Add a new `CategorySpecificInfo` wrapper by llucax · Pull Request #257 · frequenz-floss/frequenz-client-common-python

llucax · 2026-07-21T12:41:39Z

An electrical component may carry category-specific info on the wire. Known fields are decoded into typed attributes on the concrete component, but anything left over — because the category is unrecognized, or because a newer API version added fields this client doesn't know yet, was either dropped or kept as a bare dict with no indication of which variant it came from. This type gives that leftover data a documented home together with the variant name needed to interpret it.

Some of this info was saved in the category_specific_info field for some component classes, but it was not saved consistently for all, and the kind of the variant was not preserved, making it hard to inspect unrecognized components or to debug mismatched categories.

This PR introduces a small frozen dataclass pairing the protobuf category_specific_info variant name (kind) with the leftover fields that this client version did not translate into typed attributes and stores the carried category specific info on every component by making the atttribute category_specific_info of type CategorySpecificInfo | None.

The protobuf category_specific_info oneof is now decoded once into its kind plus a dict of its fields, then each concrete component drops the fields it translated into typed attributes (e.g. a battery's type), keeping the kind and any leftover. The result:

recognized components keep the variant kind with an empty leftover
UnrecognizedElectricalComponent now preserves the carried info instead of dropping it — previously the escape hatch was empty for the very class that needs it most
MismatchedCategoryElectricalComponent keeps the full info (nothing was translated), recording exactly which variant disagreed with the category
components carrying no variant get None

It also stores the declared category name on mismatched components so that MismatchedCategoryElectricalComponent.__str__ can show the declared category name instead of just the raw int.

To do so MismatchedCategoryElectricalComponent now has a new category_name attribute alongside the raw category, resolved by the converter from the protobuf enum descriptor with the long ELECTRICAL_COMPONENT_CATEGORY_ prefix stripped, so __str__ renders the short, readable name: CID42:comp1:mismatched:category=BATTERY:kind=inverter

Only the mismatched component needs it: its declared category is a recognized value that its class name hides. Unrecognized components have no name to show (that is what makes them unrecognized), and every other component encodes its category in its class name.

Copilot

Pull request overview

This PR expands the client’s forward-compatibility surface by introducing a typed home for “leftover” protobuf fields (via CategorySpecificInfo) and by making numeric and bounds handling more robust/explicit across the library. In addition to the electrical-component changes described in the PR metadata, it also introduces a new FloatInt alias and a new bounds/bounds-set hierarchy with deprecations in the metrics proto converters and MetricSample.

Changes:

Add CategorySpecificInfo and plumb category_specific_info: CategorySpecificInfo | None through all electrical components (including mismatched/unrecognized cases) during proto conversion.
Introduce FloatInt = float | int and update several “float” fields/accessors to reflect the runtime numeric tower.
Add InvalidBounds / BoundsSet / InvalidBoundsSet + new accessors (MetricSample.get_bounds_set(), ElectricalComponent.get_metric_config_bounds()), and deprecate older bounds conversion/fields.

Reviewed changes

Copilot reviewed 51 out of 51 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
tests/types/_location/test_location.py	Extends `Location` tests to cover `int` latitude/longitude via `FloatInt`.
tests/types/_location/test_invalid_longitude.py	Adds coverage that `InvalidLongitude` preserves `int` values.
tests/types/_location/test_invalid_latitude.py	Adds coverage that `InvalidLatitude` preserves `int` values.
tests/test_float.py	New tests for the `FloatInt` type alias behavior at runtime.
tests/microgrid/proto/v1alpha8/test_microgrid.py	Removes assertions about warning logging during microgrid conversion.
tests/microgrid/electrical_components/test_problematic.py	Adds string formatting tests for unrecognized/mismatched components and new `CategorySpecificInfo` usage.
tests/microgrid/electrical_components/test_power_transformer.py	Updates transformer voltage tests to accept `FloatInt`.
tests/microgrid/electrical_components/test_inverter.py	Adds `__str__` coverage for `UnrecognizedInverter`.
tests/microgrid/electrical_components/test_ev_charger.py	Adds `__str__` coverage for `UnrecognizedEvCharger`.
tests/microgrid/electrical_components/test_electrical_component_base.py	Updates base component tests for `category_specific_info` and adds bounds accessor tests.
tests/microgrid/electrical_components/test_category_specific_info.py	New unit tests for `CategorySpecificInfo` construction/equality/hashing.
tests/microgrid/electrical_components/test_battery.py	Adds `__str__` coverage for `UnrecognizedBattery`.
tests/microgrid/electrical_components/proto/v1alpha8/test_raw_storage.py	Ensures proto conversion preserves category-specific info kind/leftovers.
tests/microgrid/electrical_components/proto/v1alpha8/test_electrical_component_simple.py	Verifies category mismatch preserves `CategorySpecificInfo` and adds `category_name`.
tests/microgrid/electrical_components/proto/v1alpha8/test_electrical_component_base.py	Updates base-data expectations and adds metric-config-bounds behavior tests.
tests/microgrid/electrical_components/proto/v1alpha8/conftest.py	Adjusts fixtures for `category_specific_info=None` and removes old metadata assertion.
tests/metrics/test_sample_metric_sample.py	Migrates sample tests from `bounds` to `bounds_set`, adds deprecation-path tests and new bounds-set error handling.
tests/metrics/test_sample_metric_connection.py	Updates `MetricConnection.name` semantics to empty-string default.
tests/metrics/test_sample_aggregated_value.py	Updates aggregated metric value tests to accept `FloatInt`.
tests/metrics/test_bounds.py	Removes old bounds tests (replaced by new `tests/metrics/_bounds/*`).
tests/metrics/proto/v1alpha8/test_sample_metric_sample.py	Updates proto sample conversion tests to use `bounds_set` and invalid-bounds preservation.
tests/metrics/proto/v1alpha8/test_sample_metric_connection.py	Updates proto connection test for empty-string name semantics.
tests/metrics/proto/v1alpha8/test_bounds.py	Adds coverage for new `bounds_from_proto2` and deprecation warnings on old converters.
tests/metrics/_bounds/test_invalid_bounds.py	New tests for `InvalidBounds`.
tests/metrics/_bounds/test_invalid_bounds_set.py	New tests for `InvalidBoundsSet`.
tests/metrics/_bounds/test_invalid_bounds_set_error.py	New tests for `InvalidBoundsSetError`.
tests/metrics/_bounds/test_invalid_bounds_error.py	New tests for `InvalidBoundsError`.
tests/metrics/_bounds/test_bounds.py	New/updated tests for `Bounds` behavior (contains/bool/str, `FloatInt` handling).
tests/metrics/_bounds/test_bounds_set.py	New tests for `BoundsSet` normalization/containment/str/hash behavior.
tests/metrics/_bounds/test_base_bounds.py	New tests ensuring `BaseBounds` is abstract/non-instantiable.
tests/metrics/_bounds/init.py	New package marker for bounds tests.
src/frequenz/client/common/types/_location.py	Updates location types/accessors/errors to use `FloatInt` and handle `int` in pattern matching.
src/frequenz/client/common/microgrid/proto/v1alpha8/_microgrid.py	Removes warning aggregation/logging from `microgrid_from_proto`.
src/frequenz/client/common/microgrid/electrical_components/proto/v1alpha8/_electrical_component.py	Adds category-name lookup, introduces `CategorySpecificInfo` parsing/leftover handling, preserves invalid bounds, and switches to `bounds_from_proto2`.
src/frequenz/client/common/microgrid/electrical_components/_problematic.py	Adds/updates `__str__` for problematic components and adds `category_name` for mismatched cases.
src/frequenz/client/common/microgrid/electrical_components/_power_transformer.py	Updates transformer voltage fields to `FloatInt`.
src/frequenz/client/common/microgrid/electrical_components/_inverter.py	Adds `__str__` for `UnrecognizedInverter`.
src/frequenz/client/common/microgrid/electrical_components/_ev_charger.py	Adds `__str__` for `UnrecognizedEvCharger`.
src/frequenz/client/common/microgrid/electrical_components/_electrical_component.py	Replaces `category_specific_metadata` with `category_specific_info`, preserves invalid bounds types, adds `get_metric_config_bounds()`, and changes `__str__` formatting.
src/frequenz/client/common/microgrid/electrical_components/_category_specific_info.py	New `CategorySpecificInfo` dataclass with custom hashing.
src/frequenz/client/common/microgrid/electrical_components/_battery.py	Adds `__str__` for `UnrecognizedBattery`.
src/frequenz/client/common/microgrid/electrical_components/init.py	Exports `CategorySpecificInfo` (and currently also exports `DefaultT`).
src/frequenz/client/common/metrics/proto/v1alpha8/_sample.py	Switches proto sample conversion to produce `bounds_set` and preserves invalid bounds via type.
src/frequenz/client/common/metrics/proto/v1alpha8/_bounds.py	Adds `bounds_from_proto2`, and deprecates older bounds converters.
src/frequenz/client/common/metrics/proto/v1alpha8/init.py	Re-exports `bounds_from_proto2`.
src/frequenz/client/common/metrics/_sample.py	Introduces `bounds_set`, deprecates `bounds`, adds `get_bounds_set()`, and updates numeric types to `FloatInt`.
src/frequenz/client/common/metrics/_bounds.py	Adds `BaseBounds`, `InvalidBounds`, `BoundsSet`, `InvalidBoundsSet`, and related errors + membership/normalization logic.
src/frequenz/client/common/metrics/init.py	Exports the new bounds/bounds-set types and errors.
src/frequenz/client/common/_float.py	New `FloatInt` type alias module with rationale and examples.
src/frequenz/client/common/init.py	Re-exports `FloatInt`.
RELEASE_NOTES.md	Adds release notes for bounds deprecations, bounds-set APIs, and `FloatInt` (but currently omits the new category-specific info API).

