glean-parser 16.0.0__tar.gz → 16.2.0__tar.gz

{glean_parser-16.0.0 → glean_parser-16.2.0}/CHANGELOG.md

@@ -2,6 +2,16 @@
 
 ## Unreleased
 
+## 16.2.0
+
+- New lint: error when there are metrics whose names are too similar ([bug 1934099](https://bugzilla.mozilla.org/show_bug.cgi?id=1934099))
+- Require `array` or `object` as the root type in object metrics ([#780](https://github.com/mozilla/glean_parser/pull/780))
+- Remove 100-bucket limit for `custom_distribution` metrics ([bug 1940967](https://bugzilla.mozilla.org/show_bug.cgi?id=1940967))
+
+## 16.1.0
+
+- Allow specifying a subset of interesting metrics to actually collect. Other metrics will be built, but marked as disabled ([bug 1931277](https://bugzilla.mozilla.org/show_bug.cgi?id=1911165)).
+
 ## 16.0.0
 
 - BREAKING CHANGE: Support `follows_collection_enabled` for pings ([#776](https://github.com/mozilla/glean_parser/pull/776))

{glean_parser-16.0.0 → glean_parser-16.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: glean_parser
-Version: 16.0.0
+Version: 16.2.0
 Summary: Parser tools for Mozilla's Glean telemetry
 Home-page: https://github.com/mozilla/glean_parser
 Author: The Glean Team
@@ -79,6 +79,16 @@ $ glean_parser check < ping.json
 
 ## Unreleased
 
+## 16.2.0
+
+- New lint: error when there are metrics whose names are too similar ([bug 1934099](https://bugzilla.mozilla.org/show_bug.cgi?id=1934099))
+- Require `array` or `object` as the root type in object metrics ([#780](https://github.com/mozilla/glean_parser/pull/780))
+- Remove 100-bucket limit for `custom_distribution` metrics ([bug 1940967](https://bugzilla.mozilla.org/show_bug.cgi?id=1940967))
+
+## 16.1.0
+
+- Allow specifying a subset of interesting metrics to actually collect. Other metrics will be built, but marked as disabled ([bug 1931277](https://bugzilla.mozilla.org/show_bug.cgi?id=1911165)).
+
 ## 16.0.0
 
 - BREAKING CHANGE: Support `follows_collection_enabled` for pings ([#776](https://github.com/mozilla/glean_parser/pull/776))

{glean_parser-16.0.0 → glean_parser-16.2.0}/docs/history.md

@@ -2,6 +2,16 @@
 
 ## Unreleased
 
+## 16.2.0
+
+- New lint: error when there are metrics whose names are too similar ([bug 1934099](https://bugzilla.mozilla.org/show_bug.cgi?id=1934099))
+- Require `array` or `object` as the root type in object metrics ([#780](https://github.com/mozilla/glean_parser/pull/780))
+- Remove 100-bucket limit for `custom_distribution` metrics ([bug 1940967](https://bugzilla.mozilla.org/show_bug.cgi?id=1940967))
+
+## 16.1.0
+
+- Allow specifying a subset of interesting metrics to actually collect. Other metrics will be built, but marked as disabled ([bug 1931277](https://bugzilla.mozilla.org/show_bug.cgi?id=1911165)).
+
 ## 16.0.0
 
 - BREAKING CHANGE: Support `follows_collection_enabled` for pings ([#776](https://github.com/mozilla/glean_parser/pull/776))

{glean_parser-16.0.0 → glean_parser-16.2.0}/glean_parser/lint.py

@@ -385,6 +385,35 @@ def check_unknown_ping(
                 yield nit
 
 
+def check_name_too_similar(
+    check_name: str,
+    check_type: CheckType,
+    all_pings: Dict[str, pings.Ping],
+    all_metrics: Dict[str, metrics.Metric],
+    parser_config: Dict[str, Any],
+) -> NitGenerator:
+    """
+    Check that all metrics identifiers are suitably distinct.
+    Require that at least n-1 of the similarly-named metrics must be no_lint'd to dismiss the lint.
+
+    Current similarity test: the fully-qualified identifier differs solely in punctuation.
+    e.g. formautofill.credit_cards and formautofill.creditcards
+    """
+    seen_metrics: Dict[str, metrics.Metric] = dict()
+
+    for _, metric in all_metrics.items():
+        if check_name in metric.no_lint:
+            continue
+
+        no_punc = metric.identifier().replace("_", "").replace(".", "")
+        if no_punc in seen_metrics:
+            msg = f"Metric `{metric.identifier()}`'s name is too similar to existing metric `{seen_metrics[no_punc].identifier()}`"
+            nit = GlinterNit(check_name, metric.identifier(), msg, check_type)
+            yield nit
+
+        seen_metrics[no_punc] = metric
+
+
 # The checks that operate on an entire category of metrics:
 #    {NAME: (function, is_error)}
 CATEGORY_CHECKS: Dict[
@@ -437,6 +466,7 @@ ALL_OBJECT_CHECKS: Dict[
     ],
 ] = {
     "UNKNOWN_PING_REFERENCED": (check_unknown_ping, CheckType.error),
+    "NAME_TOO_SIMILAR": (check_name_too_similar, CheckType.error),
 }
 
 

{glean_parser-16.0.0 → glean_parser-16.2.0}/glean_parser/metrics.py

@@ -509,6 +509,20 @@ class Object(Metric):
         if None:
             raise ValueError("`structure` needed for object metric.")
 
+        # Different from `ALLOWED_TYPES`:
+        # We _require_ the root type to be an object or array.
+        allowed_types = ["object", "array"]
+        if "type" not in structure:
+            raise ValueError(
+                f"missing `type` in object structure. Allowed: {allowed_types}"
+            )
+        if structure["type"] not in allowed_types:
+            raise ValueError(
+                "invalid `type` in object structure. found: {}, allowed: {}".format(
+                    structure["type"], allowed_types
+                )
+            )
+
         structure = Object._validate_substructure(structure)
         return structure
 

{glean_parser-16.0.0 → glean_parser-16.2.0}/glean_parser/parser.py

@@ -444,6 +444,8 @@ def parse_objects(
           the metric expires.
         - `allow_missing_files`: Do not raise a `FileNotFoundError` if any of
           the input `filepaths` do not exist.
+        - `interesting`: Contains an array of interesting metrics/ping files.
+          Probes not included in these files will be marked as disabled.
     """
     if config is None:
         config = {}
@@ -465,4 +467,40 @@ def parse_objects(
             yield from _instantiate_tags(
                 all_objects, sources, content, filepath, config
             )
+
+    if config.get("interesting"):
+        # We're configured to disable probes not included in the interesting list.
+        filepaths = util.ensure_list(config.get("interesting"))
+        interesting_metrics_dict: Dict[str, Dict[str, Any]] = dict()
+        interesting_metrics_dict.setdefault("metrics", DictWrapper())
+        interesting_metrics_dict.setdefault("pings", DictWrapper())
+        for filepath in filepaths:
+            content, filetype = yield from _load_file(filepath, config)
+
+            if not isinstance(content, dict):
+                raise TypeError(f"Invalid content for {filepath}")
+
+            for category_key, category_val in sorted(content.items()):
+              if category_key.startswith("$"):
+                  continue
+
+              interesting_metrics_dict.setdefault(category_key, DictWrapper())
+
+              if not isinstance(category_val, dict):
+                  raise TypeError(f"Invalid category_val for {category_key}")
+
+              for metric_key, metric_val in sorted(category_val.items()):
+                  interesting_metrics_dict[category_key][metric_key] = metric_val
+
+        for category_key, category_val in all_objects.items():
+            if category_key == "tags":
+                continue
+
+            for metric_key, metric_val in sorted(category_val.items()):
+                category_dict = interesting_metrics_dict.get(category_key, {})
+                if metric_key not in category_dict:
+                    obj = all_objects[category_key][metric_key]
+                    if hasattr(obj, "disabled"):
+                        obj.disabled = True
+
     return _preprocess_objects(all_objects, config)

{glean_parser-16.0.0 → glean_parser-16.2.0}/glean_parser/schemas/metrics.2-0-0.schema.yaml

