feat: feature description and similar features list

package.json

@@ -20,7 +20,7 @@
     "test:versions": "node scripts/check-versions.mjs",
     "test": "npm run test:versions && npm run test:js && npm run test:python && npm run test:rust",
     "check": "npm run check -w @embedding-atlas/component -w @embedding-atlas/viewer -w @embedding-atlas/embedding-atlas -w @embedding-atlas/examples -w @embedding-atlas/backend",
-    "check-format": "prettier -c . && uvx ruff format --check --exclude '*.ipynb' --exclude '*.md'"
+    "check-format": "prettier -c . && uv run --frozen --only-group lint ruff format --check --exclude '*.ipynb' --exclude '*.md'"
   },
   "devDependencies": {
     "prettier": "^3.8.1",

packages/backend/embedding_atlas/async_map.py

@@ -2,7 +2,8 @@
 
 import asyncio
 import random
-from typing import Awaitable, Callable, TypeVar
+from collections.abc import Awaitable, Callable
+from typing import TypeVar
 
 from tqdm.auto import tqdm
 

packages/backend/embedding_atlas/cache.py

@@ -31,16 +31,17 @@
 """
 
 import base64
+import functools
 import hashlib
 import hmac
 import json
 import logging
 import secrets
 import struct
-from functools import lru_cache
+from collections.abc import Callable
 from io import BytesIO, TextIOWrapper
 from pathlib import Path
-from typing import IO, Any, Callable
+from typing import IO, Any
 
 import numpy as np
 from cryptography.hazmat.primitives import hashes
@@ -267,7 +268,7 @@ async def async_file_cache_value(
 
     if cache_path.exists():
         try:
-            with open(cache_path, "rb") as file:
+            with open(cache_path, "rb") as file:  # noqa: ASYNC230
                 data = _decrypt_data(file.read(), key=encryption_key)
 
             result = deserializer(BytesIO(data))
@@ -290,7 +291,7 @@ async def async_file_cache_value(
         serializer(value, buffer)
         encrypted_data = _encrypt_data(buffer.getvalue(), key=encryption_key)
 
-        with open(cache_path_tmp, "wb") as file:
+        with open(cache_path_tmp, "wb") as file:  # noqa: ASYNC230
             file.write(encrypted_data)
 
         cache_path_tmp.rename(cache_path)
@@ -307,7 +308,7 @@ def _resolve_cache_root(cache_root: str | Path | None = None) -> Path:
         return Path(cache_root).resolve()
 
 
-@lru_cache(maxsize=None)
+@functools.cache
 def _get_constants(cache_root: Path) -> dict[str, bytes]:
     cache_root.mkdir(parents=True, exist_ok=True)
     constants_path = cache_root / "cache_constants.json"

packages/backend/embedding_atlas/cli.py

@@ -390,6 +390,15 @@ def import_modules(names: list[str]):
     default=None,
     help="Path to a file containing labels for the embedding view. The file should be a table with columns 'x', 'y', 'text', and optionally 'level' and 'priority'",
 )
+@click.option(
+    "--state",
+    "initial_state",
+    type=JSONParamType(),
+    default=None,
+    help="Initial app state (charts, layouts, column styles, ...), as exported from the app. "
+    "Pass a JSON string or a path to a JSON file. If the state contains charts, "
+    "they replace the default charts.",
+)
 @click.option(
     "--mcp/--no-mcp",
     "enable_mcp",
@@ -441,6 +450,7 @@ def main(
     point_size: float | None,
     stop_words: str | None,
     labels: str | None,
+    initial_state: dict | None,
     enable_mcp: bool,
 ):
     apply_logging_config()
@@ -654,6 +664,7 @@ def main(
         point_size=point_size,
         stop_words=stop_words_resolved,
         labels=labels_resolved,
+        initial_state=initial_state,
         additional_tables=additional_tables_meta or None,
     )
 

packages/backend/embedding_atlas/options.py

@@ -54,15 +54,6 @@ class EmbeddingAtlasOptions(TypedDict, total=False):
     point_size:
         Override the default point size for the embedding view.
 
-    show_table:
-        Whether to display the data table when the widget opens.
-
-    show_charts:
-        Whether to display charts when the widget opens.
-
-    show_embedding:
-        Whether to display the embedding view when the widget opens.
-
     initial_state:
         The initial Embedding Atlas state.
 
@@ -89,10 +80,6 @@ class EmbeddingAtlasOptions(TypedDict, total=False):
     labels: list[dict] | None
     stop_words: list[str] | None
 
-    show_table: bool | None
-    show_charts: bool | None
-    show_embedding: bool | None
-
     initial_state: dict | None
 
     additional_tables: dict | None
@@ -149,11 +136,4 @@ def set_prop(key: str, value):
     # Additional tables (id/relation for cross-table filtering)
     set_prop("additionalTables", options.get("additional_tables"))
 
-    # Layout
-    set_prop("initialState.layoutStates.list.showTable", options.get("show_table"))
-    set_prop("initialState.layoutStates.list.showCharts", options.get("show_charts"))
-    set_prop(
-        "initialState.layoutStates.list.showEmbedding", options.get("show_embedding")
-    )
-
     return props

packages/backend/embedding_atlas/server.py

@@ -5,9 +5,9 @@
 import json
 import re
 import uuid
+from collections.abc import Callable
 from functools import lru_cache, partial
 from io import BytesIO
-from typing import Callable
 
 import duckdb
 from fastapi import FastAPI, HTTPException, Request, Response, WebSocket
@@ -282,7 +282,7 @@ async def send_request(self, request: dict) -> dict:
                 raise HTTPException(status_code=503, detail="WebSocket disconnected")
             else:
                 raise HTTPException(
-                    status_code=500, detail=f"Internal server error: {str(e)}"
+                    status_code=500, detail=f"Internal server error: {e!s}"
                 )
 
     async def send_close(self):

packages/backend/embedding_atlas/streamlit.py

@@ -59,15 +59,6 @@ def embedding_atlas(
         point_size:
             Override the default point size for the embedding view.
 
-        show_table:
-            Whether to display the data table when the widget opens.
-
-        show_charts:
-            Whether to display charts when the widget opens.
-
-        show_embedding:
-            Whether to display the embedding view when the widget opens.
-
         key:
             The key of the Streamlit widget.
 

packages/backend/embedding_atlas/utils.py

@@ -1,6 +1,7 @@
 # Copyright (c) 2025 Apple Inc. Licensed under MIT License.
 
 import logging
+import sys
 from pathlib import Path
 from typing import Any
 
@@ -35,7 +36,7 @@ def load_huggingface_data(filename: str, splits: list[str] | None) -> pd.DataFra
         print(
             "⚠️ Loading Hugging Face datasets requires the `datasets` package to be installed. Please run `pip install datasets`, then try again."
         )
-        exit(-1)
+        sys.exit(-1)
 
     if Path(filename).is_dir():
         ds: Any = load_from_disk(filename)
@@ -51,7 +52,7 @@ def load_huggingface_data(filename: str, splits: list[str] | None) -> pd.DataFra
 
     if splits is None or len(splits) == 0:
         ds_split_options = []
-        for key in ds.keys():
+        for key in ds:
             option = (f"{key} ({ds[key].num_rows} rows)", key)
             ds_split_options.append(option)
         split_question = [

packages/backend/embedding_atlas/widget.py

@@ -85,15 +85,6 @@ def __init__(
             point_size:
                 Override the default point size for the embedding view.
 
-            show_table:
-                Whether to display the data table when the widget opens.
-
-            show_charts:
-                Whether to display charts when the widget opens.
-
-            show_embedding:
-                Whether to display the embedding view when the widget opens.
-
             connection (DuckDBPyConnection, optional):
                 A DuckDB connection. Defaults to duckdb.connect().
         """

packages/backend/examples/streamlit.py

@@ -39,7 +39,6 @@ def main():
         x="projection_x",
         y="projection_y",
         neighbors="neighbors",
-        show_table=True,
     )
 
     # Show selected rows in a Streamlit data frame

packages/docs/generate_cache.py

@@ -32,7 +32,7 @@ def generate_dataset_embedding(
     modality: str = "text",
     model: str | None = "all-MiniLM-L6-v2",
     pagerank: bool = False,
-    umap_args: dict = {},
+    umap_args: dict | None = None,
 ):
     click.echo(click.style(f"Processing {url}", fg="cyan"))
 
@@ -48,6 +48,8 @@ def generate_dataset_embedding(
     _ = data_frame
     df = duckdb.query(query).to_df()
 
