From e4989782c73aaa3a5b08e84d283f162a45af2420 Mon Sep 17 00:00:00 2001
From: Dhanushkumar Jv <kumardhanush509@gmail.com>
Date: Tue, 21 Jul 2026 22:37:24 +0530
Subject: [PATCH 1/5] Add Zarr reader for CLI, library, and Globus.

---
 aidrin/file_handling/file_parser.py         |  17 +-
 aidrin/file_handling/readers/root_reader.py |  17 +
 aidrin/file_handling/readers/structured.py  |  81 +++++
 aidrin/file_handling/readers/zarr_reader.py | 344 ++++++++++++++++++++
 pyproject.toml                              |   6 +
 tests/unit/test_structured_readers.py       | 187 +++++++++++
 web/routes/core.py                          |   7 +-
 web/templates/_components/globus_panel.html |   4 +-
 8 files changed, 658 insertions(+), 5 deletions(-)
 create mode 100644 aidrin/file_handling/readers/root_reader.py
 create mode 100644 aidrin/file_handling/readers/structured.py
 create mode 100644 aidrin/file_handling/readers/zarr_reader.py
 create mode 100644 tests/unit/test_structured_readers.py

diff --git a/aidrin/file_handling/file_parser.py b/aidrin/file_handling/file_parser.py
index 2657f9dc..8d8461b5 100644
--- a/aidrin/file_handling/file_parser.py
+++ b/aidrin/file_handling/file_parser.py
@@ -8,6 +8,7 @@
 from aidrin.file_handling.readers.json_reader import jsonReader
 from aidrin.file_handling.readers.npz_reader import npzReader
 from aidrin.file_handling.readers.parquet_reader import parquetReader
+from aidrin.file_handling.readers.zarr_reader import zarrReader
 
 # Notes:
 # To add support for new file types:
@@ -15,6 +16,8 @@
 #       (and optionally .parse(), .filter()) to 'file_readers'.
 # - Register the class in READER_MAP.
 # - Add a display name and extension to SUPPORTED_FILE_TYPES for the front end.
+# - Formats that are CLI/library/Globus-only (e.g. Zarr directories) go in
+#   GLOBUS_FILE_TYPES but not SUPPORTED_FILE_TYPES (local upload dropdown).
 
 # Reader Map. Used to create file type specific parsing
 READER_MAP = {
@@ -24,10 +27,11 @@
     ".json": jsonReader,
     ".h5": hdf5Reader,
     ".parquet": parquetReader,
+    ".zarr": zarrReader,
     # Add additional file types here
 }
 
-# Supported file types. Read on front end to create select features.
+# Supported file types. Read on front end to create local upload select features.
 SUPPORTED_FILE_TYPES = [
     (".csv", "CSV"),
     (".xls, .xlsb, .xlsx, .xlsm", "Excel"),
@@ -39,6 +43,15 @@
     # (file_type,file_type_name)
 ]
 
+# Globus (and other path-based surfaces) may include directory-shaped formats
+# that are not offered in the local browser upload dropdown.
+GLOBUS_FILE_TYPES = SUPPORTED_FILE_TYPES + [
+    (".zarr", "Zarr"),
+]
+
+# File types that accept selected_keys for multi-array / multi-dataset selection.
+_SELECTION_FILE_TYPES = {".h5", ".zarr"}
+
 # logger config
 file_upload_time_log = logging.getLogger("file_upload")
 
@@ -227,7 +240,7 @@ def read_file(file_info, columns=None):
 
         # Slow path: parse the source once.
         reader_cls = READER_MAP[file_type]
-        if file_type == ".h5":
+        if file_type in _SELECTION_FILE_TYPES:
             df = reader_cls(
                 file_path, file_upload_time_log, selected_keys=selected_keys
             ).read()
diff --git a/aidrin/file_handling/readers/root_reader.py b/aidrin/file_handling/readers/root_reader.py
new file mode 100644
index 00000000..576640e4
--- /dev/null
+++ b/aidrin/file_handling/readers/root_reader.py
@@ -0,0 +1,17 @@
+from aidrin.file_handling.readers.structured import (
+    INVENTORY_UNSUPPORTED,
+    InventoryResult,
+    StructuredFileReader,
+    make_inventory,
+)
+
+
+class rootReader(StructuredFileReader):
+    """ROOT file reader (scaffold). v1 will support TTrees only via uproot."""
+
+    def inventory(self) -> InventoryResult:
+        return make_inventory(INVENTORY_UNSUPPORTED)
+
+    def read(self):
+        self.logger.warning("ROOT read is not implemented yet.")
+        return None
diff --git a/aidrin/file_handling/readers/structured.py b/aidrin/file_handling/readers/structured.py
new file mode 100644
index 00000000..bd86b8ce
--- /dev/null
+++ b/aidrin/file_handling/readers/structured.py
@@ -0,0 +1,81 @@
+"""
+Shared contract for structured file readers (HDF5, Zarr, ROOT).
+
+Structured formats may hold many arrays or tables under paths. Readers classify
+the on-disk layout via ``inventory()`` before ``read()`` loads tabular data.
+
+Inventory ``type`` values
+-------------------------
+empty
+    No readable datasets or arrays.
+single_dataset
+    One dataset; auto-read without a picker.
+multi_dataset
+    Incompatible or grouped layout; caller must select paths.
+legacy
+    Internal auto-read path for compatible multi-array layouts (and HDF5
+    PyTables stores). Never surface this label in API or UI responses.
+unsupported
+    Reader stub or format not implemented yet. Never treat as auto-read.
+"""
+
+from __future__ import annotations
+
+from typing import Any, TypedDict
+
+from aidrin.file_handling.readers.base_reader import BaseFileReader
+
+INVENTORY_EMPTY = "empty"
+INVENTORY_SINGLE = "single_dataset"
+INVENTORY_MULTI = "multi_dataset"
+INVENTORY_LEGACY = "legacy"
+INVENTORY_UNSUPPORTED = "unsupported"
+
+# Types that may be shown to users when driving picker or error messages.
+USER_FACING_INVENTORY_TYPES = frozenset(
+    {INVENTORY_EMPTY, INVENTORY_SINGLE, INVENTORY_MULTI}
+)
+
+
+class DatasetEntry(TypedDict):
+    path: str
+    shape: tuple[int, ...]
+    ndim: int
+    dtype: str
+    size: int
+
+
+class PickerGroup(TypedDict, total=False):
+    name: str
+    paths: list[str]
+    prefix: str
+
+
+class InventoryResult(TypedDict):
+    type: str
+    datasets: list[DatasetEntry]
+    groups: list[dict[str, Any]]
+
+
+def make_inventory(
+    inv_type: str,
+    datasets: list[DatasetEntry] | None = None,
+    groups: list[dict[str, Any]] | None = None,
+) -> InventoryResult:
+    """Build a normalized inventory dict."""
+    return {
+        "type": inv_type,
+        "datasets": datasets or [],
+        "groups": groups or [],
+    }
+
+
+class StructuredFileReader(BaseFileReader):
+    """Base for hierarchical array/tree formats with inventory and metadata hooks."""
+
+    def inventory(self) -> InventoryResult:
+        return make_inventory(INVENTORY_UNSUPPORTED)
+
+    def get_metadata(self) -> dict[str, Any]:
+        """Embedded file metadata for FAIR assessment (``.zattrs``, ROOT headers, etc.)."""
+        return {}
diff --git a/aidrin/file_handling/readers/zarr_reader.py b/aidrin/file_handling/readers/zarr_reader.py
new file mode 100644
index 00000000..e7edf947
--- /dev/null
+++ b/aidrin/file_handling/readers/zarr_reader.py
@@ -0,0 +1,344 @@
+"""Zarr store reader for directory paths (CLI, library, Globus).
+
+Local browser upload of Zarr stores is intentionally not registered in
+``SUPPORTED_FILE_TYPES``; use a directory path with CLI/library or Globus.
+"""
+
+from __future__ import annotations
+
+from typing import Any
+
+import numpy as np
+import pandas as pd
+
+from aidrin.file_handling.readers.structured import (
+    INVENTORY_EMPTY,
+    INVENTORY_LEGACY,
+    INVENTORY_MULTI,
+    INVENTORY_SINGLE,
+    DatasetEntry,
+    InventoryResult,
+    StructuredFileReader,
+    make_inventory,
+)
+
+
+def _require_zarr():
+    try:
+        import zarr
+    except ImportError as exc:
+        raise ImportError(
+            "Zarr support requires the 'zarr' package. "
+            "Install with: pip install 'aidrin[zarr]' or pip install zarr"
+        ) from exc
+    return zarr
+
+
+class zarrReader(StructuredFileReader):
+    """Read Zarr directory stores into pandas DataFrames."""
+
+    def __init__(self, file_path: str, logger, selected_keys=None):
+        super().__init__(file_path, logger)
+        self._explicit_selected_keys = selected_keys
+
+    def _open_store(self):
+        zarr = _require_zarr()
+        return zarr.open(self.file_path, mode="r")
+
+    def _list_group_paths(self):
+        zarr = _require_zarr()
+        root = self._open_store()
+        if isinstance(root, zarr.Array):
+            return []
+
+        groups = []
+
+        def walk(group, prefix=""):
+            for key in group.keys():
+                child = group[key]
+                path = f"{prefix}/{key}" if prefix else key
+                if isinstance(child, zarr.Group):
+                    groups.append(path)
+                    walk(child, path)
+
+        walk(root)
+        return groups
+
+    def _list_datasets(self) -> list[DatasetEntry]:
+        zarr = _require_zarr()
+        root = self._open_store()
+        datasets: list[DatasetEntry] = []
+
+        if isinstance(root, zarr.Array):
+            return [
+                {
+                    "path": "",
+                    "shape": tuple(int(s) for s in root.shape),
+                    "ndim": int(root.ndim),
+                    "dtype": str(root.dtype),
+                    "size": int(root.size),
+                }
+            ]
+
+        def walk(group, prefix=""):
+            for key in group.keys():
+                child = group[key]
+                path = f"{prefix}/{key}" if prefix else key
+                if isinstance(child, zarr.Array):
+                    datasets.append(
+                        {
+                            "path": path,
+                            "shape": tuple(int(s) for s in child.shape),
+                            "ndim": int(child.ndim),
+                            "dtype": str(child.dtype),
+                            "size": int(child.size),
+                        }
+                    )
+                elif isinstance(child, zarr.Group):
+                    walk(child, path)
+
+        walk(root)
+        return datasets
+
+    def _build_picker_groups(self, datasets: list[DatasetEntry]):
+        paths = {ds["path"] for ds in datasets if ds["path"]}
+        assigned = set()
+        groups = {}
+
+        for group_path in sorted(self._list_group_paths(), key=len, reverse=True):
+            prefix = f"{group_path}/"
+            members = sorted(p for p in paths if p.startswith(prefix) and p not in assigned)
+            if len(members) >= 2:
+                groups[group_path] = {
+                    "id": group_path,
+                    "label": group_path,
+                    "type": "zarr_group",
+                    "dataset_paths": members,
+                }
+                assigned.update(members)
+
+        return sorted(groups.values(), key=lambda group: group["label"].lower())
+
+    def _is_incompatible_root_layout(self, datasets: list[DatasetEntry]):
+        root = [ds for ds in datasets if "/" not in ds["path"] and ds["path"] != ""]
+        # Root-store single array uses path ""
+        if any(ds["path"] == "" for ds in datasets):
+            return False
+        if len(root) < 2:
+            return False
+        if not all(ds["ndim"] == 1 for ds in root):
+            return False
+        lengths = {ds["shape"][0] for ds in root if ds["shape"]}
+        return len(lengths) > 1
+
+    def _is_grouped_hierarchical_layout(self, datasets: list[DatasetEntry]):
+        nested = [ds for ds in datasets if "/" in ds["path"]]
+        if len(nested) < 4:
+            return False
+
+        groups = self._build_picker_groups(datasets)
+        if len(groups) < 2:
+            return False
+
+        lengths = set()
+        for ds in nested:
+            if ds["ndim"] == 1 and ds["shape"]:
+                lengths.add(ds["shape"][0])
+        return len(lengths) > 1
+
+    def _needs_dataset_selection(self, datasets: list[DatasetEntry]):
+        return self._is_incompatible_root_layout(datasets) or self._is_grouped_hierarchical_layout(
+            datasets
+        )
+
+    def inventory(self) -> InventoryResult:
+        try:
+            datasets = self._list_datasets()
+        except Exception as e:
+            self.logger.error("Failed to inventory Zarr store: %s", e, exc_info=True)
+            return make_inventory(INVENTORY_EMPTY)
+
+        if not datasets:
+            layout = INVENTORY_EMPTY
+        elif len(datasets) == 1:
+            layout = INVENTORY_SINGLE
+        elif self._needs_dataset_selection(datasets):
+            layout = INVENTORY_MULTI
+        else:
+            layout = INVENTORY_LEGACY
+
+        groups = self._build_picker_groups(datasets) if layout == INVENTORY_MULTI else []
+        return make_inventory(layout, datasets, groups)
+
+    def parse(self):
+        return [ds["path"] or "(root)" for ds in self._list_datasets()]
+
+    def get_metadata(self) -> dict[str, Any]:
+        zarr = _require_zarr()
+        root = self._open_store()
+        metadata: dict[str, Any] = {}
+
+        def attrs_to_dict(attrs):
+            try:
+                return dict(attrs)
+            except Exception:
+                return {}
+
+        if isinstance(root, zarr.Array):
+            metadata["(root)"] = attrs_to_dict(root.attrs)
+            return metadata
+
+        metadata["(root)"] = attrs_to_dict(root.attrs)
+
+        def walk(group, prefix=""):
+            for key in group.keys():
+                child = group[key]
+                path = f"{prefix}/{key}" if prefix else key
+                child_attrs = attrs_to_dict(child.attrs)
+                if child_attrs:
+                    metadata[path] = child_attrs
+                if isinstance(child, zarr.Group):
+                    walk(child, path)
+
+        walk(root)
+        return metadata
+
+    def _normalize_selected_keys(self, keys):
+        if isinstance(keys, str):
+            keys = [key.strip() for key in keys.split(",") if key.strip()]
+        return [str(key) for key in keys if key]
+
+    def _get_selected_dataset_keys(self):
+        if self._explicit_selected_keys is not None:
+            return self._normalize_selected_keys(self._explicit_selected_keys)
+        try:
+            from flask import session
+
+            keys = session.get("selected_keys") or []
+            return self._normalize_selected_keys(keys)
+        except RuntimeError:
+            return []
+
+    def _resolve_array(self, root, path: str):
+        zarr = _require_zarr()
+        if path in ("", "(root)"):
+            if isinstance(root, zarr.Array):
+                return root
+            self.logger.warning("Zarr root path requested but store is a group")
+            return None
+        try:
+            obj = root[path]
+        except Exception:
+            self.logger.warning("Zarr array path not found: %s", path)
+            return None
+        if not isinstance(obj, zarr.Array):
+            self.logger.warning("Zarr path is not an array: %s", path)
+            return None
+        return obj
+
+    def _array_to_frame(self, path: str, arr):
+        data = arr[:]
+        col_name = path.split("/")[-1] if path not in ("", "(root)") else "value"
+        if getattr(data, "ndim", 0) == 0:
+            df = pd.DataFrame({col_name: [data]})
+        elif data.ndim == 1:
+            df = pd.DataFrame({col_name: data})
+        else:
+            try:
+                df = pd.DataFrame(data)
+            except Exception:
+                df = pd.DataFrame(data.tolist())
+            df.columns = [str(col) for col in df.columns]
+        df.columns = [str(col) for col in df.columns]
+        return df if not df.empty else None
+
+    def _read_array_path(self, path: str):
+        root = self._open_store()
+        arr = self._resolve_array(root, path)
+        if arr is None:
+            return None
+        return self._array_to_frame(path, arr)
+
+    def _read_compatible_array_paths(self, paths):
+        if not paths:
+            return None
+
+        columns = {}
+        expected_len = None
+        root = self._open_store()
+
+        for path in paths:
+            arr = self._resolve_array(root, path)
+            if arr is None:
+                return None
+            data = np.asarray(arr[:])
+            if data.ndim != 1:
+                self.logger.warning(
+                    "Zarr multi-select requires 1D arrays; '%s' has ndim=%s",
+                    path,
+                    data.ndim,
+                )
+                return None
+            length = int(data.shape[0])
+            if expected_len is None:
+                expected_len = length
+            elif length != expected_len:
+                self.logger.warning(
+                    "Zarr selected arrays have incompatible lengths (%d vs %d)",
+                    expected_len,
+                    length,
+                )
+                return None
+            short = path.split("/")[-1] or path
+            name = short if short not in columns else path.replace("/", ".")
+            columns[name] = data
+
+        df = pd.DataFrame(columns)
+        return df if not df.empty else None
+
+    def _auto_read_all(self, datasets: list[DatasetEntry]):
+        """Auto-read when layout does not require explicit selection."""
+        if len(datasets) == 1:
+            return self._read_array_path(datasets[0]["path"] or "")
+
+        ones = [ds for ds in datasets if ds["ndim"] == 1]
+        if ones and len(ones) == len(datasets):
+            lengths = {ds["shape"][0] for ds in ones if ds["shape"]}
+            if len(lengths) == 1:
+                return self._read_compatible_array_paths([ds["path"] for ds in ones])
+
+        self.logger.warning(
+            "Zarr store has %d arrays that cannot be auto-merged; "
+            "pass selected_keys to read specific paths.",
+            len(datasets),
+        )
+        return None
+
+    def read(self):
+        try:
+            inv = self.inventory()
+            if inv["type"] == INVENTORY_EMPTY:
+                self.logger.warning("No arrays found in Zarr store")
+                return None
+
+            # Honor explicit selection for any non-empty layout (CLI/library/Globus).
+            selected = self._get_selected_dataset_keys()
+            if selected:
+                if len(selected) == 1:
+                    return self._read_array_path(selected[0])
+                return self._read_compatible_array_paths(selected)
+
+            if inv["type"] == INVENTORY_MULTI:
+                self.logger.warning(
+                    "Zarr store has %d arrays in an incompatible layout; "
+                    "refusing to flatten. Pass selected_keys to choose paths.",
+                    len(inv["datasets"]),
+                )
+                return None
+
+            return self._auto_read_all(inv["datasets"])
+        except ImportError:
+            raise
+        except Exception as e:
+            self.logger.error("Error reading Zarr store: %s", e, exc_info=True)
+            return None
diff --git a/pyproject.toml b/pyproject.toml
index 9e39ed13..edfe35fe 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -49,6 +49,9 @@ globus = [
     "globus-compute-sdk>=2.3.0",
     "globus-sdk>=3.20.0",
 ]
+zarr = [
+    "zarr>=3.0",
+]
 telemetry = [
     "opentelemetry-api>=1.20",
     "opentelemetry-sdk>=1.20",
@@ -126,6 +129,9 @@ version = {attr = "aidrin._version.__version__"}
 # ---------------------------------------------------------------------------
 