@@ -442,7 +442,6 @@ definitions:
           Required when `type`_ is `custom_distribution`.
         type: number
         minimum: 1
-        maximum: 100
 
       histogram_type:
         title: Histogram type

{glean_parser-16.0.0 → glean_parser-16.2.0}/glean_parser.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: glean_parser
-Version: 16.0.0
+Version: 16.2.0
 Summary: Parser tools for Mozilla's Glean telemetry
 Home-page: https://github.com/mozilla/glean_parser
 Author: The Glean Team
@@ -79,6 +79,16 @@ $ glean_parser check < ping.json
 
 ## Unreleased
 
+## 16.2.0
+
+- New lint: error when there are metrics whose names are too similar ([bug 1934099](https://bugzilla.mozilla.org/show_bug.cgi?id=1934099))
+- Require `array` or `object` as the root type in object metrics ([#780](https://github.com/mozilla/glean_parser/pull/780))
+- Remove 100-bucket limit for `custom_distribution` metrics ([bug 1940967](https://bugzilla.mozilla.org/show_bug.cgi?id=1940967))
+
+## 16.1.0
+
+- Allow specifying a subset of interesting metrics to actually collect. Other metrics will be built, but marked as disabled ([bug 1931277](https://bugzilla.mozilla.org/show_bug.cgi?id=1911165)).
+
 ## 16.0.0
 
 - BREAKING CHANGE: Support `follows_collection_enabled` for pings ([#776](https://github.com/mozilla/glean_parser/pull/776))

{glean_parser-16.0.0 → glean_parser-16.2.0}/glean_parser.egg-info/SOURCES.txt

@@ -126,6 +126,7 @@ tests/data/invalid.yamlx
 tests/data/jwe.yaml
 tests/data/metric-with-tags.yaml
 tests/data/mixed-expirations.yaml
+tests/data/name_too_similar.yaml
 tests/data/object.yaml
 tests/data/old_event_api.yamlx
 tests/data/ordering.yaml

{glean_parser-16.0.0 → glean_parser-16.2.0}/tests/data/all_metrics.yaml

@@ -72,8 +72,8 @@ all_metrics:
     <<: *defaults
     type: custom_distribution
     range_min: 0
-    range_max: 100
-    bucket_count: 10
+    range_max: 200
+    bucket_count: 201
     histogram_type: linear
 
   uuid:

glean_parser-16.2.0/tests/data/name_too_similar.yaml

@@ -0,0 +1,31 @@
+# Any copyright is dedicated to the Public Domain.
+# https://creativecommons.org/publicdomain/zero/1.0/
+
+---
+$schema: moz://mozilla.org/schemas/glean/metrics/2-0-0
+
+all_metrics:
+  valid_metric: &defaults
+    type: counter
+    lifetime: ping
+    description: for testing
+    bugs:
+      - https://bugzilla.mozilla.org/123
+    data_reviews:
+      - http://example.com/
+    notification_emails:
+      - CHANGE-ME@example.com
+    expires: never
+
+  validmetric:
+    <<: *defaults
+
+  similar_nolint:
+    <<: *defaults
+    no_lint:
+      - NAME_TOO_SIMILAR
+
+  similar_no_lint:
+    <<: *defaults
+    no_lint:
+      - NAME_TOO_SIMILAR

{glean_parser-16.0.0 → glean_parser-16.2.0}/tests/test_lint.py

@@ -508,6 +508,22 @@ def test_unknown_pings_lint():
     assert "does-not-exist" in nits[0].msg
 
 
+def test_name_too_similar_lint():
+    """Ensure the 'glinter' reports metrics whose names are too similar."""
+    # Note: NAME_TOO_SIMILAR is an all-object lint meaning we need pings for it to work.
+    input = [ROOT / "data" / "name_too_similar.yaml", ROOT / "data" / "pings.yaml"]
+    all_objects = parser.parse_objects(input)
+
+    errs = list(all_objects)
+    assert len(errs) == 0
+
+    nits = lint.lint_metrics(all_objects.value, parser_config={})
+    assert len(nits) == 1
+    assert nits[0].check_name == "NAME_TOO_SIMILAR"
+    assert nits[0].name == "all_metrics.validmetric"
+    assert "all_metrics.valid_metric" in nits[0].msg
+
+
 @pytest.mark.parametrize(
     "metric, num_nits",
     [