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot

Pull request overview

Copilot reviewed 51 out of 51 changed files in this pull request and generated 1 comment.

Comments suppressed due to low confidence (3)

src/frequenz/client/common/microgrid/electrical_components/_category_specific_info.py:42

CategorySpecificInfo.__hash__() hashes field values via repr(value). For nested mappings/sets, repr() can depend on insertion order / hash randomization, so two instances that compare equal (e.g. equal nested dicts with different insertion order) can end up with different hashes, violating the hash/equality contract.

        frozen_fields = tuple(
            sorted((key, repr(value)) for key, value in self.fields.items())
        )
        return hash((self.kind, frozen_fields))

src/frequenz/client/common/microgrid/electrical_components/init.py:23

DefaultT is a typing-only TypeVar used internally for overloads; re-exporting it from the package surface (__init__.py) needlessly expands the public API and makes it harder to change internal typing details later.

from ._crypto_miner import CryptoMiner
from ._diagnostic_code import ElectricalComponentDiagnosticCode
from ._electrical_component import DefaultT, ElectricalComponent
from ._electrical_component_connection import (

src/frequenz/client/common/microgrid/electrical_components/init.py:92

DefaultT (a private TypeVar) is included in __all__, effectively making it part of the public API. It should be kept internal to avoid locking in an implementation detail of type annotations.

    "CryptoMiner",
    "DcEvCharger",
    "DefaultT",
    "ElectricalComponent",
    "ElectricalComponentCategory",

llucax · 2026-07-21T12:56:05Z

This is a draft because it is based on #256, but it should be ready for review.