 [dependency-groups]
+zarr = [
+    "zarr>=3.0",
+]
 dev = [
     "pytest>=7.0",
     "pytest-flask",
diff --git a/tests/unit/test_structured_readers.py b/tests/unit/test_structured_readers.py
new file mode 100644
index 00000000..d4cab412
--- /dev/null
+++ b/tests/unit/test_structured_readers.py
@@ -0,0 +1,187 @@
+"""Unit tests for structured reader scaffold and Zarr reader."""
+
+import logging
+import sys
+import types
+
+import numpy as np
+import pytest
+
+# Compatibility shim: pkg_resources for clean Python 3.12+ environments.
+if "pkg_resources" not in sys.modules:
+    _pkg_resources = types.ModuleType("pkg_resources")
+
+    class _Dist:
+        def __init__(self):
+            self.version = "0.0.0"
+
+    _pkg_resources.get_distribution = lambda _name: _Dist()
+    sys.modules["pkg_resources"] = _pkg_resources
+
+from aidrin.file_handling.file_parser import (
+    GLOBUS_FILE_TYPES,
+    READER_MAP,
+    SUPPORTED_FILE_TYPES,
+    read_file,
+)
+from aidrin.file_handling.readers.root_reader import rootReader
+from aidrin.file_handling.readers.structured import (
+    INVENTORY_EMPTY,
+    INVENTORY_LEGACY,
+    INVENTORY_MULTI,
+    INVENTORY_SINGLE,
+    INVENTORY_UNSUPPORTED,
+    USER_FACING_INVENTORY_TYPES,
+    make_inventory,
+)
+from aidrin.file_handling.readers.zarr_reader import zarrReader
+
+zarr = pytest.importorskip("zarr")
+
+
+@pytest.fixture
+def logger():
+    return logging.getLogger("test_structured_readers")
+
+
+@pytest.mark.parametrize("reader_cls", [rootReader])
+def test_root_stub_inventory_contract(tmp_path, logger, reader_cls):
+    path = tmp_path / "placeholder"
+    path.write_text("not used", encoding="utf-8")
+    inv = reader_cls(str(path), logger).inventory()
+
+    assert set(inv.keys()) == {"type", "datasets", "groups"}
+    assert inv["type"] == INVENTORY_UNSUPPORTED
+    assert inv["datasets"] == []
+    assert inv["groups"] == []
+    assert inv["type"] not in USER_FACING_INVENTORY_TYPES
+
+
+def test_make_inventory_helper():
+    inv = make_inventory(INVENTORY_UNSUPPORTED)
+    assert inv == {"type": INVENTORY_UNSUPPORTED, "datasets": [], "groups": []}
+
+
+def test_zarr_registered_for_cli_not_local_upload():
+    assert ".zarr" in READER_MAP
+    local_exts = [ext for ext, _ in SUPPORTED_FILE_TYPES]
+    globus_exts = [ext for ext, _ in GLOBUS_FILE_TYPES]
+    assert ".zarr" not in local_exts
+    assert ".zarr" in globus_exts
+
+
+def _write_single_array_store(path):
+    arr = zarr.open(str(path), mode="w", shape=(5,), dtype="f8")
+    arr[:] = np.arange(5, dtype=np.float64)
+    arr.attrs["unit"] = "m"
+
+
+def _write_group_store(path):
+    root = zarr.open_group(str(path), mode="w")
+    root.attrs["license"] = "MIT"
+    a = root.create_array("temp", shape=(4,), dtype="f8")
+    a[:] = np.arange(4, dtype=np.float64)
+    a.attrs["unit"] = "C"
+    g = root.create_group("station")
+    b = g.create_array("x", shape=(4,), dtype="i4")
+    b[:] = np.arange(4, dtype=np.int32)
+
+
+def _write_incompatible_root_store(path):
+    root = zarr.open_group(str(path), mode="w")
+    a = root.create_array("short", shape=(3,), dtype="f8")
+    a[:] = [1.0, 2.0, 3.0]
+    b = root.create_array("long", shape=(5,), dtype="f8")
+    b[:] = np.arange(5, dtype=np.float64)
+
+
+def _write_grouped_hierarchical_store(path):
+    root = zarr.open_group(str(path), mode="w")
+    for station in ("S1", "S2"):
+        g = root.create_group(station)
+        for axis, length in (("X", 10), ("Y", 10), ("meta", 1)):
+            arr = g.create_array(axis, shape=(length,), dtype="f8")
+            arr[:] = np.arange(length, dtype=np.float64)
+
+
+def test_zarr_single_array_inventory_and_read(tmp_path, logger):
+    store = tmp_path / "single.zarr"
+    _write_single_array_store(store)
+    reader = zarrReader(str(store), logger)
+    inv = reader.inventory()
+    assert inv["type"] == INVENTORY_SINGLE
+    assert len(inv["datasets"]) == 1
+
+    df = reader.read()
+    assert df is not None
+    assert list(df.columns) == ["value"]
+    assert len(df) == 5
+    assert list(df["value"]) == [0.0, 1.0, 2.0, 3.0, 4.0]
+
+
+def test_zarr_compatible_group_auto_read(tmp_path, logger):
+    store = tmp_path / "group.zarr"
+    _write_group_store(store)
+    reader = zarrReader(str(store), logger)
+    inv = reader.inventory()
+    assert inv["type"] == INVENTORY_LEGACY
+    assert inv["type"] not in USER_FACING_INVENTORY_TYPES
+
+    df = reader.read()
+    assert df is not None
+    assert set(df.columns) == {"temp", "x"}
+    assert len(df) == 4
+
+
+def test_zarr_incompatible_root_needs_selection(tmp_path, logger):
+    store = tmp_path / "bad.zarr"
+    _write_incompatible_root_store(store)
+    reader = zarrReader(str(store), logger)
+    inv = reader.inventory()
+    assert inv["type"] == INVENTORY_MULTI
+    assert reader.read() is None
+
+    df = zarrReader(str(store), logger, selected_keys=["short"]).read()
+    assert df is not None
+    assert list(df.columns) == ["short"]
+    assert len(df) == 3
+
+
+def test_zarr_grouped_hierarchical_selection(tmp_path, logger):
+    store = tmp_path / "stations.zarr"
+    _write_grouped_hierarchical_store(store)
+    reader = zarrReader(str(store), logger)
+    inv = reader.inventory()
+    assert inv["type"] == INVENTORY_MULTI
+    assert len(inv["groups"]) >= 2
+
+    df = zarrReader(
+        str(store), logger, selected_keys=["S1/X", "S1/Y"]
+    ).read()
+    assert df is not None
+    assert set(df.columns) == {"X", "Y"}
+    assert len(df) == 10
+
+
+def test_zarr_get_metadata(tmp_path, logger):
+    store = tmp_path / "meta.zarr"
+    _write_group_store(store)
+    meta = zarrReader(str(store), logger).get_metadata()
+    assert meta["(root)"]["license"] == "MIT"
+    assert meta["temp"]["unit"] == "C"
+
+
+def test_zarr_read_file_integration(tmp_path):
+    store = tmp_path / "cli.zarr"
+    _write_single_array_store(store)
+    df = read_file((str(store), "cli.zarr", ".zarr"))
+    assert df is not None
+    assert len(df) == 5
+
+
+def test_zarr_empty_store(tmp_path, logger):
+    store = tmp_path / "empty.zarr"
+    zarr.open_group(str(store), mode="w")
+    inv = zarrReader(str(store), logger).inventory()
+    assert inv["type"] == INVENTORY_EMPTY
+    assert zarrReader(str(store), logger).read() is None
diff --git a/web/routes/core.py b/web/routes/core.py
index 4b5ba87d..37b5eec8 100644
--- a/web/routes/core.py
+++ b/web/routes/core.py
@@ -16,7 +16,11 @@
     url_for,
 )
 from werkzeug.utils import secure_filename
-from aidrin.file_handling.file_parser import SUPPORTED_FILE_TYPES, READER_MAP
+from aidrin.file_handling.file_parser import (
+    GLOBUS_FILE_TYPES,
+    READER_MAP,
+    SUPPORTED_FILE_TYPES,
+)
 from aidrin.file_handling.readers.hdf5_reader import hdf5Reader
 from web.routes.utils import (
     clear_all_user_cache,
@@ -144,6 +148,7 @@ def inspector():
             uploaded_file_name=effective_file_name or "",
             file_type=effective_file_type or "",
             supported_file_types=SUPPORTED_FILE_TYPES,
+            globus_file_types=GLOBUS_FILE_TYPES,
             file_preview=file_preview if file_preview is not None else [],
             current_checked_keys=current_checked_keys
             if current_checked_keys is not None
diff --git a/web/templates/_components/globus_panel.html b/web/templates/_components/globus_panel.html
index 865e41e3..038b8caa 100644
--- a/web/templates/_components/globus_panel.html
+++ b/web/templates/_components/globus_panel.html
@@ -38,7 +38,7 @@ <h3 class="text-base font-semibold text-gray-900 dark:text-white mb-2">Connect t
       <!-- Remote file path -->
       <div class="mb-4">
         <label class="block text-sm font-medium mb-1.5 text-gray-900 dark:text-white">File path on remote endpoint</label>
-        <input type="text" id="globus-file-path" placeholder="e.g., /data/datasets/adult.csv"
+        <input type="text" id="globus-file-path" placeholder="e.g., /data/datasets/adult.csv or /data/store.zarr"
                class="w-full px-3 py-2 text-sm border border-gray-300 rounded-lg bg-gray-50 dark:bg-gray-700 dark:border-gray-600 dark:text-white focus:ring-blue-500 focus:border-blue-500"
                style="max-width: none;" />
       </div>
@@ -47,7 +47,7 @@ <h3 class="text-base font-semibold text-gray-900 dark:text-white mb-2">Connect t
       <div class="mb-4">
         <label class="block text-sm font-medium mb-1.5 text-gray-900 dark:text-white">File type</label>
         <select id="globus-file-type" class="w-full" style="max-width: none;">
-          {% for ext, label in supported_file_types %}
+          {% for ext, label in globus_file_types %}
             <option value="{{ ext }}">{{ label }} ({{ ext }})</option>
           {% endfor %}
         </select>

