Switch from pykwalify to StrictYAML

Author: ben-edna

CHANGELOG.rst

@@ -2,6 +2,8 @@ Release 0.12.0 (unreleased)
 ====================================
 
 * Internal refactoring: introduce superproject & subproject (#896)
+* Switch from pykwalify to StrictYAML (#0)
+* Show line number when manifest validation fails (#36)
 
 Release 0.11.0 (released 2026-01-03)
 ====================================

dfetch/manifest/schema.py

@@ -0,0 +1,42 @@
+"""StrictYAML schema for the manifest."""
+
+from strictyaml import Bool, Enum, Float, Int, Map, Optional, Seq, Str
+
+NUMBER = Int() | Float()
+
+REMOTE_SCHEMA = Map(
+    {
+        "name": Str(),
+        "url-base": Str(),
+        Optional("default"): Bool(),
+    }
+)
+
+PROJECT_SCHEMA = Map(
+    {
+        "name": Str(),
+        Optional("dst"): Str(),
+        Optional("branch"): Str(),
+        Optional("tag"): Str(),
+        Optional("revision"): Str(),
+        Optional("url"): Str(),
+        Optional("repo-path"): Str(),
+        Optional("remote"): Str(),
+        Optional("patch"): Str(),
+        Optional("vcs"): Enum(["git", "svn"]),
+        Optional("src"): Str(),
+        Optional("ignore"): Seq(Str()),
+    }
+)
+
+MANIFEST_SCHEMA = Map(
+    {
+        "manifest": Map(
+            {
+                "version": NUMBER,
+                Optional("remotes"): Seq(REMOTE_SCHEMA),
+                "projects": Seq(PROJECT_SCHEMA),
+            }
+        )
+    }
+)

dfetch/manifest/validate.py

@@ -1,27 +1,50 @@
-"""Validate manifests."""
+"""Validate manifests using StrictYAML."""
+
+from __future__ import annotations
 
 import logging
+from collections.abc import Mapping
+from typing import Any, cast
+
+from strictyaml import StrictYAMLError, YAMLValidationError, load
+
+from dfetch.manifest.schema import MANIFEST_SCHEMA
 
-import pykwalify
-from pykwalify.core import Core, SchemaError
-from yaml.scanner import ScannerError
 
-import dfetch.resources
+def _ensure_unique(seq: list[dict[str, Any]], key: str, context: str) -> None:
+    """Ensure values for `key` are unique within a sequence of dicts."""
+    values = [item.get(key) for item in seq if key in item]
+    seen = set()
+    dups = {v for v in values if (v in seen) or seen.add(v)}  # type: ignore
+    if dups:
+        dup_list = ", ".join(sorted(map(str, dups)))
+        raise RuntimeError(f"Duplicate {context}.{key} value(s): {dup_list}")
 
 
 def validate(path: str) -> None:
-    """Validate the given manifest."""
-    logging.getLogger(pykwalify.__name__).setLevel(logging.CRITICAL)
+    """Validate the given manifest file against the StrictYAML schema.
 
-    with dfetch.resources.schema_path() as schema_path:
-        try:
-            validator = Core(source_file=path, schema_files=[str(schema_path)])
-        except ScannerError as err:
-            raise RuntimeError(f"{schema_path} is not a valid YAML file!") from err
+    Raises:
+        RuntimeError: if the file is not valid YAML or violates the schema/uniqueness constraints.
+    """
+    logging.getLogger("strictyaml").setLevel(logging.CRITICAL)
 
     try:
-        validator.validate(raise_exception=True)
-    except SchemaError as err:
-        raise RuntimeError(
-            str(err.msg)  # pyright: ignore[reportAttributeAccessIssue, reportCallIssue]
-        ) from err
+        with open(path, encoding="utf-8") as f:
+            loaded_manifest = load(f.read(), schema=MANIFEST_SCHEMA)
+    except YAMLValidationError as err:
+        # More specific: schema mismatch (missing required keys, wrong types, unknown keys).
+        raise RuntimeError(str(err)) from err
+    except StrictYAMLError as err:
+        # Broader: parsing errors and strictness violations (e.g., forbidden constructs).
+        raise RuntimeError(f"{path} is not a valid YAML file: {err}") from err
+
+    data: dict[str, Any] = cast(dict[str, Any], loaded_manifest.data)
+    manifest: Mapping[str, Any] = data["manifest"]  # required
+    projects: list[dict[str, Any]] = manifest["projects"]  # required
+    remotes: list[dict[str, Any]] = manifest.get("remotes", []) or []  # option
+
+    # Enforce cross-item uniqueness constraints originally in pykwalify:
+    _ensure_unique(remotes, "name", "manifest.remotes")
+    _ensure_unique(projects, "name", "manifest.projects")
+    _ensure_unique(projects, "dst", "manifest.projects")  # only those that have 'dst'

dfetch/resources/__init__.py

@@ -17,9 +17,9 @@ def _resource_path(filename: str) -> ContextManager[Path]:
     )
 
 
-def schema_path() -> ContextManager[Path]:
-    """Get path to schema."""
-    return _resource_path("schema.yaml")
+def template_path() -> ContextManager[Path]:
+    """Get path to template."""
+    return _resource_path("template.yaml")
 
 
 TEMPLATE_PATH = _resource_path("template.yaml")

dfetch/resources/schema.yaml

@@ -90,7 +90,11 @@ Feature: Updated project has dependencies
             Dfetch (0.11.0)
               SomeProject         : Fetched v1
             SomeProject/dfetch.yaml: Schema validation failed:
-             - Value 'very-invalid-manifest' is not a dict. Value path: ''.
+
+                "very-invalid-manifest\n"
+                 ^ (line: 1)
+
+            found arbitrary text
             """
         And 'MyProject' looks like:
             """

features/updated-project-has-dependencies.feature

@@ -41,7 +41,7 @@ classifiers = [
 dependencies = [
     "PyYAML==6.0.3",
     "coloredlogs==15.0.1",
-    "pykwalify==1.8.0",
+    "strictyaml==1.7.3",
     "halo==0.0.31",
     "colorama==0.4.6",
     "typing-extensions==4.15.0",

pyproject.toml

@@ -8,18 +8,18 @@
 import dfetch.resources
 
 
-def test_schema_path() -> None:
-    """Test that schema path can be used as context manager."""
+def test_template_path() -> None:
+    """Test that template path can be used as context manager."""
 
-    with dfetch.resources.schema_path() as schema_path:
-        assert os.path.isfile(schema_path)
+    with dfetch.resources.template_path() as template_path:
+        assert os.path.isfile(template_path)
 
 
-def test_call_schema_path_twice() -> None:
+def test_call_template_path_twice() -> None:
     """Had a lot of problems with calling contextmanager twice."""
 
-    with dfetch.resources.schema_path() as schema_path:
-        assert os.path.isfile(schema_path)
+    with dfetch.resources.template_path() as template_path:
+        assert os.path.isfile(template_path)
 
-    with dfetch.resources.schema_path() as schema_path:
-        assert os.path.isfile(schema_path)
+    with dfetch.resources.template_path() as template_path:
+        assert os.path.isfile(template_path)