Give `Bounds` membership-testing capability, so callers can write `value in bounds` to check whether a value falls within the range. The semantics are: * both bounds are inclusive (`lower <= value <= upper`), * a `None` bound means unbounded in that direction, and * `None` is a bound marker only, never a value, so `None in bounds` is always `False`. The method lives on `Bounds` alone, not on `BaseBounds` or `InvalidBounds`: malformed bounds must not be range-checked — preserving them as `InvalidBounds` is precisely so callers inspect the raw values instead of testing membership against a broken range. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Let `Bounds` answer "do these bounds restrict anything?" so that a `Bounds | None` field can be tested uniformly: both `None` and a fully unbounded `Bounds()` become falsy, so `if not bounds:` reads as "unbounded". Any set `lower` or `upper` makes the bounds truthy. `is_bounded()` is the explicit spelling of that truthiness, following the method style used elsewhere (e.g. `Microgrid.is_active()`), for callers who prefer a named predicate over relying on `bool()`. Emptiness is decided by `is not None`, not by the value's own truthiness, so `Bounds(lower=0.0, upper=0.0)` is correctly bounded rather than being mistaken for unbounded because `0.0` is falsy. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Introduce a normalized union of `Bounds` for efficient membership testing. On construction the bounds are sorted by lower value and overlapping or touching bounds are merged (inclusive, so `[1, 5]` and `[5, 10]` become `[1, 10]`), so the stored `bounds` are canonical: sorted and pairwise non-overlapping. Membership checks perform a binary search over the normalized bounds. `BoundsSet` is a domain-specialized set, not a mathematical one: the empty set is the *unbounded* set. It contains every value and is falsy, so `not bounds_set` reliably means "unbounded". To keep that invariant honest even when unboundedness arrives split across bounds (e.g. the wire sends `[None, 5]` and `[3, None]`), a union that covers the whole space collapses to the empty set, giving unboundedness a single canonical representation. `__contains__` special-cases the empty set to contain everything, and `__bool__` / `is_bounded()` report emptiness. The type deliberately omits `__iter__` and `__len__`: the normalized sequence is public as `BoundsSet.bounds`, and querying membership by iterating it would disagree with `value in bounds_set` for the unbounded set, so `in` is kept the single authoritative operation. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Add the set-level counterpart to `InvalidBounds`, mirroring the `Bounds` / `InvalidBounds` split one level up. A collection of bounds that contains any `InvalidBounds` cannot be normalized into a well-formed `BoundsSet` — malformed ranges have no meaningful sort or merge order — so it is preserved as an `InvalidBoundsSet` instead of silently dropping the bad entries. The type keeps all of the raw bounds (valid and invalid alike) in their original wire order, with no normalization, so callers can inspect exactly what was received. It deliberately provides no membership test: range-checking malformed data is exactly the mistake this type exists to prevent. It is truthy by default, since malformed data is not the same as "unbounded" — only an empty `BoundsSet` means unbounded. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Replace `MetricSample.bounds: list[Bounds]` with `bounds_set: BoundsSet | InvalidBoundsSet`. The metric bounds are a union of ranges, so a normalized `BoundsSet` models them better than a raw list, and — mirroring `bounds_from_proto2` returning `Bounds | InvalidBounds` — malformed wire data is now preserved as an `InvalidBoundsSet` instead of being silently dropped. `bounds` is kept as deprecated for backwards-compatibility. A hand-written `__init__` still accepts the deprecated `bounds=` keyword (emitting a `DeprecationWarning` and building a `BoundsSet` from it), and a deprecated read-only `bounds` property returns the valid `Bounds` from `bounds_set`, so both existing readers and constructors keep working (with warnings). `init=False` plus a manual `__init__` is required because an `InitVar` named `bounds` would collide with the `bounds` property. We could have named the new field `bounds2`, but we decided for `bounds_set` instead: it is a permanent, self-documenting name, so the next minor just drops the deprecated `bounds` property and parameter and leaves `bounds_set` in place — no rename needed. Since the new field is not a 1-1 translation of the protocol message (it goes through normalization) it makes sense to have a different name than the protobuf field. The compatibility property returns only the valid bounds (dropping malformed ones, as the old field effectively did) but the merged, normalized bounds rather than the raw wire list; that small divergence is acceptable for a deprecated shim. On the proto side `_metric_bounds_from_proto` becomes `_bounds_set_from_proto`, returning `BoundsSet | InvalidBoundsSet`. It no longer needs the `metric` argument or the `major_issues` / `minor_issues` side channels: validity is encoded in the returned type (as done for `metric_config_bounds`), so the "bounds ... is invalid, ignoring these bounds" major issue is gone and the invalid bounds are preserved instead of dropped. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Add the set-level counterpart to `InvalidBoundsError`, for a semantic accessor that resolves a `BoundsSet | InvalidBoundsSet` field to a valid `BoundsSet` and instead sees an `InvalidBoundsSet`. It carries the offending set on its `bounds_set` attribute so callers can inspect the raw wire data, and is an `InvalidAttributeError` (hence also a `ValueError`) like the rest of the accessor errors. This is the error `MetricSample.get_bounds_set()` will raise; it is added on its own first. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