From a8206072820e7c12c324d9ee99d990fa82132e5d Mon Sep 17 00:00:00 2001
From: Dhanushkumar Jv <kumardhanush509@gmail.com>
Date: Tue, 21 Jul 2026 23:43:14 +0530
Subject: [PATCH 2/5] Fix Zarr optional extra for uv lock and Python 3.10.

---
 aidrin/file_handling/readers/zarr_reader.py |   4 +-
 pyproject.toml                              |   5 +-
 uv.lock                                     | 202 +++++++++++++++-----
 3 files changed, 158 insertions(+), 53 deletions(-)

diff --git a/aidrin/file_handling/readers/zarr_reader.py b/aidrin/file_handling/readers/zarr_reader.py
index e7edf947..228209f5 100644
--- a/aidrin/file_handling/readers/zarr_reader.py
+++ b/aidrin/file_handling/readers/zarr_reader.py
@@ -28,8 +28,8 @@ def _require_zarr():
         import zarr
     except ImportError as exc:
         raise ImportError(
-            "Zarr support requires the 'zarr' package. "
-            "Install with: pip install 'aidrin[zarr]' or pip install zarr"
+            "Zarr support requires the 'zarr' package on Python >=3.11. "
+            "Install with: pip install 'aidrin[zarr]' or pip install 'zarr>=3.0.8,<3.2'"
         ) from exc
     return zarr
 
diff --git a/pyproject.toml b/pyproject.toml
index edfe35fe..77f4de25 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -49,8 +49,9 @@ globus = [
     "globus-compute-sdk>=2.3.0",
     "globus-sdk>=3.20.0",
 ]
+# Zarr 3.x needs Python >=3.11 (3.2+ needs >=3.12). Pin <3.2 so 3.11 stays resolvable.
 zarr = [
-    "zarr>=3.0",
+    "zarr>=3.0.8,<3.2; python_version >= '3.11'",
 ]
 telemetry = [
     "opentelemetry-api>=1.20",
@@ -130,7 +131,7 @@ version = {attr = "aidrin._version.__version__"}
 
 [dependency-groups]
 zarr = [
-    "zarr>=3.0",
+    "zarr>=3.0.8,<3.2; python_version >= '3.11'",
 ]
 dev = [
     "pytest>=7.0",
diff --git a/uv.lock b/uv.lock
index 0025d3ba..8845b871 100644
--- a/uv.lock
+++ b/uv.lock
@@ -104,6 +104,9 @@ telemetry = [
     { name = "opentelemetry-instrumentation-flask" },
     { name = "opentelemetry-sdk" },
 ]
+zarr = [
+    { name = "zarr", marker = "python_full_version >= '3.11'" },
+]
 
 [package.dev-dependencies]
 agentic = [
@@ -144,6 +147,9 @@ mcp = [
     { name = "pymupdf" },
     { name = "python-dotenv" },
 ]
+zarr = [
+    { name = "zarr", marker = "python_full_version >= '3.11'" },
+]
 
 [package.metadata]
 requires-dist = [
@@ -201,8 +207,9 @@ requires-dist = [
     { name = "sphinx", marker = "extra == 'docs'", specifier = "==6.2.1" },
     { name = "sphinx-rtd-theme", marker = "extra == 'docs'", specifier = "==1.2.2" },
     { name = "sphinxemoji", marker = "extra == 'docs'" },
+    { name = "zarr", marker = "python_full_version >= '3.11' and extra == 'zarr'", specifier = ">=3.0.8,<3.2" },
 ]
-provides-extras = ["globus", "telemetry", "llm", "mcp", "agentic", "dev", "docs"]
+provides-extras = ["globus", "zarr", "telemetry", "llm", "mcp", "agentic", "dev", "docs"]
 
 [package.metadata.requires-dev]
 agentic = [
@@ -241,6 +248,7 @@ mcp = [
     { name = "pymupdf", specifier = ">=1.24.0" },
     { name = "python-dotenv", specifier = ">=1.0.0" },
 ]
+zarr = [{ name = "zarr", marker = "python_full_version >= '3.11'", specifier = ">=3.0.8,<3.2" }]
 
 [[package]]
 name = "aiohappyeyeballs"
@@ -813,7 +821,7 @@ resolution-markers = [
     "python_full_version < '3.11'",
 ]
 dependencies = [
-    { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" },
+    { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" } },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/66/54/eb9bfc647b19f2009dd5c7f5ec51c4e6ca831725f1aea7a993034f483147/contourpy-1.3.2.tar.gz", hash = "sha256:b6945942715a034c671b7fc54f9588126b0b8bf23db2696e3ca8328f3ff0ab54", size = 13466130, upload-time = "2025-04-15T17:47:53.79Z" }
 wheels = [
@@ -898,7 +906,7 @@ resolution-markers = [
     "(python_full_version == '3.11.*' and platform_machine != 'x86_64' and sys_platform == 'darwin') or (python_full_version == '3.11.*' and sys_platform != 'darwin' and sys_platform != 'emscripten' and sys_platform != 'win32')",
 ]
 dependencies = [
-    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11'" },
+    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" } },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/58/01/1253e6698a07380cd31a736d248a3f2a50a7c88779a1813da27503cadc2a/contourpy-1.3.3.tar.gz", hash = "sha256:083e12155b210502d0bca491432bb04d56dc3432f95a979b429f2848c3dbe880", size = 13466174, upload-time = "2025-07-26T12:03:12.549Z" }
 wheels = [
@@ -1189,6 +1197,18 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/8d/14/69b4bad34e3f250afe29a854da03acb6747711f3df06c359fa053fae4e76/docutils-0.18.1-py2.py3-none-any.whl", hash = "sha256:23010f129180089fbcd3bc08cfefccb3b890b0050e1ca00c867036e9d161b98c", size = 570050, upload-time = "2021-11-23T17:49:38.556Z" },
 ]
 
+[[package]]
+name = "donfig"
+version = "0.8.1.post1"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "pyyaml" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/25/71/80cc718ff6d7abfbabacb1f57aaa42e9c1552bfdd01e64ddd704e4a03638/donfig-0.8.1.post1.tar.gz", hash = "sha256:3bef3413a4c1c601b585e8d297256d0c1470ea012afa6e8461dc28bfb7c23f52", size = 19506, upload-time = "2024-05-23T14:14:31.513Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/0c/d5/c5db1ea3394c6e1732fb3286b3bd878b59507a8f77d32a2cebda7d7b7cd4/donfig-0.8.1.post1-py3-none-any.whl", hash = "sha256:2a3175ce74a06109ff9307d90a230f81215cbac9a751f4d1c6194644b8204f9d", size = 21592, upload-time = "2024-05-23T14:13:55.283Z" },
+]
+
 [[package]]
 name = "dython"
 version = "0.7.12"
@@ -1527,6 +1547,41 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/1a/f0/67d4b279d5a19324792e29499856160d3a478e864cfec3919d23ebc88268/globus_sdk-4.7.0-py3-none-any.whl", hash = "sha256:6f2a15cff130c93ca70ddc25a8156ae636865850d9a9c9dbb7ffc365e70930e2", size = 439273, upload-time = "2026-05-20T16:20:27.683Z" },
 ]
 
+[[package]]
+name = "google-crc32c"
+version = "1.8.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/03/41/4b9c02f99e4c5fb477122cd5437403b552873f014616ac1d19ac8221a58d/google_crc32c-1.8.0.tar.gz", hash = "sha256:a428e25fb7691024de47fecfbff7ff957214da51eddded0da0ae0e0f03a2cf79", size = 14192, upload-time = "2025-12-16T00:35:25.142Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/95/ac/6f7bc93886a823ab545948c2dd48143027b2355ad1944c7cf852b338dc91/google_crc32c-1.8.0-cp310-cp310-macosx_12_0_arm64.whl", hash = "sha256:0470b8c3d73b5f4e3300165498e4cf25221c7eb37f1159e221d1825b6df8a7ff", size = 31296, upload-time = "2025-12-16T00:19:07.261Z" },
+    { url = "https://files.pythonhosted.org/packages/f7/97/a5accde175dee985311d949cfcb1249dcbb290f5ec83c994ea733311948f/google_crc32c-1.8.0-cp310-cp310-macosx_12_0_x86_64.whl", hash = "sha256:119fcd90c57c89f30040b47c211acee231b25a45d225e3225294386f5d258288", size = 30870, upload-time = "2025-12-16T00:29:17.669Z" },
+    { url = "https://files.pythonhosted.org/packages/3d/63/bec827e70b7a0d4094e7476f863c0dbd6b5f0f1f91d9c9b32b76dcdfeb4e/google_crc32c-1.8.0-cp310-cp310-manylinux1_x86_64.manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_5_x86_64.whl", hash = "sha256:6f35aaffc8ccd81ba3162443fabb920e65b1f20ab1952a31b13173a67811467d", size = 33214, upload-time = "2025-12-16T00:40:19.618Z" },
+    { url = "https://files.pythonhosted.org/packages/63/bc/11b70614df04c289128d782efc084b9035ef8466b3d0a8757c1b6f5cf7ac/google_crc32c-1.8.0-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:864abafe7d6e2c4c66395c1eb0fe12dc891879769b52a3d56499612ca93b6092", size = 33589, upload-time = "2025-12-16T00:40:20.7Z" },
+    { url = "https://files.pythonhosted.org/packages/3e/00/a08a4bc24f1261cc5b0f47312d8aebfbe4b53c2e6307f1b595605eed246b/google_crc32c-1.8.0-cp310-cp310-win_amd64.whl", hash = "sha256:db3fe8eaf0612fc8b20fa21a5f25bd785bc3cd5be69f8f3412b0ac2ffd49e733", size = 34437, upload-time = "2025-12-16T00:35:19.437Z" },
+    { url = "https://files.pythonhosted.org/packages/5d/ef/21ccfaab3d5078d41efe8612e0ed0bfc9ce22475de074162a91a25f7980d/google_crc32c-1.8.0-cp311-cp311-macosx_12_0_arm64.whl", hash = "sha256:014a7e68d623e9a4222d663931febc3033c5c7c9730785727de2a81f87d5bab8", size = 31298, upload-time = "2025-12-16T00:20:32.241Z" },
+    { url = "https://files.pythonhosted.org/packages/c5/b8/f8413d3f4b676136e965e764ceedec904fe38ae8de0cdc52a12d8eb1096e/google_crc32c-1.8.0-cp311-cp311-macosx_12_0_x86_64.whl", hash = "sha256:86cfc00fe45a0ac7359e5214a1704e51a99e757d0272554874f419f79838c5f7", size = 30872, upload-time = "2025-12-16T00:33:58.785Z" },
+    { url = "https://files.pythonhosted.org/packages/f6/fd/33aa4ec62b290477181c55bb1c9302c9698c58c0ce9a6ab4874abc8b0d60/google_crc32c-1.8.0-cp311-cp311-manylinux1_x86_64.manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_5_x86_64.whl", hash = "sha256:19b40d637a54cb71e0829179f6cb41835f0fbd9e8eb60552152a8b52c36cbe15", size = 33243, upload-time = "2025-12-16T00:40:21.46Z" },
+    { url = "https://files.pythonhosted.org/packages/71/03/4820b3bd99c9653d1a5210cb32f9ba4da9681619b4d35b6a052432df4773/google_crc32c-1.8.0-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:17446feb05abddc187e5441a45971b8394ea4c1b6efd88ab0af393fd9e0a156a", size = 33608, upload-time = "2025-12-16T00:40:22.204Z" },
+    { url = "https://files.pythonhosted.org/packages/7c/43/acf61476a11437bf9733fb2f70599b1ced11ec7ed9ea760fdd9a77d0c619/google_crc32c-1.8.0-cp311-cp311-win_amd64.whl", hash = "sha256:71734788a88f551fbd6a97be9668a0020698e07b2bf5b3aa26a36c10cdfb27b2", size = 34439, upload-time = "2025-12-16T00:35:20.458Z" },
+    { url = "https://files.pythonhosted.org/packages/e9/5f/7307325b1198b59324c0fa9807cafb551afb65e831699f2ce211ad5c8240/google_crc32c-1.8.0-cp312-cp312-macosx_12_0_arm64.whl", hash = "sha256:4b8286b659c1335172e39563ab0a768b8015e88e08329fa5321f774275fc3113", size = 31300, upload-time = "2025-12-16T00:21:56.723Z" },
+    { url = "https://files.pythonhosted.org/packages/21/8e/58c0d5d86e2220e6a37befe7e6a94dd2f6006044b1a33edf1ff6d9f7e319/google_crc32c-1.8.0-cp312-cp312-macosx_12_0_x86_64.whl", hash = "sha256:2a3dc3318507de089c5384cc74d54318401410f82aa65b2d9cdde9d297aca7cb", size = 30867, upload-time = "2025-12-16T00:38:31.302Z" },
+    { url = "https://files.pythonhosted.org/packages/ce/a9/a780cc66f86335a6019f557a8aaca8fbb970728f0efd2430d15ff1beae0e/google_crc32c-1.8.0-cp312-cp312-manylinux1_x86_64.manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_5_x86_64.whl", hash = "sha256:14f87e04d613dfa218d6135e81b78272c3b904e2a7053b841481b38a7d901411", size = 33364, upload-time = "2025-12-16T00:40:22.96Z" },
+    { url = "https://files.pythonhosted.org/packages/21/3f/3457ea803db0198c9aaca2dd373750972ce28a26f00544b6b85088811939/google_crc32c-1.8.0-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:cb5c869c2923d56cb0c8e6bcdd73c009c36ae39b652dbe46a05eb4ef0ad01454", size = 33740, upload-time = "2025-12-16T00:40:23.96Z" },
+    { url = "https://files.pythonhosted.org/packages/df/c0/87c2073e0c72515bb8733d4eef7b21548e8d189f094b5dad20b0ecaf64f6/google_crc32c-1.8.0-cp312-cp312-win_amd64.whl", hash = "sha256:3cc0c8912038065eafa603b238abf252e204accab2a704c63b9e14837a854962", size = 34437, upload-time = "2025-12-16T00:35:21.395Z" },
+    { url = "https://files.pythonhosted.org/packages/d1/db/000f15b41724589b0e7bc24bc7a8967898d8d3bc8caf64c513d91ef1f6c0/google_crc32c-1.8.0-cp313-cp313-macosx_12_0_arm64.whl", hash = "sha256:3ebb04528e83b2634857f43f9bb8ef5b2bbe7f10f140daeb01b58f972d04736b", size = 31297, upload-time = "2025-12-16T00:23:20.709Z" },
+    { url = "https://files.pythonhosted.org/packages/d7/0d/8ebed0c39c53a7e838e2a486da8abb0e52de135f1b376ae2f0b160eb4c1a/google_crc32c-1.8.0-cp313-cp313-macosx_12_0_x86_64.whl", hash = "sha256:450dc98429d3e33ed2926fc99ee81001928d63460f8538f21a5d6060912a8e27", size = 30867, upload-time = "2025-12-16T00:43:14.628Z" },
+    { url = "https://files.pythonhosted.org/packages/ce/42/b468aec74a0354b34c8cbf748db20d6e350a68a2b0912e128cabee49806c/google_crc32c-1.8.0-cp313-cp313-manylinux1_x86_64.manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_5_x86_64.whl", hash = "sha256:3b9776774b24ba76831609ffbabce8cdf6fa2bd5e9df37b594221c7e333a81fa", size = 33344, upload-time = "2025-12-16T00:40:24.742Z" },
+    { url = "https://files.pythonhosted.org/packages/1c/e8/b33784d6fc77fb5062a8a7854e43e1e618b87d5ddf610a88025e4de6226e/google_crc32c-1.8.0-cp313-cp313-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:89c17d53d75562edfff86679244830599ee0a48efc216200691de8b02ab6b2b8", size = 33694, upload-time = "2025-12-16T00:40:25.505Z" },
+    { url = "https://files.pythonhosted.org/packages/92/b1/d3cbd4d988afb3d8e4db94ca953df429ed6db7282ed0e700d25e6c7bfc8d/google_crc32c-1.8.0-cp313-cp313-win_amd64.whl", hash = "sha256:57a50a9035b75643996fbf224d6661e386c7162d1dfdab9bc4ca790947d1007f", size = 34435, upload-time = "2025-12-16T00:35:22.107Z" },
+    { url = "https://files.pythonhosted.org/packages/21/88/8ecf3c2b864a490b9e7010c84fd203ec8cf3b280651106a3a74dd1b0ca72/google_crc32c-1.8.0-cp314-cp314-macosx_12_0_arm64.whl", hash = "sha256:e6584b12cb06796d285d09e33f63309a09368b9d806a551d8036a4207ea43697", size = 31301, upload-time = "2025-12-16T00:24:48.527Z" },
+    { url = "https://files.pythonhosted.org/packages/36/c6/f7ff6c11f5ca215d9f43d3629163727a272eabc356e5c9b2853df2bfe965/google_crc32c-1.8.0-cp314-cp314-macosx_12_0_x86_64.whl", hash = "sha256:f4b51844ef67d6cf2e9425983274da75f18b1597bb2c998e1c0a0e8d46f8f651", size = 30868, upload-time = "2025-12-16T00:48:12.163Z" },
+    { url = "https://files.pythonhosted.org/packages/56/15/c25671c7aad70f8179d858c55a6ae8404902abe0cdcf32a29d581792b491/google_crc32c-1.8.0-cp314-cp314-manylinux1_x86_64.manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_5_x86_64.whl", hash = "sha256:b0d1a7afc6e8e4635564ba8aa5c0548e3173e41b6384d7711a9123165f582de2", size = 33381, upload-time = "2025-12-16T00:40:26.268Z" },
+    { url = "https://files.pythonhosted.org/packages/42/fa/f50f51260d7b0ef5d4898af122d8a7ec5a84e2984f676f746445f783705f/google_crc32c-1.8.0-cp314-cp314-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:8b3f68782f3cbd1bce027e48768293072813469af6a61a86f6bb4977a4380f21", size = 33734, upload-time = "2025-12-16T00:40:27.028Z" },
+    { url = "https://files.pythonhosted.org/packages/08/a5/7b059810934a09fb3ccb657e0843813c1fee1183d3bc2c8041800374aa2c/google_crc32c-1.8.0-cp314-cp314-win_amd64.whl", hash = "sha256:d511b3153e7011a27ab6ee6bb3a5404a55b994dc1a7322c0b87b29606d9790e2", size = 34878, upload-time = "2025-12-16T00:35:23.142Z" },
+    { url = "https://files.pythonhosted.org/packages/52/c5/c171e4d8c44fec1422d801a6d2e5d7ddabd733eeda505c79730ee9607f07/google_crc32c-1.8.0-pp311-pypy311_pp73-manylinux1_x86_64.manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_5_x86_64.whl", hash = "sha256:87fa445064e7db928226b2e6f0d5304ab4cd0339e664a4e9a25029f384d9bb93", size = 28615, upload-time = "2025-12-16T00:40:29.298Z" },
+    { url = "https://files.pythonhosted.org/packages/9c/97/7d75fe37a7a6ed171a2cf17117177e7aab7e6e0d115858741b41e9dd4254/google_crc32c-1.8.0-pp311-pypy311_pp73-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:f639065ea2042d5c034bf258a9f085eaa7af0cd250667c0635a3118e8f92c69c", size = 28800, upload-time = "2025-12-16T00:40:30.322Z" },
+]
+
 [[package]]
 name = "googleapis-common-protos"
 version = "1.75.0"
@@ -2748,6 +2803,38 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/32/36/8be7118ffd4c8440881046eac3d0982cc5ab42909508cf5d67024d62a2e4/numba-0.65.1-cp314-cp314t-win_amd64.whl", hash = "sha256:20609346e3bd75204950dcbbfe383a8d7dbf4902f442aedbf00f97fef4aa8f38", size = 2758237, upload-time = "2026-04-24T02:02:54.612Z" },
 ]
 
