRED*
+  - `--regex` - Regex which will cause the hook to fail if it matches any of the
+    values returned by the JMESPath query. *REQUIRED*
+  - `--allow-multiple-documents` - Allow YAML files which use the
+    [multi-document syntax](http://www.yaml.org/spec/1.2/spec.html#YAML)

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = yamlpolicy
-version = 1.0.0
+version = 1.2.0
 description = Allows an organization to specify YAML usage policy.
 long_description = file: README.md
 long_description_content_type = text/markdown
@@ -23,13 +23,15 @@ classifiers =
 [options]
 packages = find:
 install_requires =
+    jmespath>=1.0.0
     ruamel.yaml>=0.16.12
 python_requires = >=3.7.1
 
 [options.entry_points]
 console_scripts =
     bannedk8skinds = bannedk8skinds.bannedk8skinds:main
     disallowunquoted = disallowunquoted.disallowunquoted:main
+    valueregex = valueregex.valueregex:main
 
 [options.packages.find]
 exclude =

valueregex/valueregex.py

@@ -0,0 +1,75 @@
+import argparse
+import re
+from typing import List, Optional, Sequence
+
+import jmespath
+import ruamel.yaml
+
+yaml = ruamel.yaml.YAML(typ="safe")
+
+
+def main(argv: Optional[Sequence[str]] = None) -> int:
+    parser: argparse.ArgumentParser = argparse.ArgumentParser(add_help=False)
+    required: argparse.ArgumentGroup = parser.add_argument_group("required arguments")
+    optional: argparse.ArgumentGroup = parser.add_argument_group("optional arguments")
+    optional.add_argument(
+        "-h",
+        "--help",
+        action="help",
+        default=argparse.SUPPRESS,
+        help="show this help message and exit",
+    )
+    optional.add_argument(
+        "-m",
+        "--multi",
+        "--allow-multiple-documents",
+        action="store_true",
+    )
+    required.add_argument(
+        "-r",
+        "--regex",
+        required=True,
+    )
+    required.add_argument(
+        "-j",
+        "--jmespath",
+        required=True,
+    )
+    required.add_argument("filenames", nargs="*", help="Filenames to check.")
+    args: argparse.Namespace = parser.parse_args(argv)
+
+    regex: re.Pattern = re.compile(args.regex)
+    search: str = args.jmespath
+
+    retval: int = 0
+    for filename in args.filenames:
+        try:
+            with open(filename, encoding="UTF-8") as f:
+                if args.multi:
+                    docs = yaml.load_all(f)
+                else:
+                    docs = [yaml.load(f)]
+                for doc in docs:
+                    for val in jmespath.search(search, doc) or []:
+                        match: re.Match = regex.search(val)
+                        if match:
+                            print(
+                                f'{filename}: Restricted value found for JMESPath "{search}" =  {match.group(0).rstrip()}'
+                            )
+                            retval = 1
+        except ruamel.yaml.YAMLError as exc:
+            print(exc)
+            retval = 1
+        except jmespath.exceptions.ParseError as exc:
+            print(exc)
+            retval = 1
+        except TypeError:
+            print(
+                "JMESPath expression returned non-string values. Ensure your expression returns a flat list of strings."
+            )
+            retval = 1
+    return retval
+
+
+if __name__ == "__main__":
+    exit(main())