codeant-ai-for-open-source[bot] commented on code in PR #40099:
URL: https://github.com/apache/superset/pull/40099#discussion_r3238423327
##########
superset/mcp_service/chart/tool/get_chart_preview.py:
##########
@@ -422,22 +436,27 @@ def generate(self) -> VegaLitePreview | ChartError:
# Build columns list: include both x_axis and groupby
columns = _build_query_columns(form_data)
+ prepare_form_data_for_query(
+ form_data,
+ self.chart.datasource_id,
+ self.chart.datasource_type,
+ )
+ query_dict = {
+ "columns": columns,
+ "metrics": form_data.get("metrics", []),
+ "row_limit": 1000, # More data for visualization
+ "order_desc": True,
+ }
+ apply_form_data_filters_to_query(query_dict, form_data)
+
# Create query context for data retrieval
factory = QueryContextFactory()
query_context = factory.create(
datasource={
"id": self.chart.datasource_id,
"type": self.chart.datasource_type,
},
- queries=[
- {
- "filters": form_data.get("filters", []),
- "columns": columns,
- "metrics": form_data.get("metrics", []),
- "row_limit": 1000, # More data for visualization
- "order_desc": True,
- }
- ],
+ queries=[query_dict],
form_data=form_data,
force=self.request.force_refresh,
)
Review Comment:
**Suggestion:** Preview generation now injects a handcrafted query using
only `columns` and `metrics`, where `metrics` is read only from the plural key.
Charts that store measures/dimensions in alternative fields (such as singular
`metric`, `entity`/`series`, or secondary series fields) can produce
empty/incorrect previews even though the chart is valid. Reuse the same
chart-type-aware metric/groupby resolution used by data/SQL paths before
constructing preview queries. [incomplete implementation]
<details>
<summary><b>Severity Level:</b> Major ⚠️</summary>
```mdx
- ❌ MCP table previews can be empty for valid bubble charts.
- ❌ Vega-Lite previews drop mixed_timeseries secondary series data.
- ⚠️ LLM chart previews may misrepresent underlying saved charts.
```
</details>
<details>
<summary><b>Steps of Reproduction ✅ </b></summary>
```mdx
1. Save a chart whose measures live in non-`metrics` fields, for example a
`bubble` chart
using `x`/`y`/`size` metrics or a big-number-style chart using singular
`metric` (these
chart-type-specific fields are documented in the data tool fallback at
superset/mcp_service/chart/tool/get_chart_data.py:37-68 and bubble handling
at 50-57).
2. From an MCP client, call `get_chart_preview`
(superset/mcp_service/chart/tool/get_chart_preview.py:17-23, 43-62) with
`format="vega_lite"` or `format="table"` so the request is routed to
`VegaLitePreviewStrategy.generate` or `TablePreviewStrategy.generate`.
3. Inside the selected strategy, `form_data` is loaded from
`self.chart.params` and passed
through `prepare_form_data_for_query` (TablePreviewStrategy at 327-333,
VegaLitePreviewStrategy at 439-443), then `query_dict` is built with
`columns` from
`_build_query_columns(form_data)` and `metrics = form_data.get("metrics",
[])` at lines
335-337 and 444-447, so alternate metric fields like `metric`,
`x`/`y`/`size`, or
`metrics_b` for `mixed_timeseries` are ignored.
4. `QueryContextFactory.create(..., queries=[query_dict],
form_data=form_data, ...)` is
called (superset/mcp_service/chart/tool/get_chart_preview.py:342-351,
453-462;
superset/common/query_context_factory.py:45-89) and `ChartDataCommand.run()`
executes with
this incomplete query, which for charts lacking a `metrics` array or
standard `groupby`
yields empty or erroring results, even though the same chart returns data
correctly via
SQL/data paths that use richer metric/groupby resolution
(superset/mcp_service/chart/tool/get_chart_sql.py:78-88, 122-146, 184-200).