`MetricSample.bounds_set` is the lower-level, forward-compatible field: callers reading it must narrow the `BoundsSet | InvalidBoundsSet` union themselves on every access, which is easy to get wrong and easy to skip. Add a higher-level accessor that does the narrowing once: `get_bounds_set()` returns the `BoundsSet` unchanged for well-formed data and turns an `InvalidBoundsSet` into an `InvalidBoundsSetError`, whose `bounds_set` attribute preserves the raw set for inspection. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

`float("nan")` compares `False` against everything, so `nan < lower` and `nan > upper` are both `False` and the membership checks fell through to `True`: `nan in Bounds(1, 5)` and `nan in a_bounds_set` both reported the value as contained. For metric bounds that is a real hazard — a `NaN` sample would be silently treated as within range. Guard both `Bounds.__contains__` and `BoundsSet.__contains__` with `math.isnan()` so `NaN` is never contained, matching how `None` is already rejected. `math` is already imported for the bisect key. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

The property's docstring promises normalized, merged bounds, but it only delivered that for a `BoundsSet` (whose `bounds` are already normalized). For an `InvalidBoundsSet` — whose `bounds` are the raw, unmerged wire data — it returned the valid entries unmerged, so the same property behaved inconsistently depending on the arm of the union. Run the valid entries through `BoundsSet` in both cases, so the deprecated property always returns the merged, normalized bounds it documents, regardless of whether the underlying set is valid. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

PEP 484's numeric tower makes `int` assignable to any `float`-annotated parameter or field, even under `mypy --strict`, while at runtime `isinstance(1, float)` is `False`. Any field annotated `float` can therefore silently store an `int`, and code unwrapping it with `match … case float():` falls through to `assert_never()`, calls to `float`-only methods like `hex()` crash, and `type()`-based dispatch misbehaves — despite everything type-checking cleanly. There is no clean fix in current Python (see the discussion in frequenz-floss#250): runtime coercion at ingress was prototyped and measured ~2.3× slower on hot-path types, structural `Protocol` tricks don't close the widened variable and `Sequence` covariance holes, and narrowing match arms one by one leaves the annotation lying. So the decision is to stop lying instead: annotate every such value as `float | int`, which is what PEP 484 actually admits, at zero runtime cost. It is also forward-compatible with the typing-council proposal to make `float` mean `float | int` (python/typing-council#46). Add the `FloatInt` alias as the canonical spelling of that union. The alias gives the workaround a single documented home — explaining the trap, the `bool ⊂ int` leak. Follow-up commits migrate the existing `float` annotations to it. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

