chore: use pydantic as schema provider

config/config.schema.json

@@ -0,0 +1,83 @@
+{
+  "additionalProperties": false,
+  "properties": {
+    "countries": {
+      "default": [
+        "benin",
+        "togo"
+      ],
+      "description": "tbd",
+      "items": {
+        "type": "string"
+      },
+      "minItems": 1,
+      "type": "array"
+    },
+    "retrieve": {
+      "additionalProperties": false,
+      "properties": {
+        "source": {
+          "default": "geofabrik",
+          "description": "tbd",
+          "enum": [
+            "geofabrik",
+            "overpass"
+          ],
+          "type": "string"
+        },
+        "primary_name": {
+          "default": "power",
+          "description": "tbd",
+          "type": "string"
+        },
+        "features": {
+          "default": [
+            "substation",
+            "line"
+          ],
+          "description": "tbd",
+          "items": {
+            "type": "string"
+          },
+          "minItems": 1,
+          "type": "array"
+        },
+        "force_redownload": {
+          "default": false,
+          "description": "tbd",
+          "type": "boolean"
+        },
+        "mp": {
+          "default": true,
+          "description": "tbd",
+          "type": "boolean"
+        },
+        "stream_backend": {
+          "default": true,
+          "description": "tbd",
+          "type": "boolean"
+        },
+        "cache_primary": {
+          "default": false,
+          "description": "tbd",
+          "type": "boolean"
+        },
+        "target_date": {
+          "anyOf": [
+            {
+              "type": "string"
+            },
+            {
+              "type": "null"
+            }
+          ],
+          "default": null,
+          "description": "tbd"
+        }
+      },
+      "description": "tbd"
+    }
+  },
+  "title": "ConfigSchema",
+  "type": "object"
+}

config/config.yaml

@@ -1,15 +1,17 @@
+%YAML 1.1
+---
+# yaml-language-server: $schema=./config.schema.json
 countries:
-  - benin
-  - togo
-
+- benin
+- togo
 retrieve:
   source: geofabrik
   primary_name: power
   features:
-    - substation
-    - line
+  - substation
+  - line
   force_redownload: false
   mp: true
   stream_backend: true
   cache_primary: false
-  target_date: null
+  target_date:

pixi.toml

@@ -13,7 +13,8 @@ clio-tools = ">=2026.03.30"
 conda = ">=25.0.0"
 ipdb = ">=0.13.13"
 ipykernel = ">=6.29.5"
-jsonschema = ">=4.0.0"
+pydantic = ">=2.0"
+"ruamel.yaml" = ">=0.18"
 mypy = ">=1.15.0"
 pytest = ">=8.3.5"
 python = ">=3.12"
@@ -25,3 +26,4 @@ earth-osm = ">=3.0.2"
 
 [tasks]
 test-integration = {cmd = "pytest tests/integration_test.py"}
+generate-config = {cmd = "python workflow/scripts/_schema.py"}

tests/integration_test.py

@@ -31,7 +31,7 @@ def test_interface_file(module_path):
         "LICENSE",
         "README.md",
         "config/config.yaml",
-        "workflow/internal/config.schema.yaml",
+        "config/config.schema.json",
         "tests/integration/Snakefile",
     ],
 )

workflow/Snakefile

@@ -1,6 +1,10 @@
+import sys
+
 import yaml
+from snakemake.utils import min_version
 
-from snakemake.utils import min_version, validate
+sys.path.insert(0, workflow.basedir)
+from scripts._schema import validate_config
 
 min_version("9.19")
 
@@ -15,8 +19,8 @@ pathvars:
 configfile: workflow.source_path("../config/config.yaml")
 
 
-# Validate the configuration using the schema file.
-validate(config, workflow.source_path("internal/config.schema.yaml"))
+# Validate the configuration using the Pydantic schema.
+config = validate_config(config)
 
 # Load internal settings separately so users cannot modify them.
 with open(workflow.source_path("internal/settings.yaml"), "r") as f: