Skip to content

A reactive notebook for Python — run reproducible experiments, query with SQL, execute as a script, deploy as an app, and version with git. Stored as pure Python. All in a modern, AI-native editor.

share types and function between table and dataframe for download (#6497)

Shahmir Varqha666fc6906b75
diff --git a/frontend/src/components/data-table/schemas.ts b/frontend/src/components/data-table/schemas.ts
new file mode 100644
index 0000000000..76ee9e3d87
--- /dev/null
+++ b/frontend/src/components/data-table/schemas.ts
@@ -0,0 +1,16 @@
+/* Copyright 2024 Marimo. All rights reserved. */
+
+import z from "zod";
+import { rpc } from "@/plugins/core/rpc";
+
+export type DownloadAsArgs = (req: {
+ format: "csv" | "json" | "parquet";
+}) => Promise<string>;
+
+export const DownloadAsSchema = rpc
+ .input(
+ z.object({
+ format: z.enum(["csv", "json", "parquet"]),
+ }),
+ )
+ .output(z.string());
diff --git a/frontend/src/plugins/impl/DataTablePlugin.tsx b/frontend/src/plugins/impl/DataTablePlugin.tsx
index dcf7d8bb92..4381498dfb 100644
--- a/frontend/src/plugins/impl/DataTablePlugin.tsx
+++ b/frontend/src/plugins/impl/DataTablePlugin.tsx
@@ -36,6 +36,10 @@ import {
import { usePanelOwnership } from "@/components/data-table/hooks/use-panel-ownership";
import { LoadingTable } from "@/components/data-table/loading-table";
import { RowViewerPanel } from "@/components/data-table/row-viewer-panel/row-viewer";
+import {
+ type DownloadAsArgs,
+ DownloadAsSchema,
+} from "@/components/data-table/schemas";
import {
type BinValues,
type ColumnHeaderStats,
@@ -191,7 +195,7 @@ interface Data<T> {
// eslint-disable-next-line @typescript-eslint/consistent-type-definitions
type DataTableFunctions = {
- download_as: (req: { format: "csv" | "json" | "parquet" }) => Promise<string>;
+ download_as: DownloadAsArgs;
get_column_summaries: <T>(
opts: ColumnSummariesArgs,
) => Promise<ColumnSummaries<T>>;
@@ -265,9 +269,7 @@ export const DataTablePlugin = createPlugin<S>("marimo-table")
}),
)
.withFunctions<DataTableFunctions>({
- download_as: rpc
- .input(z.object({ format: z.enum(["csv", "json", "parquet"]) }))
- .output(z.string()),
+ download_as: DownloadAsSchema,
get_column_summaries: rpc
.input(z.object({ precompute: z.boolean() }))
.output(
diff --git a/frontend/src/plugins/impl/data-frames/DataFramePlugin.tsx b/frontend/src/plugins/impl/data-frames/DataFramePlugin.tsx
index e89cee1795..b3018dcc28 100644
--- a/frontend/src/plugins/impl/data-frames/DataFramePlugin.tsx
+++ b/frontend/src/plugins/impl/data-frames/DataFramePlugin.tsx
@@ -4,6 +4,10 @@ import { isEqual } from "lodash-es";
import { Code2Icon, DatabaseIcon, FunctionSquareIcon } from "lucide-react";
import { type JSX, memo, useEffect, useRef, useState } from "react";
import { z } from "zod";
+import {
+ type DownloadAsArgs,
+ DownloadAsSchema,
+} from "@/components/data-table/schemas";
import type { FieldTypesWithExternalType } from "@/components/data-table/types";
import { ReadonlyCode } from "@/components/editor/code/readonly-python-code";
import { Spinner } from "@/components/icons/spinner";
@@ -68,7 +72,7 @@ type PluginFunctions = {
data: TableData<T>;
total_rows: number;
}>;
- download_as: (req: { format: "csv" | "json" | "parquet" }) => Promise<string>;
+ download_as: DownloadAsArgs;
};
// Value is selection, but it is not currently exposed to the user
@@ -127,13 +131,7 @@ export const DataFramePlugin = createPlugin<S>("marimo-dataframe")
total_rows: z.number(),
}),
),
- download_as: rpc
- .input(
- z.object({
- format: z.enum(["csv", "json", "parquet"]),
- }),
- )
- .output(z.string()),
+ download_as: DownloadAsSchema,
})
.renderer((props) => (
<TableProviders>
@@ -152,7 +150,7 @@ interface DataTableProps extends Data, PluginFunctions {
setValue: (value: S) => void;
host: HTMLElement;
showDownload: boolean;
- download_as: (req: { format: "csv" | "json" | "parquet" }) => Promise<string>;
+ download_as: DownloadAsArgs;
}
const EMPTY: Transformations = {
diff --git a/marimo/_plugins/ui/_impl/dataframes/dataframe.py b/marimo/_plugins/ui/_impl/dataframes/dataframe.py
index f5646227f7..f987659e4b 100644
--- a/marimo/_plugins/ui/_impl/dataframes/dataframe.py
+++ b/marimo/_plugins/ui/_impl/dataframes/dataframe.py
@@ -13,7 +13,6 @@ from typing import (