The protobuf message field is required, so the string will always come. Adding `None` as a possibility in the type system forces user code to deal with `None`, and adds an ambiguity between `None` and `""`. So, as we did with other `name` or similar string fields, it is better to just leave it as a pure `str` and let users deal with the empty string more easily, without the need to special-case `None`. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Introduce a small frozen dataclass pairing the protobuf `category_specific_info` variant name (`kind`) with the leftover fields that this client version did not translate into typed attributes. An electrical component may carry category-specific info on the wire. Known fields are decoded into typed attributes on the concrete component, but anything left over — because the category is unrecognized, or because a newer API version added fields this client doesn't know yet — was either dropped or kept as a bare `dict` with no indication of which variant it came from. This type gives that leftover data a documented home together with the variant name needed to interpret it. It is hashable like the components that will carry it. The leftover `fields` mapping may hold arbitrary, possibly unhashable values, and folding them into the hash — even through `repr()` — is unsafe: values that compare equal can hash or repr differently (e.g. `1 == 1.0 == True`), which would break the invariant that equal objects hash equally. So `fields` is excluded from the hash and instances hash on `kind` alone, mirroring `metric_config_bounds` on the component. Collisions between instances that share a `kind` but differ in `fields` are then possible, but harmless and highly unlikely in practice. A follow-up commit adopts it as the `category_specific_info` field on every electrical component. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Adopt `CategorySpecificInfo | None` as the `category_specific_info` field on `ElectricalComponent`, replacing the bare `Mapping[str, Any]` that only the mismatched component ever populated. The protobuf `category_specific_info` oneof is now decoded once into its `kind` plus a dict of its fields, then each concrete component drops the fields it translated into typed attributes (e.g. a battery's `type`), keeping the `kind` and any leftover. The result: * recognized components keep the variant `kind` with an empty leftover; * `UnrecognizedElectricalComponent` now preserves the carried info instead of dropping it — previously the escape hatch was empty for the very class that needs it most; * `MismatchedCategoryElectricalComponent` keeps the full info (nothing was translated), recording exactly which variant disagreed with the category; * components carrying no variant get `None`. Because `CategorySpecificInfo` is hashable, the field no longer needs `hash=False` and now participates in the component hash like the other fields. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Replace the `{id}<{Class}>{name}` rendering with `{id}:{name}:{Class}`, and override `__str__` on the problematic subclasses so the raw wire value that makes them problematic stays visible in the marker: * recognized components render as `CID42:bat1:LiIonBattery` (the name is always literal, so an empty name reads as `CID42::LiIonBattery`) * `UnrecognizedBattery`/`UnrecognizedEvCharger`/`UnrecognizedInverter` render as `CID42:bat1:Battery:type=99` — the hardcoded base label plus the raw type * `UnrecognizedElectricalComponent` renders as `CID42:comp1:category=999` — the raw, unrecognized category * `MismatchedCategoryElectricalComponent` renders as `CID42:comp1:mismatched:category=5:kind=inverter` — the declared category (as its enum name) against the carried variant kind. Each subclass overrides `__str__` in full rather than sharing a `__str__` hook on the base: the format is a single short f-string, the tests pin it down, and the flat overrides read better than the indirection a hook would need. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

`MismatchedCategoryElectricalComponent.__str__` could only show the declared category as its raw int (`category=5`), because resolving the name would mean using the deprecated `ElectricalComponentCategory` wrapper enum. Store a `category_name` alongside the raw `category`, resolved by the converter from the protobuf enum descriptor (which is not deprecated) with the long `ELECTRICAL_COMPONENT_CATEGORY_` prefix stripped, so `__str__` renders the short, readable name: `CID42:comp1:mismatched:category=BATTERY:kind=inverter` Only the mismatched component needs it: its declared category is a recognized value that its class name hides. Unrecognized components have no name to show (that is what makes them unrecognized), and every other component encodes its category in its class name. Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Copilot AI review requested due to automatic review settings July 21, 2026 12:41

Copilot started reviewing on behalf of llucax July 21, 2026 12:42 View session

llucax force-pushed the comp-specific-info branch from 996ca26 to a9bce0f Compare July 21, 2026 12:46

Copilot AI reviewed Jul 21, 2026

View reviewed changes

Copilot AI review requested due to automatic review settings July 21, 2026 12:47

Copilot started reviewing on behalf of llucax July 21, 2026 12:48 View session

Copilot AI reviewed Jul 21, 2026

View reviewed changes

Comment thread src/frequenz/client/common/microgrid/electrical_components/_electrical_component.py

llucax mentioned this pull request Jul 21, 2026

Improve exception messages #258

Draft

llucax force-pushed the comp-specific-info branch 4 times, most recently from 4b0aad6 to 14739e2 Compare July 22, 2026 09:12

llucax added 12 commits August 11, 2026 12:20

Update release notes

43ea977

Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Use FloatInt in metric bounds

558e20e

Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

llucax added 9 commits August 11, 2026 12:21

Use FloatInt in metric samples

f2c4f36

Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Use FloatInt in Location

f7011dc

Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Use FloatInt in PowerTransformer

380314f

Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

Update release notes

c2395af

Signed-off-by: Leandro Lucarella <luca-frequenz@llucax.com>

llucax force-pushed the comp-specific-info branch from 14739e2 to baa37c5 Compare August 11, 2026 11:24

Navigation Menu

Uh oh!

Add a new `CategorySpecificInfo` wrapper - #257

Add a new `CategorySpecificInfo` wrapper#257
llucax wants to merge 21 commits into
frequenz-floss:v0.x.xfrom
llucax:comp-specific-info

llucax commented Jul 21, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

llucax commented Jul 21, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

llucax commented Jul 21, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

llucax commented Jul 21, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants