codeant-ai-for-open-source[bot] commented on code in PR #40954:
URL: https://github.com/apache/superset/pull/40954#discussion_r3391462137
##########
superset-frontend/plugins/plugin-chart-echarts/src/Timeseries/transformers.ts:
##########
@@ -319,9 +322,12 @@ export function transformSeries(
* same as the original series, otherwise uses separate colors
* */
const itemStyle: ItemStyleOption = {
- color: timeShiftColor
- ? colorScale(colorScaleKey, sliceId)
- : colorScale(seriesKey || forecastSeries.name, sliceId),
+ color: isAnomaly
+ ? (theme?.colorError ??
+ colorScale(seriesKey || forecastSeries.name, sliceId))
+ : timeShiftColor
+ ? colorScale(colorScaleKey, sliceId)
+ : colorScale(seriesKey || forecastSeries.name, sliceId),
Review Comment:
**Suggestion:** The new anomaly color assignment does not actually guarantee
red anomaly points when `colorByPrimaryAxis` is enabled, because this chart
path bypasses series-level `itemStyle` and applies per-point colors instead.
This breaks the intended behavior ("anomalies are red") in a valid chart
configuration. Ensure anomaly series bypass `applyColorByPrimaryAxis` coloring
(or force red in that branch) so anomaly markers stay red consistently. [logic
error]
<details>
<summary><b>Severity Level:</b> Major ⚠️</summary>
```mdx
- ⚠️ Timeseries anomaly points not reliably highlighted in red.
- ⚠️ Color-by-X-axis charts misrepresent anomaly visual semantics.
- ⚠️ Users may overlook anomalies due to inconsistent coloring.
```
</details>
<details>
<summary><b>Steps of Reproduction ✅ </b></summary>
```mdx
1. Configure a Time-series ECharts chart (e.g., Regular Line) that uses the
Timeseries
plugin so that `transformProps()` at
`superset-frontend/plugins/plugin-chart-echarts/src/Timeseries/transformProps.ts:220-307`
is used to build the series list and call `transformSeries()` for each
metric column.
2. In the chart control panel, enable anomaly detection so the backend
produces
`metric__anomaly` columns (verified by tests in
`superset/tests/unit_tests/pandas_postprocessing/test_anomaly.py`) and
enable the "Color
by X-axis / primary axis" option so `colorByPrimaryAxis` is `true` in the
`EchartsTimeseriesFormData` passed to `transformProps()` and forwarded into
`transformSeries()` (`transformProps.ts:256-293`).
3. At runtime, for the anomaly series, `transformSeries()` in
`superset-frontend/plugins/plugin-chart-echarts/src/Timeseries/transformers.ts:270-275`
sets `forecastSeries.type === ForecastSeriesEnum.Anomaly`, making
`isAnomaly` true and
computing a red series-level `itemStyle.color` using `theme?.colorError` in
the snippet at
lines 324-333; however, because `colorByPrimaryAxis` is true and
`Array.isArray(data)` is
true, the returned series has its `data` replaced by the result of
`applyColorByPrimaryAxis()` (`transformers.ts:379-387`) and omits the
`itemStyle` entirely
via `...(colorByPrimaryAxis ? {} : { itemStyle })` at line 397.
4. `applyColorByPrimaryAxis()` (`transformers.ts:172-195`) colors each point
by its
primary-axis value using `colorScale(colorKey, sliceId)` and ignores
anomaly-specific
styling, so in the rendered chart anomaly points are colored by X-axis
category rather
than consistently red, diverging from the intended "red anomaly markers"
behavior even
though the anomaly path is exercised through normal chart configuration.
```
</details>
[Fix in
Cursor](https://app.codeant.ai/fix-in-ide?tool=cursor&prompt_id=0ba6085e1a714987bfae29a7a6d2c563&service=github&base_url=https%3A%2F%2Fgithub.com&org=apache&repo=apache%2Fsuperset)
| [Fix in VSCode
Claude](https://app.codeant.ai/fix-in-ide?tool=vscode-claude&prompt_id=0ba6085e1a714987bfae29a7a6d2c563&service=github&base_url=https%3A%2F%2Fgithub.com&org=apache&repo=apache%2Fsuperset)
*(Use Cmd/Ctrl + Click for best experience)*
<details>
<summary><b>Prompt for AI Agent 🤖 </b></summary>
```mdx
This is a comment left during a code review.
**Path:**
superset-frontend/plugins/plugin-chart-echarts/src/Timeseries/transformers.ts
**Line:** 325:330
**Comment:**
*Logic Error: The new anomaly color assignment does not actually
guarantee red anomaly points when `colorByPrimaryAxis` is enabled, because this
chart path bypasses series-level `itemStyle` and applies per-point colors
instead. This breaks the intended behavior ("anomalies are red") in a valid
chart configuration. Ensure anomaly series bypass `applyColorByPrimaryAxis`
coloring (or force red in that branch) so anomaly markers stay red consistently.
Validate the correctness of the flagged issue. If correct, How can I resolve
this? If you propose a fix, implement it and please make it concise.
Once fix is implemented, also check other comments on the same PR, and ask
user if the user wants to fix the rest of the comments as well. if said yes,
then fetch all the comments validate the correctness and implement a minimal fix
```
</details>
<a
href='https://app.codeant.ai/feedback?pr_url=https%3A%2F%2Fgithub.com%2Fapache%2Fsuperset%2Fpull%2F40954&comment_hash=0854226ae046ab47b648129bbb32d74a9348e3a010685abcf19b6f1045418784&reaction=like'>👍</a>
| <a
href='https://app.codeant.ai/feedback?pr_url=https%3A%2F%2Fgithub.com%2Fapache%2Fsuperset%2Fpull%2F40954&comment_hash=0854226ae046ab47b648129bbb32d74a9348e3a010685abcf19b6f1045418784&reaction=dislike'>👎</a>
##########
superset-frontend/plugins/plugin-chart-echarts/src/utils/forecast.ts:
##########
@@ -26,14 +26,25 @@ import {
} from '../types';
import { sanitizeHtml } from './series';
-const seriesTypeRegex = new RegExp(
+const forecastSuffixRegex = new RegExp(
`(.+)(${ForecastSeriesEnum.ForecastLower}|${ForecastSeriesEnum.ForecastTrend}|${ForecastSeriesEnum.ForecastUpper})$`,
);
export const extractForecastSeriesContext = (
seriesName: OptionName,
): ForecastSeriesContext => {
const name = seriesName as string;
- const regexMatch = seriesTypeRegex.exec(name);
+
+ // Check for anomaly suffix first, then resolve nested forecast suffix
+ if (name.endsWith(ForecastSeriesEnum.Anomaly)) {
+ const stripped = name.slice(0, -ForecastSeriesEnum.Anomaly.length);
+ const forecastMatch = forecastSuffixRegex.exec(stripped);
+ return {
+ name: forecastMatch ? forecastMatch[1] : stripped,
+ type: ForecastSeriesEnum.Anomaly,
+ };
Review Comment:
**Suggestion:** The anomaly parsing now collapses nested forecast anomaly
series (for example `metric__yhat__anomaly`) to the same base key as regular
anomaly series (`metric__anomaly`). When both exist (such as forecast enabled
with zero future periods), tooltip aggregation merges them into one entry and
overwrites values/colors, producing incorrect tooltip output. Preserve distinct
keys for nested anomaly series (e.g., keep `stripped` as the name) and only
normalize for display later. [logic error]
<details>
<summary><b>Severity Level:</b> Major ⚠️</summary>
```mdx
- ⚠️ Timeseries charts can drop one of two anomaly series.
- ⚠️ Tooltips cannot distinguish original vs forecast anomalies.
- ⚠️ Users may misinterpret which values were anomalous.
```
</details>
<details>
<summary><b>Steps of Reproduction ✅ </b></summary>
```mdx
1. Enable Prophet forecast and anomaly detection together for a timeseries
metric in a
chart that uses the ECharts Timeseries plugin; configure forecast so that
the original
metric has no NaNs in the returned dataframe (e.g., zero or no future
periods). Backend
tests in
`superset/tests/unit_tests/pandas_postprocessing/test_anomaly.py:test_anomaly_detection_with_forecast`
confirm this produces both `metric__anomaly` and `metric__yhat__anomaly`
columns.
2. On the frontend, `transformProps()` in
`superset-frontend/plugins/plugin-chart-echarts/src/Timeseries/transformProps.ts:73-75`
calls `rebaseForecastDatum(data, verboseMap)`
(`superset-frontend/plugins/plugin-chart-echarts/src/utils/forecast.ts:138-170`).
Inside
`rebaseForecastDatum`, each column key is passed through
`extractForecastSeriesContext()`
(`forecast.ts:32-53`); for both `metric__anomaly` and
`metric__yhat__anomaly`, the code at
lines 38-44 detects the `__anomaly` suffix, strips it, optionally strips any
`__yhat*`
suffix via `forecastSuffixRegex`, and returns `ForecastSeriesContext { name:
'metric',
type: ForecastSeriesEnum.Anomaly }` for both columns.
3. Still in `rebaseForecastDatum`, when `verboseMap` has an entry for the
base metric key
`'metric'`, both anomaly columns compute the same `verboseKey`
`${verboseMap['metric']}${ForecastSeriesEnum.Anomaly}`
(`forecast.ts:147-151`). The second
anomaly column processed (typically `metric__yhat__anomaly`, as in the
backend test column
order) overwrites the first in `newRow`, so only one anomaly column remains
in the rebased
data and the other anomaly series' values are silently lost before
`extractSeries()`
builds the chart series
(`superset-frontend/plugins/plugin-chart-echarts/src/utils/series.ts:70-103`).
4. When hovering the chart, the tooltip formatter in
`superset-frontend/plugins/plugin-chart-echarts/src/Timeseries/transformProps.ts:1015-1079`
calls `extractForecastValuesFromTooltipParams()` (`forecast.ts:68-97`),
which again uses
`extractForecastSeriesContext(seriesId)` and aggregates by `context.name`.
Because both
base and forecast anomaly series are normalized to the same `name`
`'metric'`, any
remaining cases where both anomaly series survive (e.g., if `verboseMap`
lacked the base
key or in other consumers such as
`MixedTimeseries/transformProps.ts:790-797`) will be
merged into a single `ForecastValue` entry, causing one series' anomaly
value/color to
overwrite the other and preventing tooltips from distinguishing base vs
forecast
anomalies.
```
</details>
[Fix in
Cursor](https://app.codeant.ai/fix-in-ide?tool=cursor&prompt_id=1728d023c2bd4454a0962a94a0b03644&service=github&base_url=https%3A%2F%2Fgithub.com&org=apache&repo=apache%2Fsuperset)
| [Fix in VSCode
Claude](https://app.codeant.ai/fix-in-ide?tool=vscode-claude&prompt_id=1728d023c2bd4454a0962a94a0b03644&service=github&base_url=https%3A%2F%2Fgithub.com&org=apache&repo=apache%2Fsuperset)
*(Use Cmd/Ctrl + Click for best experience)*
<details>
<summary><b>Prompt for AI Agent 🤖 </b></summary>
```mdx
This is a comment left during a code review.
**Path:**
superset-frontend/plugins/plugin-chart-echarts/src/utils/forecast.ts
**Line:** 38:44
**Comment:**
*Logic Error: The anomaly parsing now collapses nested forecast anomaly
series (for example `metric__yhat__anomaly`) to the same base key as regular
anomaly series (`metric__anomaly`). When both exist (such as forecast enabled
with zero future periods), tooltip aggregation merges them into one entry and
overwrites values/colors, producing incorrect tooltip output. Preserve distinct
keys for nested anomaly series (e.g., keep `stripped` as the name) and only
normalize for display later.
Validate the correctness of the flagged issue. If correct, How can I resolve
this? If you propose a fix, implement it and please make it concise.
Once fix is implemented, also check other comments on the same PR, and ask
user if the user wants to fix the rest of the comments as well. if said yes,
then fetch all the comments validate the correctness and implement a minimal fix
```
</details>
<a
href='https://app.codeant.ai/feedback?pr_url=https%3A%2F%2Fgithub.com%2Fapache%2Fsuperset%2Fpull%2F40954&comment_hash=b31d42ab423e69c95cccb51294d63ee06ba7e1b3c48a5110484a4b5ec567bab4&reaction=like'>👍</a>
| <a
href='https://app.codeant.ai/feedback?pr_url=https%3A%2F%2Fgithub.com%2Fapache%2Fsuperset%2Fpull%2F40954&comment_hash=b31d42ab423e69c95cccb51294d63ee06ba7e1b3c48a5110484a4b5ec567bab4&reaction=dislike'>👎</a>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