+    if umap_args is None:
+        umap_args = {}
     umap_args = {"random_state": 42} | umap_args
 
     df = compute_projection(

packages/docs/streamlit.md

@@ -24,7 +24,6 @@ df = compute_projection(df, inputs="description", modality="text",
 value = embedding_atlas(
     df, text="description",
     x="projection_x", y="projection_y", neighbors="neighbors",
-    show_table=True
 )
 ```
 

packages/umap/benchmark/benchmark/__main__.py

@@ -6,7 +6,6 @@
 @click.group()
 def cli():
     """UMAP / NNDescent benchmark suite."""
-    pass
 
 
 @cli.command()

packages/umap/benchmark/benchmark/align.py

@@ -22,7 +22,7 @@ def _kabsch(emb, ref):
     with warnings.catch_warnings():
         warnings.simplefilter("ignore")
         H = E.T @ R
-        U, S, Vt = np.linalg.svd(H)
+        U, _, Vt = np.linalg.svd(H)
     if np.any(~np.isfinite(U)) or np.any(~np.isfinite(Vt)):
         return emb.astype(np.float32), np.inf
 

packages/umap/benchmark/benchmark/common.py

@@ -60,7 +60,9 @@ def run_cmd(cmd, desc="", timeout=None, threads="multi"):
     print(f"{'=' * 60}")
     t0 = time.perf_counter()
     try:
-        result = subprocess.run(cmd, timeout=timeout, env=bench_env(threads))
+        result = subprocess.run(
+            cmd, timeout=timeout, env=bench_env(threads), check=False
+        )
         elapsed = time.perf_counter() - t0
         if result.returncode != 0:
             print(f"  FAILED (exit code {result.returncode}) in {elapsed:.1f}s")
@@ -100,7 +102,7 @@ def __init__(self, path, columns):
         Path(path).parent.mkdir(parents=True, exist_ok=True)
         self.path = path
         self.columns = columns
-        self._file = open(path, "w", newline="")
+        self._file = open(path, "w", newline="")  # noqa: SIM115
         self._writer = csv.DictWriter(
             self._file, fieldnames=columns, extrasaction="ignore"
         )

packages/umap/benchmark/benchmark/html_report.py

@@ -223,7 +223,7 @@ def _load_embeddings(ds_dir, rows):
         # Align to reference (Rust CPU preferred)
         ref_name = next((name for _, name in _IMPL_ORDER if name in embeddings), None)
         ref_emb = embeddings[ref_name]
-        for name in embeddings:
+        for name in embeddings:  # noqa: PLC0206
             if name != ref_name:
                 embeddings[name] = align_to_reference(embeddings[name], ref_emb)
 
@@ -233,7 +233,7 @@ def _load_embeddings(ds_dir, rows):
         maxs = all_coords.max(axis=0)
         ranges = maxs - mins
         ranges[ranges == 0] = 1
-        for name in embeddings:
+        for name in embeddings:  # noqa: PLC0206
             embeddings[name] = ((embeddings[name] - mins) / ranges).astype(np.float32)
 
         # Serialize

packages/viewer/src/assets/chart_icons.ts

@@ -6,6 +6,7 @@ import chart_cards from "./chart-cards.svg?raw";
 import chart_content_viewer from "./chart-content-viewer.svg?raw";
 import chart_ecdf from "./chart-ecdf.svg?raw";
 import chart_embedding from "./chart-embedding.svg?raw";
+import chart_features_list from "./chart-features-list.svg?raw";
 import chart_h_bar from "./chart-h-bar.svg?raw";
 import chart_heatmap from "./chart-heatmap.svg?raw";
 import chart_line from "./chart-line.svg?raw";
@@ -24,6 +25,7 @@ export const chartIcons: Record<string, string> = {
   "chart-content-viewer": chart_content_viewer,
   "chart-ecdf": chart_ecdf,
   "chart-embedding": chart_embedding,
+  "chart-features-list": chart_features_list,
   "chart-h-bar": chart_h_bar,
   "chart-heatmap": chart_heatmap,
   "chart-line": chart_line,

packages/viewer/src/assets/icons.ts

@@ -38,6 +38,8 @@ export { default as IconPin } from "~icons/fluent/pin-16-filled";
 export { default as IconPinRegular } from "~icons/fluent/pin-16-regular";
 export { default as IconSelected } from "~icons/fluent/checkmark-circle-16-filled";
 export { default as IconUnselected } from "~icons/fluent/circle-16-regular";
+export { default as IconInfo } from "~icons/fluent/info-16-filled";
+export { default as IconInfoRegular } from "~icons/fluent/info-16-regular";
 
 // Icons for list layout view toggles (hide/show parts)
 export { default as IconEmbeddingView } from "~icons/fluent/data-scatter-20-filled";