Started two new rules

Author: forman

docs/rule-ref.md

@@ -5,12 +5,32 @@ New rules will be added by upcoming XRLint releases.
 
 ## Core Rules
 
+### :material-lightbulb: `conventions`
+
+Datasets should identify the applicable conventions using the `Conventions` attribute.
+ The rule has an optional configuration parameter `match` which is a regex pattern that the value of the `Conventions` attribute must match, if any.
+[:material-information-variant:](https://cfconventions.org/cf-conventions/cf-conventions.html#identification-of-conventions)
+
+Contained in:  `all`-:material-lightning-bolt:
+
 ### :material-bug: `coords-for-dims`
 
 Dimensions of data variables should have corresponding coordinates.
 
 Contained in:  `all`-:material-lightning-bolt: `recommended`-:material-lightning-bolt:
 
+### :material-lightbulb: `dataset-description`
+
+A dataset should provide information about where the data came from and what has been done to it. This information is mainly for the benefit of human readers. The rule accepts the following configuration parameters:
+
+- `global_attrs`: list of global attribute names. Defaults to `['title', 'history']`.
+- `var_attrs`: list of variable attribute names. Defaults to `['institution', 'source', 'references', 'comment']`.
+- `ignored_vars`: list of ignored variables (regex patterns). Defaults to `['crs', 'spatial_ref']`.
+
+[:material-information-variant:](https://cfconventions.org/cf-conventions/cf-conventions.html#description-of-file-contents)
+
+Contained in:  `all`-:material-lightning-bolt:
+
 ### :material-lightbulb: `dataset-title-attr`
 
 Datasets should be given a non-empty title.
@@ -67,6 +87,7 @@ Contained in:  `all`-:material-lightning-bolt: `recommended`-:material-lightning
 ### :material-lightbulb: `var-units-attr`
 
 Every variable should have a valid 'units' attribute.
+[:material-information-variant:](https://cfconventions.org/cf-conventions/cf-conventions.html#units)
 
 Contained in:  `all`-:material-lightning-bolt: `recommended`-:material-alert:
 

xrlint/plugins/core/rules/conventions.py

@@ -0,0 +1,50 @@
+import re
+
+
+from xrlint.node import DatasetNode
+from xrlint.plugins.core.plugin import plugin
+from xrlint.rule import RuleContext, RuleOp, RuleExit
+from xrlint.util.schema import schema
+
+DEFAULT_ATTR_NAMES = ["Conventions", "title", "source", "history", "name"]
+
+
+@plugin.define_rule(
+    "conventions",
+    version="1.0.0",
+    type="suggestion",
+    description=(
+        "Datasets should identify the applicable conventions"
+        " using the `Conventions` attribute.\n"
+        " The rule has an optional configuration parameter `match` which"
+        " is a regex pattern that the value of the `Conventions` attribute"
+        " must match, if any."
+    ),
+    docs_url=(
+        "https://cfconventions.org/cf-conventions/cf-conventions.html"
+        "#identification-of-conventions"
+    ),
+    schema=schema(
+        "object", properties=dict(match=schema("string", title="Regex pattern"))
+    ),
+)
+class Conventions(RuleOp):
+    def __init__(self, match: str | None = None):
+        self.match = re.compile(match) if match else None
+
+    def dataset(self, ctx: RuleContext, node: DatasetNode):
+        if "Conventions" not in node.dataset.attrs:
+            ctx.report("Missing attribute 'Conventions'.")
+        else:
+            conventions_spec = node.dataset.attrs.get("Conventions")
+            if not isinstance(conventions_spec, str):
+                ctx.report(
+                    f"Invalid attribute 'Conventions':"
+                    f" expected string, got value {conventions_spec!r}."
+                )
+            elif self.match is not None and not self.match.match(conventions_spec):
+                ctx.report(
+                    f"Invalid attribute 'Conventions':"
+                    f" {conventions_spec!r} doesn't match pattern {self.match.pattern!r}."
+                )
+        raise RuleExit