+[[package]]
+name = "numcodecs"
+version = "0.16.5"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" } },
+    { name = "typing-extensions" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/44/bd/8a391e7c356366224734efd24da929cc4796fff468bfb179fe1af6548535/numcodecs-0.16.5.tar.gz", hash = "sha256:0d0fb60852f84c0bd9543cc4d2ab9eefd37fc8efcc410acd4777e62a1d300318", size = 6276387, upload-time = "2025-11-21T02:49:48.986Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/af/85/1ac101a40ead81eaa1c7dc49a8827a30e2e436211b43ebdc63c590eb1347/numcodecs-0.16.5-cp311-cp311-macosx_10_13_x86_64.whl", hash = "sha256:78382dcea50622f2ef1e6e7a71dbe7f861d8fe376b27b7c297c26907304fef1e", size = 1621795, upload-time = "2025-11-21T02:49:17.418Z" },
+    { url = "https://files.pythonhosted.org/packages/0e/cc/0d97ef55dda48cb0f93d7b92d761208e7a99bd2eea6b0e859426e6a99a21/numcodecs-0.16.5-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:e2d04a19cb57a3c519b4127ac377cca6471aee1990d7c18f5b1e3a4fe1306689", size = 1153030, upload-time = "2025-11-21T02:49:19.089Z" },
+    { url = "https://files.pythonhosted.org/packages/5e/41/e120ee1b390730ac5987cde2afd82e2b8442cec315ab40b94b0373e93e73/numcodecs-0.16.5-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:c043af648eb280cd61785c99c22ff5c3c3460f906eb51a8511327c4f5111b283", size = 8510503, upload-time = "2025-11-21T02:49:20.324Z" },
+    { url = "https://files.pythonhosted.org/packages/54/4b/195ac84cc8f6077b4f0f421e8daee21b7f1bd88cb7716414234379fe68ec/numcodecs-0.16.5-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:c398919ef2eb0e56b8e97456f622640bfd3deed06de3acc976989cbcb22628a3", size = 9123428, upload-time = "2025-11-21T02:49:22.328Z" },
+    { url = "https://files.pythonhosted.org/packages/0f/5b/af02c417954f46e5c7bd5163ac251f535877d909fce54861c99ae197f6f6/numcodecs-0.16.5-cp311-cp311-win_amd64.whl", hash = "sha256:3820860ed302d4d84a1c66e70981ff959d5eb712555be4e7d8ced49888594773", size = 801542, upload-time = "2025-11-21T02:49:24.265Z" },
+    { url = "https://files.pythonhosted.org/packages/75/cc/55420f3641a67f78392dc0bc5d02cb9eb0a9dcebf2848d1ac77253ca61fa/numcodecs-0.16.5-cp312-cp312-macosx_10_13_x86_64.whl", hash = "sha256:24e675dc8d1550cd976a99479b87d872cb142632c75cc402fea04c08c4898523", size = 1656287, upload-time = "2025-11-21T02:49:25.755Z" },
+    { url = "https://files.pythonhosted.org/packages/f5/6c/86644987505dcb90ba6d627d6989c27bafb0699f9fd00187e06d05ea8594/numcodecs-0.16.5-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:94ddfa4341d1a3ab99989d13b01b5134abb687d3dab2ead54b450aefe4ad5bd6", size = 1148899, upload-time = "2025-11-21T02:49:26.87Z" },
+    { url = "https://files.pythonhosted.org/packages/97/1e/98aaddf272552d9fef1f0296a9939d1487914a239e98678f6b20f8b0a5c8/numcodecs-0.16.5-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:b554ab9ecf69de7ca2b6b5e8bc696bd9747559cb4dd5127bd08d7a28bec59c3a", size = 8534814, upload-time = "2025-11-21T02:49:28.547Z" },
+    { url = "https://files.pythonhosted.org/packages/fb/53/78c98ef5c8b2b784453487f3e4d6c017b20747c58b470393e230c78d18e8/numcodecs-0.16.5-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:ad1a379a45bd3491deab8ae6548313946744f868c21d5340116977ea3be5b1d6", size = 9173471, upload-time = "2025-11-21T02:49:30.444Z" },
+    { url = "https://files.pythonhosted.org/packages/1c/20/2fdec87fc7f8cec950d2b0bea603c12dc9f05b4966dc5924ba5a36a61bf6/numcodecs-0.16.5-cp312-cp312-win_amd64.whl", hash = "sha256:845a9857886ffe4a3172ba1c537ae5bcc01e65068c31cf1fce1a844bd1da050f", size = 801412, upload-time = "2025-11-21T02:49:32.123Z" },
+    { url = "https://files.pythonhosted.org/packages/38/38/071ced5a5fd1c85ba0e14ba721b66b053823e5176298c2f707e50bed11d9/numcodecs-0.16.5-cp313-cp313-macosx_10_13_x86_64.whl", hash = "sha256:25be3a516ab677dad890760d357cfe081a371d9c0a2e9a204562318ac5969de3", size = 1654359, upload-time = "2025-11-21T02:49:33.673Z" },
+    { url = "https://files.pythonhosted.org/packages/d1/c0/5f84ba7525577c1b9909fc2d06ef11314825fc4ad4378f61d0e4c9883b4a/numcodecs-0.16.5-cp313-cp313-macosx_11_0_arm64.whl", hash = "sha256:0107e839ef75b854e969cb577e140b1aadb9847893937636582d23a2a4c6ce50", size = 1144237, upload-time = "2025-11-21T02:49:35.294Z" },
+    { url = "https://files.pythonhosted.org/packages/0b/00/787ea5f237b8ea7bc67140c99155f9c00b5baf11c49afc5f3bfefa298f95/numcodecs-0.16.5-cp313-cp313-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:015a7c859ecc2a06e2a548f64008c0ec3aaecabc26456c2c62f4278d8fc20597", size = 8483064, upload-time = "2025-11-21T02:49:36.454Z" },
+    { url = "https://files.pythonhosted.org/packages/c4/e6/d359fdd37498e74d26a167f7a51e54542e642ea47181eb4e643a69a066c3/numcodecs-0.16.5-cp313-cp313-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:84230b4b9dad2392f2a84242bd6e3e659ac137b5a1ce3571d6965fca673e0903", size = 9126063, upload-time = "2025-11-21T02:49:38.018Z" },
+    { url = "https://files.pythonhosted.org/packages/27/72/6663cc0382ddbb866136c255c837bcb96cc7ce5e83562efec55e1b995941/numcodecs-0.16.5-cp313-cp313-win_amd64.whl", hash = "sha256:5088145502ad1ebf677ec47d00eb6f0fd600658217db3e0c070c321c85d6cf3d", size = 799275, upload-time = "2025-11-21T02:49:39.558Z" },
+    { url = "https://files.pythonhosted.org/packages/3c/9e/38e7ca8184c958b51f45d56a4aeceb1134ecde2d8bd157efadc98502cc42/numcodecs-0.16.5-cp314-cp314-macosx_10_13_x86_64.whl", hash = "sha256:b05647b8b769e6bc8016e9fd4843c823ce5c9f2337c089fb5c9c4da05e5275de", size = 1654721, upload-time = "2025-11-21T02:49:40.602Z" },
+    { url = "https://files.pythonhosted.org/packages/a1/37/260fa42e7b2b08e6e00ad632f8dd620961a60a459426c26cea390f8c68d0/numcodecs-0.16.5-cp314-cp314-macosx_11_0_arm64.whl", hash = "sha256:3832bd1b5af8bb3e413076b7d93318c8e7d7b68935006b9fa36ca057d1725a8f", size = 1146887, upload-time = "2025-11-21T02:49:41.721Z" },
+    { url = "https://files.pythonhosted.org/packages/4e/15/e2e1151b5a8b14a15dfd4bb4abccce7fff7580f39bc34092780088835f3a/numcodecs-0.16.5-cp314-cp314-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:49f7b7d24f103187f53135bed28bb9f0ed6b2e14c604664726487bb6d7c882e1", size = 8476987, upload-time = "2025-11-21T02:49:43.363Z" },
+    { url = "https://files.pythonhosted.org/packages/6d/30/16a57fc4d9fb0ba06c600408bd6634f2f1753c54a7a351c99c5e09b51ee2/numcodecs-0.16.5-cp314-cp314-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:aec9736d81b70f337d89c4070ee3ffeff113f386fd789492fa152d26a15043e4", size = 9102377, upload-time = "2025-11-21T02:49:45.508Z" },
+    { url = "https://files.pythonhosted.org/packages/31/a5/a0425af36c20d55a3ea884db4b4efca25a43bea9214ba69ca7932dd997b4/numcodecs-0.16.5-cp314-cp314-win_amd64.whl", hash = "sha256:b16a14303800e9fb88abc39463ab4706c037647ac17e49e297faa5f7d7dbbf1d", size = 819022, upload-time = "2025-11-21T02:49:47.39Z" },
+]
+
 [[package]]
 name = "numpy"
 version = "2.2.6"
@@ -3263,10 +3350,10 @@ resolution-markers = [
     "python_full_version < '3.11'",
 ]
 dependencies = [
-    { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" },
-    { name = "python-dateutil", marker = "python_full_version < '3.11'" },
-    { name = "pytz", marker = "python_full_version < '3.11'" },
-    { name = "tzdata", marker = "python_full_version < '3.11'" },
+    { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" } },
+    { name = "python-dateutil" },
+    { name = "pytz" },
+    { name = "tzdata" },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/33/01/d40b85317f86cf08d853a4f495195c73815fdf205eef3993821720274518/pandas-2.3.3.tar.gz", hash = "sha256:e05e1af93b977f7eafa636d043f9f94c7ee3ac81af99c13508215942e64c993b", size = 4495223, upload-time = "2025-09-29T23:34:51.853Z" }
 wheels = [
@@ -3342,9 +3429,9 @@ resolution-markers = [
     "(python_full_version == '3.11.*' and platform_machine != 'x86_64' and sys_platform == 'darwin') or (python_full_version == '3.11.*' and sys_platform != 'darwin' and sys_platform != 'emscripten' and sys_platform != 'win32')",
 ]
 dependencies = [
-    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11'" },
-    { name = "python-dateutil", marker = "python_full_version >= '3.11'" },
-    { name = "tzdata", marker = "(python_full_version >= '3.11' and sys_platform == 'emscripten') or (python_full_version >= '3.11' and sys_platform == 'win32')" },
+    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" } },
+    { name = "python-dateutil" },
+    { name = "tzdata", marker = "sys_platform == 'emscripten' or sys_platform == 'win32'" },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/f8/87/4341c6252d1c47b08768c3d25ac487362bf403f0313ddae4a2a26c9b1b4c/pandas-3.0.3.tar.gz", hash = "sha256:696a4a00a2a2a35d4e5deb3fc946641b96c944f02230e4f76137fe35d806c4fc", size = 4651414, upload-time = "2026-05-11T18:54:29.21Z" }
 wheels = [
@@ -4652,10 +4739,10 @@ resolution-markers = [
     "python_full_version < '3.11'",
 ]
 dependencies = [
-    { name = "joblib", marker = "python_full_version < '3.11'" },
-    { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" },
-    { name = "scipy", version = "1.15.3", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" },
-    { name = "threadpoolctl", marker = "python_full_version < '3.11'" },
+    { name = "joblib" },
+    { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" } },
+    { name = "scipy", version = "1.15.3", source = { registry = "https://pypi.org/simple" } },
+    { name = "threadpoolctl" },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/98/c2/a7855e41c9d285dfe86dc50b250978105dce513d6e459ea66a6aeb0e1e0c/scikit_learn-1.7.2.tar.gz", hash = "sha256:20e9e49ecd130598f1ca38a1d85090e1a600147b9c02fa6f15d69cb53d968fda", size = 7193136, upload-time = "2025-09-09T08:21:29.075Z" }
 wheels = [
@@ -4714,10 +4801,10 @@ resolution-markers = [
     "(python_full_version == '3.11.*' and platform_machine != 'x86_64' and sys_platform == 'darwin') or (python_full_version == '3.11.*' and sys_platform != 'darwin' and sys_platform != 'emscripten' and sys_platform != 'win32')",
 ]
 dependencies = [
-    { name = "joblib", marker = "python_full_version >= '3.11'" },
-    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11'" },
-    { name = "scipy", version = "1.17.1", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11'" },
-    { name = "threadpoolctl", marker = "python_full_version >= '3.11'" },
+    { name = "joblib" },
+    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" } },
+    { name = "scipy", version = "1.17.1", source = { registry = "https://pypi.org/simple" } },
+    { name = "threadpoolctl" },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/0e/d4/40988bf3b8e34feec1d0e6a051446b1f66225f8529b9309becaeef62b6c4/scikit_learn-1.8.0.tar.gz", hash = "sha256:9bccbb3b40e3de10351f8f5068e105d0f4083b1a65fa07b6634fbc401a6287fd", size = 7335585, upload-time = "2025-12-10T07:08:53.618Z" }
 wheels = [
@@ -4767,7 +4854,7 @@ resolution-markers = [
     "python_full_version < '3.11'",
 ]
 dependencies = [
-    { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" },
+    { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" } },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/0f/37/6964b830433e654ec7485e45a00fc9a27cf868d622838f6b6d9c5ec0d532/scipy-1.15.3.tar.gz", hash = "sha256:eae3cf522bc7df64b42cad3925c876e1b0b6c35c1337c93e12c0f366f55b0eaf", size = 59419214, upload-time = "2025-05-08T16:13:05.955Z" }
 wheels = [
@@ -4841,7 +4928,7 @@ resolution-markers = [
     "(python_full_version == '3.11.*' and platform_machine != 'x86_64' and sys_platform == 'darwin') or (python_full_version == '3.11.*' and sys_platform != 'darwin' and sys_platform != 'emscripten' and sys_platform != 'win32')",
 ]
 dependencies = [
-    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11'" },
+    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" } },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/7a/97/5a3609c4f8d58b039179648e62dd220f89864f56f7357f5d4f45c29eb2cc/scipy-1.17.1.tar.gz", hash = "sha256:95d8e012d8cb8816c226aef832200b1d45109ed4464303e997c5b13122b297c0", size = 30573822, upload-time = "2026-02-23T00:26:24.851Z" }
 wheels = [
@@ -4944,20 +5031,20 @@ resolution-markers = [
     "python_full_version < '3.11'",
 ]
 dependencies = [
-    { name = "cloudpickle", marker = "(python_full_version < '3.11' and platform_machine != 'x86_64') or (python_full_version < '3.11' and sys_platform != 'darwin') or (platform_machine == 'x86_64' and sys_platform == 'darwin')" },
-    { name = "numba", marker = "(python_full_version < '3.11' and platform_machine != 'x86_64') or (python_full_version < '3.11' and sys_platform != 'darwin') or (platform_machine == 'x86_64' and sys_platform == 'darwin')" },
+    { name = "cloudpickle" },
+    { name = "numba" },
     { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" },
     { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11' and platform_machine == 'x86_64' and sys_platform == 'darwin'" },
-    { name = "packaging", marker = "(python_full_version < '3.11' and platform_machine != 'x86_64') or (python_full_version < '3.11' and sys_platform != 'darwin') or (platform_machine == 'x86_64' and sys_platform == 'darwin')" },
+    { name = "packaging" },
     { name = "pandas", version = "2.3.3", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" },
     { name = "pandas", version = "3.0.3", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11' and platform_machine == 'x86_64' and sys_platform == 'darwin'" },
     { name = "scikit-learn", version = "1.7.2", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" },
     { name = "scikit-learn", version = "1.8.0", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11' and platform_machine == 'x86_64' and sys_platform == 'darwin'" },
     { name = "scipy", version = "1.15.3", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" },
     { name = "scipy", version = "1.17.1", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11' and platform_machine == 'x86_64' and sys_platform == 'darwin'" },
-    { name = "slicer", marker = "(python_full_version < '3.11' and platform_machine != 'x86_64') or (python_full_version < '3.11' and sys_platform != 'darwin') or (platform_machine == 'x86_64' and sys_platform == 'darwin')" },
-    { name = "tqdm", marker = "(python_full_version < '3.11' and platform_machine != 'x86_64') or (python_full_version < '3.11' and sys_platform != 'darwin') or (platform_machine == 'x86_64' and sys_platform == 'darwin')" },
-    { name = "typing-extensions", marker = "(python_full_version < '3.11' and platform_machine != 'x86_64') or (python_full_version < '3.11' and sys_platform != 'darwin') or (platform_machine == 'x86_64' and sys_platform == 'darwin')" },
+    { name = "slicer" },
+    { name = "tqdm" },
+    { name = "typing-extensions" },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/dc/c6/9823a7f483aa9f3179fc359c10d22da9e418b1a7a3fc99a42b705d05e82a/shap-0.49.1.tar.gz", hash = "sha256:1114ecd804fff29f50d522ce6031082fcf42fe4a32fb1b5da233b2415d784c8c", size = 4084725, upload-time = "2025-10-14T10:04:49.75Z" }
 wheels = [
@@ -4997,17 +5084,17 @@ resolution-markers = [
     "(python_full_version == '3.11.*' and platform_machine != 'x86_64' and sys_platform == 'darwin') or (python_full_version == '3.11.*' and sys_platform != 'darwin' and sys_platform != 'emscripten' and sys_platform != 'win32')",
 ]
 dependencies = [
-    { name = "cloudpickle", marker = "(python_full_version == '3.11.*' and platform_machine != 'x86_64') or (python_full_version == '3.11.*' and sys_platform != 'darwin')" },
-    { name = "llvmlite", marker = "(python_full_version == '3.11.*' and platform_machine != 'x86_64') or (python_full_version == '3.11.*' and sys_platform != 'darwin')" },
-    { name = "numba", marker = "(python_full_version == '3.11.*' and platform_machine != 'x86_64') or (python_full_version == '3.11.*' and sys_platform != 'darwin')" },
-    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" }, marker = "(python_full_version == '3.11.*' and platform_machine != 'x86_64') or (python_full_version == '3.11.*' and sys_platform != 'darwin')" },
-    { name = "packaging", marker = "(python_full_version == '3.11.*' and platform_machine != 'x86_64') or (python_full_version == '3.11.*' and sys_platform != 'darwin')" },
-    { name = "pandas", version = "3.0.3", source = { registry = "https://pypi.org/simple" }, marker = "(python_full_version == '3.11.*' and platform_machine != 'x86_64') or (python_full_version == '3.11.*' and sys_platform != 'darwin')" },
-    { name = "scikit-learn", version = "1.8.0", source = { registry = "https://pypi.org/simple" }, marker = "(python_full_version == '3.11.*' and platform_machine != 'x86_64') or (python_full_version == '3.11.*' and sys_platform != 'darwin')" },
-    { name = "scipy", version = "1.17.1", source = { registry = "https://pypi.org/simple" }, marker = "(python_full_version == '3.11.*' and platform_machine != 'x86_64') or (python_full_version == '3.11.*' and sys_platform != 'darwin')" },
-    { name = "slicer", marker = "(python_full_version == '3.11.*' and platform_machine != 'x86_64') or (python_full_version == '3.11.*' and sys_platform != 'darwin')" },
-    { name = "tqdm", marker = "(python_full_version == '3.11.*' and platform_machine != 'x86_64') or (python_full_version == '3.11.*' and sys_platform != 'darwin')" },
-    { name = "typing-extensions", marker = "(python_full_version == '3.11.*' and platform_machine != 'x86_64') or (python_full_version == '3.11.*' and sys_platform != 'darwin')" },
+    { name = "cloudpickle" },
+    { name = "llvmlite" },
+    { name = "numba" },
+    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" } },
+    { name = "packaging" },
+    { name = "pandas", version = "3.0.3", source = { registry = "https://pypi.org/simple" } },
+    { name = "scikit-learn", version = "1.8.0", source = { registry = "https://pypi.org/simple" } },
+    { name = "scipy", version = "1.17.1", source = { registry = "https://pypi.org/simple" } },
+    { name = "slicer" },
+    { name = "tqdm" },
+    { name = "typing-extensions" },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/a4/0a/4a3ee4b1a3654f2a9ae038a64bb3e91a42af3da07577d69b65241f010970/shap-0.51.0.tar.gz", hash = "sha256:cfa17ff213657c9d50285aa923d79b0037a62e2ee1a31bc3eec7e196b00bdb59", size = 4108336, upload-time = "2026-03-04T09:18:19.985Z" }
 wheels = [
@@ -5060,16 +5147,16 @@ resolution-markers = [
     "(python_full_version == '3.12.*' and platform_machine != 'x86_64' and sys_platform == 'darwin') or (python_full_version == '3.12.*' and sys_platform != 'darwin' and sys_platform != 'emscripten' and sys_platform != 'win32')",
 ]
 dependencies = [
-    { name = "cloudpickle", marker = "(python_full_version >= '3.12' and platform_machine != 'x86_64') or (python_full_version >= '3.12' and sys_platform != 'darwin')" },
-    { name = "llvmlite", marker = "(python_full_version >= '3.12' and platform_machine != 'x86_64') or (python_full_version >= '3.12' and sys_platform != 'darwin')" },
-    { name = "numba", marker = "(python_full_version >= '3.12' and platform_machine != 'x86_64') or (python_full_version >= '3.12' and sys_platform != 'darwin')" },
-    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" }, marker = "(python_full_version >= '3.12' and platform_machine != 'x86_64') or (python_full_version >= '3.12' and sys_platform != 'darwin')" },
-    { name = "packaging", marker = "(python_full_version >= '3.12' and platform_machine != 'x86_64') or (python_full_version >= '3.12' and sys_platform != 'darwin')" },
-    { name = "pandas", version = "3.0.3", source = { registry = "https://pypi.org/simple" }, marker = "(python_full_version >= '3.12' and platform_machine != 'x86_64') or (python_full_version >= '3.12' and sys_platform != 'darwin')" },
-    { name = "scikit-learn", version = "1.8.0", source = { registry = "https://pypi.org/simple" }, marker = "(python_full_version >= '3.12' and platform_machine != 'x86_64') or (python_full_version >= '3.12' and sys_platform != 'darwin')" },
-    { name = "scipy", version = "1.17.1", source = { registry = "https://pypi.org/simple" }, marker = "(python_full_version >= '3.12' and platform_machine != 'x86_64') or (python_full_version >= '3.12' and sys_platform != 'darwin')" },
-    { name = "slicer", marker = "(python_full_version >= '3.12' and platform_machine != 'x86_64') or (python_full_version >= '3.12' and sys_platform != 'darwin')" },
-    { name = "tqdm", marker = "(python_full_version >= '3.12' and platform_machine != 'x86_64') or (python_full_version >= '3.12' and sys_platform != 'darwin')" },
+    { name = "cloudpickle" },
+    { name = "llvmlite" },
+    { name = "numba" },
+    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" } },
+    { name = "packaging" },
+    { name = "pandas", version = "3.0.3", source = { registry = "https://pypi.org/simple" } },
+    { name = "scikit-learn", version = "1.8.0", source = { registry = "https://pypi.org/simple" } },
+    { name = "scipy", version = "1.17.1", source = { registry = "https://pypi.org/simple" } },
+    { name = "slicer" },
+    { name = "tqdm" },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/fc/0a/aa278f42c08cb47f2bb503085be0c521da2886929c6605b6105748a7590f/shap-0.52.0.tar.gz", hash = "sha256:81d4ae478f67f8122de1bb411dc4e3ddff0604cbc27dc9cb8ea66d5c73462fd2", size = 4192842, upload-time = "2026-05-28T14:17:49.011Z" }
 wheels = [
@@ -5671,8 +5758,8 @@ name = "uvicorn"
 version = "0.49.0"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
-    { name = "click", marker = "python_full_version < '3.11' or sys_platform != 'emscripten'" },
-    { name = "h11", marker = "python_full_version < '3.11' or sys_platform != 'emscripten'" },
+    { name = "click" },
+    { name = "h11" },
     { name = "typing-extensions", marker = "python_full_version < '3.11'" },
 ]
 sdist = { url = "https://files.pythonhosted.org/packages/c4/1f/fa18009dea8469069cca78a4e877a008ab78f08b064bfc9ab891579077ff/uvicorn-0.49.0.tar.gz", hash = "sha256:ebf4271aa580d9de97f93192d4595176df6e91f9aae919ca73e4fc07df1e66a3", size = 91284, upload-time = "2026-06-03T22:01:30.448Z" }
@@ -6128,6 +6215,23 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/fd/4d/4b880086bd0d3e034d25647be1d830afc3e3f610e98c4ab3490af6b1b6d5/yarl-1.24.2-py3-none-any.whl", hash = "sha256:2783d9226db8797636cd6896e4de81feed252d1db72265686c9558d97a4d94b9", size = 53576, upload-time = "2026-05-19T21:31:03.909Z" },
 ]
 