```
</details>
[Fix in
Cursor](https://app.codeant.ai/fix-in-ide?tool=cursor&prompt=This%20is%20a%20comment%20left%20during%20a%20code%20review.%0A%0A%2A%2APath%3A%2A%2A%20superset%2Fmcp_service%2Fchart%2Ftool%2Fget_chart_preview.py%0A%2A%2ALine%3A%2A%2A%20444%3A462%0A%2A%2AComment%3A%2A%2A%0A%09%2AIncomplete%20Implementation%3A%20Preview%20generation%20now%20injects%20a%20handcrafted%20query%20using%20only%20%60columns%60%20and%20%60metrics%60%2C%20where%20%60metrics%60%20is%20read%20only%20from%20the%20plural%20key.%20Charts%20that%20store%20measures%2Fdimensions%20in%20alternative%20fields%20%28such%20as%20singular%20%60metric%60%2C%20%60entity%60%2F%60series%60%2C%20or%20secondary%20series%20fields%29%20can%20produce%20empty%2Fincorrect%20previews%20even%20though%20the%20chart%20is%20valid.%20Reuse%20the%20same%20chart-type-aware%20metric%2Fgroupby%20resolution%20used%20by%20data%2FSQL%20paths%20before%20constructing%20preview%20queries.%0A%0AValidate%20the%20correctness%20of%20the%20flagged
%20issue.%20If%20correct%2C%20How%20can%20I%20resolve%20this%3F%20If%20you%20propose%20a%20fix%2C%20implement%20it%20and%20please%20make%20it%20concise.%0AOnce%20fix%20is%20implemented%2C%20also%20check%20other%20comments%20on%20the%20same%20PR%2C%20and%20ask%20user%20if%20the%20user%20wants%20to%20fix%20the%20rest%20of%20the%20comments%20as%20well.%20if%20said%20yes%2C%20then%20fetch%20all%20the%20comments%20validate%20the%20correctness%20and%20implement%20a%20minimal%20fix%0A)
| [Fix in VSCode
Claude](https://app.codeant.ai/fix-in-ide?tool=vscode-claude&prompt=This%20is%20a%20comment%20left%20during%20a%20code%20review.%0A%0A%2A%2APath%3A%2A%2A%20superset%2Fmcp_service%2Fchart%2Ftool%2Fget_chart_preview.py%0A%2A%2ALine%3A%2A%2A%20444%3A462%0A%2A%2AComment%3A%2A%2A%0A%09%2AIncomplete%20Implementation%3A%20Preview%20generation%20now%20injects%20a%20handcrafted%20query%20using%20only%20%60columns%60%20and%20%60metrics%60%2C%20where%20%60metrics%60%20is%20read%20only%20from%20the%20pl
ural%20key.%20Charts%20that%20store%20measures%2Fdimensions%20in%20alternative%20fields%20%28such%20as%20singular%20%60metric%60%2C%20%60entity%60%2F%60series%60%2C%20or%20secondary%20series%20fields%29%20can%20produce%20empty%2Fincorrect%20previews%20even%20though%20the%20chart%20is%20valid.%20Reuse%20the%20same%20chart-type-aware%20metric%2Fgroupby%20resolution%20used%20by%20data%2FSQL%20paths%20before%20constructing%20preview%20queries.%0A%0AValidate%20the%20correctness%20of%20the%20flagged%20issue.%20If%20correct%2C%20How%20can%20I%20resolve%20this%3F%20If%20you%20propose%20a%20fix%2C%20implement%20it%20and%20please%20make%20it%20concise.%0AOnce%20fix%20is%20implemented%2C%20also%20check%20other%20comments%20on%20the%20same%20PR%2C%20and%20ask%20user%20if%20the%20user%20wants%20to%20fix%20the%20rest%20of%20the%20comments%20as%20well.%20if%20said%20yes%2C%20then%20fetch%20all%20the%20comments%20validate%20the%20correctness%20and%20implement%20a%20minimal%20fix%0A)
*(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/mcp_service/chart/tool/get_chart_preview.py
**Line:** 444:462
**Comment:**
*Incomplete Implementation: Preview generation now injects a
handcrafted query using only `columns` and `metrics`, where `metrics` is read
only from the plural key. Charts that store measures/dimensions in alternative
fields (such as singular `metric`, `entity`/`series`, or secondary series
fields) can produce empty/incorrect previews even though the chart is valid.
Reuse the same chart-type-aware metric/groupby resolution used by data/SQL
paths before constructing preview queries.
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%2F40099&comment_hash=80b26baa560ee433a51a60f0f361a83bd734f47402879ecf0bad1644c80de692&reaction=like'>👍</a>
| <a
href='https://app.codeant.ai/feedback?pr_url=https%3A%2F%2Fgithub.com%2Fapache%2Fsuperset%2Fpull%2F40099&comment_hash=80b26baa560ee433a51a60f0f361a83bd734f47402879ecf0bad1644c80de692&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]