Union,
)
-import marimo._output.data.data as mo_data
from marimo._output.rich_help import mddoc
from marimo._plugins.ui._core.ui_element import UIElement
from marimo._plugins.ui._impl.dataframes.transforms.apply import (
@@ -38,6 +37,7 @@ from marimo._plugins.ui._impl.tables.table_manager import (
from marimo._plugins.ui._impl.tables.utils import (
get_table_manager,
)
+from marimo._plugins.ui._impl.utils.dataframe import download_as
from marimo._plugins.validators import (
validate_no_integer_columns,
validate_page_size,
@@ -298,18 +298,7 @@ class dataframe(UIElement[dict[str, Any], DataFrameType]):
# Get the table manager for the transformed data
manager = self._get_cached_table_manager(df, self._limit)
-
- ext = args.format
- if ext == "csv":
- return mo_data.csv(manager.to_csv()).url
- elif ext == "json":
- return mo_data.json(manager.to_json()).url
- elif ext == "parquet":
- return mo_data.parquet(manager.to_parquet()).url
- else:
- raise ValueError(
- "format must be one of 'csv', 'json', or 'parquet'."
- )
+ return download_as(manager, args.format)
def _apply_filters_query_sort(
self,
diff --git a/marimo/_plugins/ui/_impl/table.py b/marimo/_plugins/ui/_impl/table.py
index 2976e248c8..3a478e0c43 100644
--- a/marimo/_plugins/ui/_impl/table.py
+++ b/marimo/_plugins/ui/_impl/table.py
@@ -48,7 +48,11 @@ from marimo._plugins.ui._impl.tables.table_manager import (
TableManager,
)
from marimo._plugins.ui._impl.tables.utils import get_table_manager
-from marimo._plugins.ui._impl.utils.dataframe import ListOrTuple, TableData
+from marimo._plugins.ui._impl.utils.dataframe import (
+ ListOrTuple,
+ TableData,
+ download_as,
+)
from marimo._plugins.validators import (
validate_no_integer_columns,
validate_page_size,
@@ -806,17 +810,9 @@ class table(
# Remove the selection column before downloading
if isinstance(manager_candidate, TableManager):
- manager = manager_candidate.drop_columns([INDEX_COLUMN_NAME])
-
- ext = args.format
- if ext == "csv":
- return mo_data.csv(manager.to_csv()).url
- elif ext == "json":
- return mo_data.json(manager.to_json()).url
- elif ext == "parquet":
- return mo_data.parquet(manager.to_parquet()).url
- else:
- raise ValueError("format must be one of 'csv' or 'json'.")
+ return download_as(
+ manager_candidate, args.format, drop_marimo_index=True
+ )
else:
raise NotImplementedError(
"Download is not supported for this table format."
diff --git a/marimo/_plugins/ui/_impl/utils/dataframe.py b/marimo/_plugins/ui/_impl/utils/dataframe.py
index 5cd2ccc7fe..7366833c71 100644
--- a/marimo/_plugins/ui/_impl/utils/dataframe.py
+++ b/marimo/_plugins/ui/_impl/utils/dataframe.py
@@ -1,16 +1,16 @@
# Copyright 2024 Marimo. All rights reserved.
from __future__ import annotations
-from typing import (
- TypeVar,
- Union,
-)
+from typing import Any, TypeVar, Union
from narwhals.typing import IntoDataFrame
from marimo import _loggers
+from marimo._output.data import data as mo_data
from marimo._output.mime import MIME
from marimo._plugins.core.web_component import JSONType
+from marimo._plugins.ui._impl.tables.selection import INDEX_COLUMN_NAME
+from marimo._plugins.ui._impl.tables.table_manager import TableManager
LOGGER = _loggers.marimo_logger()
@@ -26,3 +26,35 @@ TableData = Union[
dict[str, ListOrTuple[JSONType]],
IntoDataFrame,
]
+
+
+def download_as(
+ manager: TableManager[Any], ext: str, drop_marimo_index: bool = False
+) -> str:
+ """Download the table data in the specified format.
+
+ Args:
+ manager (TableManager[Any]): The table manager to download.
+ ext (str): The format to download the table data in.
+ drop_marimo_index (bool, optional): Whether to drop the marimo selection column.
+ Defaults to False.
+
+ Raises:
+ ValueError: If unrecognized format.
+ NotImplementedError: If the table format is not supported.
+
+ Returns:
+ str: The URL to download the table data.
+ """
+ if drop_marimo_index:
+ # Remove the selection column if exists
+ manager = manager.drop_columns([INDEX_COLUMN_NAME])
+
+ if ext == "csv":
+ return mo_data.csv(manager.to_csv()).url
+ elif ext == "json":
+ return mo_data.json(manager.to_json()).url
+ elif ext == "parquet":
+ return mo_data.parquet(manager.to_parquet()).url
+ else:
+ raise ValueError("format must be one of 'csv', 'json', or 'parquet'.")