{glean_parser-16.0.0 → glean_parser-16.2.0}/tests/test_parser.py

@@ -573,7 +573,7 @@ def test_custom_distribution():
                     "type": "custom_distribution",
                     "range_min": 0,
                     "range_max": 60000,
-                    "bucket_count": 100,
+                    "bucket_count": 200,
                     "histogram_type": "exponential",
                 }
             }
@@ -591,7 +591,6 @@ def test_custom_distribution():
             "category": {
                 "metric": {
                     "type": "custom_distribution",
-                    "gecko_datapoint": "FROM_GECKO",
                 }
             }
         }
@@ -603,27 +602,6 @@ def test_custom_distribution():
     assert len(errors) == 1
     assert "`custom_distribution` is missing required parameters" in errors[0]
 
-    # Test maximum bucket_count is enforced
-    contents = [
-        {
-            "category": {
-                "metric": {
-                    "type": "custom_distribution",
-                    "gecko_datapoint": "FROM_GECKO",
-                    "range_max": 60000,
-                    "bucket_count": 101,
-                    "histogram_type": "exponential",
-                }
-            }
-        }
-    ]
-
-    contents = [util.add_required(x) for x in contents]
-    all_metrics = parser.parse_objects(contents)
-    errors = list(all_metrics)
-    assert len(errors) == 1
-    assert "101 is greater than" in errors[0]
-
     # Test that correct usage
     contents = [
         {
@@ -631,7 +609,7 @@ def test_custom_distribution():
                 "metric": {
                     "type": "custom_distribution",
                     "range_max": 60000,
-                    "bucket_count": 100,
+                    "bucket_count": 200,
                     "histogram_type": "exponential",
                 }
             }
@@ -645,7 +623,7 @@ def test_custom_distribution():
     distribution = all_metrics.value["category"]["metric"]
     assert distribution.range_min == 1
     assert distribution.range_max == 60000
-    assert distribution.bucket_count == 100
+    assert distribution.bucket_count == 200
     assert distribution.histogram_type == metrics.HistogramType.exponential
 
 
@@ -727,6 +705,40 @@ def test_do_not_disable_expired():
     assert metrics["category"]["metric"].disabled is False
 
 
+def test_interesting_disables_others():
+    # Test that if we get an "interesting" config param, only the metrics in
+    # the provided config files are enabled.
+
+    # We'll have aall_metrics.yaml available loaded.
+    all_metrics = ROOT / "data" / "all_metrics.yaml"
+    # But, only the metrics in metric-with-tags.yaml are marked as interesting
+    # so every other metric will be marked as disabled.
+    metrics_with_tags_yaml = ROOT / "data" / "metric-with-tags.yaml"
+    interesting = [Path(path) for path in [ metrics_with_tags_yaml ]]
+    config = {"interesting": interesting}
+
+    all_metrics = parser.parse_objects([all_metrics, metrics_with_tags_yaml], config)
+    errors = list(all_metrics)
+    assert len(errors) == 0
+
+    interesting_metrics = parser.parse_objects([metrics_with_tags_yaml])
+    errors = list(interesting_metrics)
+    assert len(errors) == 0
+
+    metrics = all_metrics.value
+    for category_key, category_val in metrics.items():
+        if category_key == "tags":
+          continue
+
+        for metric_key, metric_val in sorted(category_val.items()):
+            cat = interesting_metrics.value.get(category_key)
+            if cat and cat.get(metric_key):
+              disabled = cat.get(metric_key).disabled
+            else:
+              disabled = True
+            assert metrics[category_key][metric_key].disabled is disabled
+
+
 def test_send_in_pings_restrictions():
     """Test that invalid ping names are disallowed in `send_in_pings`."""
     all_metrics = parser.parse_objects(ROOT / "data" / "invalid-ping-names.yaml")
@@ -1239,3 +1251,11 @@ def test_object_invalid():
     errors = list(all_metrics)
     assert len(errors) == 1
     assert "`items` not allowed in object structure" in errors[0]
+
+    structure = {"type": "string"}
+    contents = [{"category": {"metric": {"type": "object", "structure": structure}}}]
+    contents = [util.add_required(x) for x in contents]
+    all_metrics = parser.parse_objects(contents)
+    errors = list(all_metrics)
+    assert len(errors) == 1
+    assert "invalid `type` in object structure." in errors[0]