xrlint/plugins/core/rules/dataset_description.py

@@ -0,0 +1,86 @@
+import re
+
+from xrlint.node import DataArrayNode, DatasetNode
+from xrlint.plugins.core.plugin import plugin
+from xrlint.rule import RuleContext, RuleOp
+from xrlint.util.schema import schema
+
+DEFAULT_GLOBAL_ATTRS = ["title", "history"]
+DEFAULT_VAR_ATTRS = ["institution", "source", "references", "comment"]
+DEFAULT_IGNORES = ["crs", "spatial_ref"]
+
+
+@plugin.define_rule(
+    "dataset-description",
+    version="1.0.0",
+    type="suggestion",
+    description=(
+        "A dataset should provide information about where the data came"
+        " from and what has been done to it."
+        " This information is mainly for the benefit of human readers."
+        " The rule accepts the following configuration parameters:\n\n"
+        "- `global_attrs`: list of global attribute names."
+        f" Defaults to `{DEFAULT_GLOBAL_ATTRS}`.\n"
+        "- `var_attrs`: list of variable attribute names."
+        f" Defaults to `{DEFAULT_VAR_ATTRS}`.\n"
+        "- `ignored_vars`: list of ignored variables (regex patterns)."
+        f" Defaults to `{DEFAULT_IGNORES}`.\n"
+        ""
+    ),
+    docs_url=(
+        "https://cfconventions.org/cf-conventions/cf-conventions.html"
+        "#description-of-file-contents"
+    ),
+    schema=schema(
+        "object",
+        properties={
+            "global_attrs": schema(
+                "array",
+                items=schema("string"),
+                default=DEFAULT_GLOBAL_ATTRS,
+                title="Global dataset attribute names",
+            ),
+            "var_attrs": schema(
+                "array",
+                items=schema("string"),
+                default=DEFAULT_VAR_ATTRS,
+                title="Data variable attribute names",
+            ),
+            "ignored_vars": schema(
+                "array",
+                items=schema("string"),
+                default=DEFAULT_IGNORES,
+                title="Ignored variables (regex name patterns)",
+            ),
+        },
+    ),
+)
+class DatasetDescription(RuleOp):
+    def __init__(
+        self, global_attrs: list[str] | None = None, var_attrs: list[str] | None = None
+    ):
+        self.global_attrs = global_attrs if global_attrs else DEFAULT_GLOBAL_ATTRS
+        self.var_attrs = var_attrs if var_attrs else DEFAULT_VAR_ATTRS
+        self.ignored_vars = [
+            re.compile(p) for p in (var_attrs if var_attrs else DEFAULT_VAR_ATTRS)
+        ]
+
+    def dataset(self, ctx: RuleContext, node: DatasetNode):
+        dataset = node.dataset
+        for attr_name in self.global_attrs:
+            if attr_name not in dataset.attrs:
+                ctx.report(f"Missing dataset attribute {attr_name!r}.")
+
+    def data_array(self, ctx: RuleContext, node: DataArrayNode):
+        dataset = ctx.dataset
+        if node.name not in dataset.data_vars:
+            return
+
+        for m in self.ignored_vars:
+            if m.match(str(node.name)):
+                return
+
+        var = node.data_array
+        for attr_name in self.var_attrs:
+            if attr_name not in var.attrs and attr_name not in dataset.attrs:
+                ctx.report(f"Missing data variable attribute {attr_name!r}.")

xrlint/plugins/core/rules/var_units_attr.py

@@ -8,6 +8,7 @@
     version="1.0.0",
     type="suggestion",
     description="Every variable should have a valid 'units' attribute.",
+    docs_url="https://cfconventions.org/cf-conventions/cf-conventions.html#units",
 )
 class VarUnitsAttr(RuleOp):
     def data_array(self, ctx: RuleContext, node: DataArrayNode):