+[[package]]
+name = "zarr"
+version = "3.1.6"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "donfig" },
+    { name = "google-crc32c" },
+    { name = "numcodecs" },
+    { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" } },
+    { name = "packaging" },
+    { name = "typing-extensions" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/31/5a/b8a0cf39a14c770c30bd1f2d120c54000c8cd9e84e8e79f38d9a7ce58071/zarr-3.1.6.tar.gz", hash = "sha256:d95e72cbea4b90e9a70679468b8266400331756232576ae2b43400ac5108d0eb", size = 386531, upload-time = "2026-03-23T17:25:18.748Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/de/7c/ba8ca8cbe9dbef8e83a95fc208fed8e6686c98b4719aaa0aa7f3d31fe390/zarr-3.1.6-py3-none-any.whl", hash = "sha256:b5a82c5079d1c3d4ee8f06746fa3b9a98a7d804300fa3f4be154362a33e1207e", size = 295655, upload-time = "2026-03-23T17:25:17.189Z" },
+]
+
 [[package]]
 name = "zstandard"
 version = "0.25.0"

From 0bdb0d3b8a8beb06d58f1952ec13c017e4e80176 Mon Sep 17 00:00:00 2001
From: Dhanushkumar Jv <kumardhanush509@gmail.com>
Date: Fri, 31 Jul 2026 22:42:09 +0530
Subject: [PATCH 3/5] Add Zarr subset/reduce and clearer summarize errors

---
 aidrin/file_handling/file_parser.py         |  24 +++-
 aidrin/file_handling/readers/zarr_reader.py | 130 ++++++++++++++++++--
 aidrin/headless/api.py                      |   7 ++
 tests/unit/test_structured_readers.py       |  66 +++++++++-
 4 files changed, 211 insertions(+), 16 deletions(-)

diff --git a/aidrin/file_handling/file_parser.py b/aidrin/file_handling/file_parser.py
index 1a49431b..08d2f996 100644
--- a/aidrin/file_handling/file_parser.py
+++ b/aidrin/file_handling/file_parser.py
@@ -200,7 +200,7 @@ def filter_file(file_info, kept_keys):
 # Parses the uploaded file into a pandas database
 
 
-def read_file(file_info, columns=None):
+def read_file(file_info, columns=None, subset=None, reduce=None):
     """
 
     Parses a given file into pandas Dataframe.
@@ -217,8 +217,14 @@ def read_file(file_info, columns=None):
             -file_path: str, relative or absolute path of the file.
             -file_name: str, file name.
             -file_type: str, file format. Passed from front end select value.
+            - optional selected_keys (4th element) for HDF5/Zarr path selection.
     columns: list of str, optional
         When given, only these columns are returned.
+    subset: dict, optional
+        Zarr only. Map of axis index -> slice or int, applied before reduce.
+    reduce: str, optional
+        Zarr only. ``\"spatial_mean\"`` averages all axes except axis 0 so a
+        multi-dim array becomes a 1D column suitable for existing metrics.
     Returns
     ----------
     pd.Dataframe, None, or str
@@ -248,7 +254,9 @@ def read_file(file_info, columns=None):
         cache_path = _frame_cache_path(file_path) if cacheable else None
 
         # Fast path: reload a fresh cache without touching the source.
-        if cacheable and os.path.exists(cache_path):
+        # Skip cache when Zarr subset/reduce options are set (options change the frame).
+        use_cache = cacheable and subset is None and reduce is None
+        if use_cache and os.path.exists(cache_path):
             try:
                 import pandas as pd
 
@@ -263,7 +271,15 @@ def read_file(file_info, columns=None):
 
         # Slow path: parse the source once.
         reader_cls = READER_MAP[file_type]
-        if file_type in _SELECTION_FILE_TYPES:
+        if file_type == ".zarr":
+            df = reader_cls(
+                file_path,
+                file_upload_time_log,
+                selected_keys=selected_keys,
+                subset=subset,
+                reduce=reduce,
+            ).read()
+        elif file_type in _SELECTION_FILE_TYPES:
             df = reader_cls(
                 file_path, file_upload_time_log, selected_keys=selected_keys
             ).read()
@@ -274,7 +290,7 @@ def read_file(file_info, columns=None):
         # Best-effort cache for future calls. Failure is non-fatal: we still
         # return the freshly parsed frame below, so a read-only directory or an
         # unserialisable frame never turns a successful parse into an error.
-        if cacheable:
+        if use_cache:
             _write_frame_cache(cache_path, df)
 
         if columns is not None:
diff --git a/aidrin/file_handling/readers/zarr_reader.py b/aidrin/file_handling/readers/zarr_reader.py
index 228209f5..7f3c7967 100644
--- a/aidrin/file_handling/readers/zarr_reader.py
+++ b/aidrin/file_handling/readers/zarr_reader.py
@@ -34,12 +34,25 @@ def _require_zarr():
     return zarr
 
 
+# Reduce a multi-dim array to 1D by averaging all axes except axis 0 (time-first).
+REDUCE_SPATIAL_MEAN = "spatial_mean"
+_VALID_REDUCE = {None, "", REDUCE_SPATIAL_MEAN}
+
+
 class zarrReader(StructuredFileReader):
-    """Read Zarr directory stores into pandas DataFrames."""
+    """Read Zarr directory stores into pandas DataFrames.
+
+    Multi-dimensional arrays can be sliced and reduced before conversion::
 
-    def __init__(self, file_path: str, logger, selected_keys=None):
+        zarrReader(path, logger, selected_keys=["tmax"], subset={0: slice(0, 14)},
+                   reduce="spatial_mean")
+    """
+
+    def __init__(self, file_path: str, logger, selected_keys=None, subset=None, reduce=None):
         super().__init__(file_path, logger)
         self._explicit_selected_keys = selected_keys
+        self._subset = subset
+        self._reduce = reduce
 
     def _open_store(self):
         zarr = _require_zarr()
@@ -236,19 +249,113 @@ def _resolve_array(self, root, path: str):
             return None
         return obj
 
+    def _normalize_subset(self, subset, ndim: int):
+        """Build a numpy index tuple from ``{axis: slice|int}``."""
+        if not subset:
+            return None
+        if not isinstance(subset, dict):
+            raise ValueError("subset must be a dict of axis -> slice or int")
+
+        index = [slice(None)] * ndim
+        for axis, selector in subset.items():
+            try:
+                axis_i = int(axis)
+            except (TypeError, ValueError) as exc:
+                raise ValueError(f"subset axis must be an int, got {axis!r}") from exc
+            if axis_i < 0:
+                axis_i += ndim
+            if axis_i < 0 or axis_i >= ndim:
+                raise ValueError(f"subset axis {axis} out of range for ndim={ndim}")
+            if not isinstance(selector, (slice, int, np.integer)):
+                raise ValueError(
+                    f"subset[{axis}] must be slice or int, got {type(selector).__name__}"
+                )
+            index[axis_i] = selector
+        return tuple(index)
+
+    def _apply_subset(self, data, subset):
+        if subset is None:
+            return data
+        index = self._normalize_subset(subset, data.ndim)
+        if index is None:
+            return data
+        return data[index]
+
+    def _apply_reduce(self, data, reduce):
+        """Reduce multi-dim data to 1D (or scalar) for tabular metrics."""
+        mode = (reduce or "").strip().lower() or None
+        if mode not in _VALID_REDUCE:
+            raise ValueError(
+                f"Unsupported reduce={reduce!r}; use None or '{REDUCE_SPATIAL_MEAN}'"
+            )
+        if mode is None:
+            return data
+        if data.ndim <= 1:
+            return data
+        # Keep axis 0 (typically time); average remaining spatial/member axes.
+        axes = tuple(range(1, data.ndim))
+        return np.nanmean(data, axis=axes)
+
+    def _prepare_array_data(self, arr):
+        """Load array, apply optional subset/reduce, return numpy data."""
+        data = np.asarray(arr[:])
+        try:
+            data = self._apply_subset(data, self._subset)
+            data = self._apply_reduce(data, self._reduce)
+        except ValueError as exc:
+            self.logger.warning("%s", exc)
+            return None
+
+        if getattr(data, "ndim", 0) >= 3 and not self._reduce:
+            self.logger.warning(
+                "Zarr array has ndim=%s after subset; pass reduce='%s' "
+                "(mean over axes 1..) to build a 1D table column, or subset further.",
+                data.ndim,
+                REDUCE_SPATIAL_MEAN,
+            )
+            return None
+        return data
+
+    def _column_name_from_path(self, path: str, used_names):
+        full = path.strip("/") if path not in ("", "(root)") else "value"
+        if not full:
+            full = "value"
+        if full not in used_names:
+            return full
+        short = path.split("/")[-1] or full
+        if short not in used_names:
+            return short
+        dotted = full.replace("/", ".")
+        if dotted not in used_names:
+            return dotted
+        suffix = 2
+        while f"{full}_{suffix}" in used_names:
+            suffix += 1
+        return f"{full}_{suffix}"
+
     def _array_to_frame(self, path: str, arr):
-        data = arr[:]
-        col_name = path.split("/")[-1] if path not in ("", "(root)") else "value"
+        data = self._prepare_array_data(arr)
+        if data is None:
+            return None
+
+        col_name = self._column_name_from_path(path, set())
         if getattr(data, "ndim", 0) == 0:
             df = pd.DataFrame({col_name: [data]})
         elif data.ndim == 1:
             df = pd.DataFrame({col_name: data})
-        else:
+        elif data.ndim == 2:
             try:
                 df = pd.DataFrame(data)
             except Exception:
                 df = pd.DataFrame(data.tolist())
-            df.columns = [str(col) for col in df.columns]
+            df.columns = [f"{col_name}_{i}" for i in range(df.shape[1])]
+        else:
+            self.logger.warning(
+                "Cannot convert ndim=%s Zarr array '%s' to a DataFrame",
+                getattr(data, "ndim", None),
+                path,
+            )
+            return None
         df.columns = [str(col) for col in df.columns]
         return df if not df.empty else None
 
@@ -271,10 +378,14 @@ def _read_compatible_array_paths(self, paths):
             arr = self._resolve_array(root, path)
             if arr is None:
                 return None
-            data = np.asarray(arr[:])
+            data = self._prepare_array_data(arr)
+            if data is None:
+                return None
+            data = np.asarray(data)
             if data.ndim != 1:
                 self.logger.warning(
-                    "Zarr multi-select requires 1D arrays; '%s' has ndim=%s",
+                    "Zarr multi-select requires 1D arrays after subset/reduce; "
+                    "'%s' has ndim=%s",
                     path,
                     data.ndim,
                 )
@@ -289,8 +400,7 @@ def _read_compatible_array_paths(self, paths):
                     length,
                 )
                 return None
-            short = path.split("/")[-1] or path
-            name = short if short not in columns else path.replace("/", ".")
+            name = self._column_name_from_path(path, set(columns))
             columns[name] = data
 
         df = pd.DataFrame(columns)
diff --git a/aidrin/headless/api.py b/aidrin/headless/api.py
index dd22f180..c5a4d075 100644
--- a/aidrin/headless/api.py
+++ b/aidrin/headless/api.py
@@ -284,6 +284,13 @@ def summarize_dataset(
     ext = f".{file_type}" if file_type else path.suffix.lower()
     df = read_file((file_path, path.name, ext))
 
+    if df is None or isinstance(df, str):
+        detail = df if isinstance(df, str) else (
+            "Unable to build a table from this file. For multi-array Zarr/HDF5 stores, "
+            "select compatible paths (selected_keys) or reduce multi-dim arrays first."
+        )
+        raise ValueError(detail)
+
     num_cols = df.select_dtypes(include="number").columns.tolist()
     cat_cols = df.select_dtypes(include="object").columns.tolist()
     missing = df.isnull().sum()
diff --git a/tests/unit/test_structured_readers.py b/tests/unit/test_structured_readers.py
index d4cab412..f3aa0b3c 100644
--- a/tests/unit/test_structured_readers.py
+++ b/tests/unit/test_structured_readers.py
@@ -129,7 +129,7 @@ def test_zarr_compatible_group_auto_read(tmp_path, logger):
 
     df = reader.read()
     assert df is not None
-    assert set(df.columns) == {"temp", "x"}
+    assert set(df.columns) == {"temp", "station/x"}
     assert len(df) == 4
 
 
@@ -159,7 +159,7 @@ def test_zarr_grouped_hierarchical_selection(tmp_path, logger):
         str(store), logger, selected_keys=["S1/X", "S1/Y"]
     ).read()
     assert df is not None
-    assert set(df.columns) == {"X", "Y"}
+    assert set(df.columns) == {"S1/X", "S1/Y"}
     assert len(df) == 10
 
 
@@ -185,3 +185,65 @@ def test_zarr_empty_store(tmp_path, logger):
     inv = zarrReader(str(store), logger).inventory()
     assert inv["type"] == INVENTORY_EMPTY
     assert zarrReader(str(store), logger).read() is None
+
+
+def _write_multidim_store(path):
+    """3D array shaped like (time, lat, lon) for subset/reduce tests."""
+    root = zarr.open_group(str(path), mode="w")
+    # time=5, lat=4, lon=3 — values = time index for easy spatial-mean checks
+    data = np.zeros((5, 4, 3), dtype=np.float64)
+    for t in range(5):
+        data[t, :, :] = float(t)
+    arr = root.create_array("tmax_grid", shape=data.shape, dtype="f8")
+    arr[:] = data
+
+
+def test_zarr_multidim_requires_reduce(tmp_path, logger):
+    store = tmp_path / "grid.zarr"
+    _write_multidim_store(store)
+    df = zarrReader(str(store), logger, selected_keys=["tmax_grid"]).read()
+    assert df is None
+
+
+def test_zarr_spatial_mean_reduce(tmp_path, logger):
+    store = tmp_path / "grid.zarr"
+    _write_multidim_store(store)
+    df = zarrReader(
+        str(store),
+        logger,
+        selected_keys=["tmax_grid"],
+        reduce="spatial_mean",
+    ).read()
+    assert df is not None
+    assert list(df.columns) == ["tmax_grid"]
+    assert len(df) == 5
+    assert list(df["tmax_grid"]) == [0.0, 1.0, 2.0, 3.0, 4.0]
+
+
+def test_zarr_subset_then_spatial_mean(tmp_path, logger):
+    store = tmp_path / "grid.zarr"
+    _write_multidim_store(store)
+    # First 3 time steps, subset of lat/lon
+    df = zarrReader(
+        str(store),
+        logger,
+        selected_keys=["tmax_grid"],
+        subset={0: slice(0, 3), 1: slice(0, 2), 2: slice(0, 2)},
+        reduce="spatial_mean",
+    ).read()
+    assert df is not None
+    assert len(df) == 3
+    assert list(df["tmax_grid"]) == [0.0, 1.0, 2.0]
+
+
+def test_zarr_read_file_passes_subset_reduce(tmp_path):
+    store = tmp_path / "grid.zarr"
+    _write_multidim_store(store)
+    df = read_file(
+        (str(store), "grid.zarr", ".zarr", ["tmax_grid"]),
+        reduce="spatial_mean",
+        subset={0: slice(1, 4)},
+    )
+    assert df is not None
+    assert len(df) == 3
+    assert list(df["tmax_grid"]) == [1.0, 2.0, 3.0]

From d592c7e2e8fc3f7792d0c54f64bd68409c14510b Mon Sep 17 00:00:00 2001
From: Dhanushkumar Jv <kumardhanush509@gmail.com>
Date: Tue, 4 Aug 2026 18:22:40 +0530
Subject: [PATCH 4/5] Drop Zarr mean/reduce; add CLI selected-keys

---
 aidrin/file_handling/file_parser.py         | 23 +-----
 aidrin/file_handling/readers/zarr_reader.py | 82 +++------------------
 aidrin/headless/api.py                      | 44 +++++++++--
 aidrin/headless/cli.py                      | 35 ++++++++-
 aidrin/headless/config.py                   |  3 +
 aidrin/headless/runners.py                  | 28 ++++++-
 docs/source/cli_usage.rst                   | 11 +++
 tests/unit/test_cli.py                      | 60 +++++++++++++++
 tests/unit/test_structured_readers.py       | 53 +++----------
 9 files changed, 195 insertions(+), 144 deletions(-)

diff --git a/aidrin/file_handling/file_parser.py b/aidrin/file_handling/file_parser.py
index 08d2f996..dbe1afe4 100644
--- a/aidrin/file_handling/file_parser.py
+++ b/aidrin/file_handling/file_parser.py
@@ -200,7 +200,7 @@ def filter_file(file_info, kept_keys):
 # Parses the uploaded file into a pandas database
 
 
-def read_file(file_info, columns=None, subset=None, reduce=None):
+def read_file(file_info, columns=None):
     """
 
     Parses a given file into pandas Dataframe.
@@ -220,11 +220,6 @@ def read_file(file_info, columns=None, subset=None, reduce=None):
             - optional selected_keys (4th element) for HDF5/Zarr path selection.
     columns: list of str, optional
         When given, only these columns are returned.
-    subset: dict, optional
-        Zarr only. Map of axis index -> slice or int, applied before reduce.
-    reduce: str, optional
-        Zarr only. ``\"spatial_mean\"`` averages all axes except axis 0 so a
-        multi-dim array becomes a 1D column suitable for existing metrics.
     Returns
     ----------
     pd.Dataframe, None, or str
@@ -254,9 +249,7 @@ def read_file(file_info, columns=None, subset=None, reduce=None):
         cache_path = _frame_cache_path(file_path) if cacheable else None
 
         # Fast path: reload a fresh cache without touching the source.
-        # Skip cache when Zarr subset/reduce options are set (options change the frame).
-        use_cache = cacheable and subset is None and reduce is None
-        if use_cache and os.path.exists(cache_path):
+        if cacheable and os.path.exists(cache_path):
             try:
                 import pandas as pd
 
@@ -271,15 +264,7 @@ def read_file(file_info, columns=None, subset=None, reduce=None):
 
         # Slow path: parse the source once.
         reader_cls = READER_MAP[file_type]
-        if file_type == ".zarr":
-            df = reader_cls(
-                file_path,
-                file_upload_time_log,
-                selected_keys=selected_keys,
-                subset=subset,
-                reduce=reduce,
-            ).read()
-        elif file_type in _SELECTION_FILE_TYPES:
+        if file_type in _SELECTION_FILE_TYPES:
             df = reader_cls(
                 file_path, file_upload_time_log, selected_keys=selected_keys
             ).read()
@@ -290,7 +275,7 @@ def read_file(file_info, columns=None, subset=None, reduce=None):
         # Best-effort cache for future calls. Failure is non-fatal: we still
         # return the freshly parsed frame below, so a read-only directory or an
         # unserialisable frame never turns a successful parse into an error.
-        if use_cache:
+        if cacheable:
             _write_frame_cache(cache_path, df)
 
         if columns is not None:
diff --git a/aidrin/file_handling/readers/zarr_reader.py b/aidrin/file_handling/readers/zarr_reader.py
index 7f3c7967..19a28cb9 100644
--- a/aidrin/file_handling/readers/zarr_reader.py
+++ b/aidrin/file_handling/readers/zarr_reader.py
@@ -34,25 +34,17 @@ def _require_zarr():
     return zarr
 
 
-# Reduce a multi-dim array to 1D by averaging all axes except axis 0 (time-first).
-REDUCE_SPATIAL_MEAN = "spatial_mean"
-_VALID_REDUCE = {None, "", REDUCE_SPATIAL_MEAN}
-
-
 class zarrReader(StructuredFileReader):
     """Read Zarr directory stores into pandas DataFrames.
 
-    Multi-dimensional arrays can be sliced and reduced before conversion::
-
-        zarrReader(path, logger, selected_keys=["tmax"], subset={0: slice(0, 14)},
-                   reduce="spatial_mean")
+    Same selection model as HDF5: pass ``selected_keys`` to choose array paths.
+    Only tabular-friendly arrays (0D/1D, or a single 2D array) are converted;
+    higher-dimensional arrays are refused so metrics see raw values, not aggregates.
     """
 
-    def __init__(self, file_path: str, logger, selected_keys=None, subset=None, reduce=None):
+    def __init__(self, file_path: str, logger, selected_keys=None):
         super().__init__(file_path, logger)
         self._explicit_selected_keys = selected_keys
-        self._subset = subset
-        self._reduce = reduce
 
     def _open_store(self):
         zarr = _require_zarr()
@@ -249,69 +241,14 @@ def _resolve_array(self, root, path: str):
             return None
         return obj
 
-    def _normalize_subset(self, subset, ndim: int):
-        """Build a numpy index tuple from ``{axis: slice|int}``."""
-        if not subset:
-            return None
-        if not isinstance(subset, dict):
-            raise ValueError("subset must be a dict of axis -> slice or int")
-
-        index = [slice(None)] * ndim
-        for axis, selector in subset.items():
-            try:
-                axis_i = int(axis)
-            except (TypeError, ValueError) as exc:
-                raise ValueError(f"subset axis must be an int, got {axis!r}") from exc
-            if axis_i < 0:
-                axis_i += ndim
-            if axis_i < 0 or axis_i >= ndim:
-                raise ValueError(f"subset axis {axis} out of range for ndim={ndim}")
-            if not isinstance(selector, (slice, int, np.integer)):
-                raise ValueError(
-                    f"subset[{axis}] must be slice or int, got {type(selector).__name__}"
-                )
-            index[axis_i] = selector
-        return tuple(index)
-
-    def _apply_subset(self, data, subset):
-        if subset is None:
-            return data
-        index = self._normalize_subset(subset, data.ndim)
-        if index is None:
-            return data
-        return data[index]
-
-    def _apply_reduce(self, data, reduce):
-        """Reduce multi-dim data to 1D (or scalar) for tabular metrics."""
-        mode = (reduce or "").strip().lower() or None
-        if mode not in _VALID_REDUCE:
-            raise ValueError(
-                f"Unsupported reduce={reduce!r}; use None or '{REDUCE_SPATIAL_MEAN}'"
-            )
-        if mode is None:
-            return data
-        if data.ndim <= 1:
-            return data
-        # Keep axis 0 (typically time); average remaining spatial/member axes.
-        axes = tuple(range(1, data.ndim))
-        return np.nanmean(data, axis=axes)
-
     def _prepare_array_data(self, arr):
-        """Load array, apply optional subset/reduce, return numpy data."""
+        """Load array data; refuse ndim >= 3 (no averaging / reshape for metrics)."""
         data = np.asarray(arr[:])
-        try:
-            data = self._apply_subset(data, self._subset)
-            data = self._apply_reduce(data, self._reduce)
-        except ValueError as exc:
-            self.logger.warning("%s", exc)
-            return None
-
-        if getattr(data, "ndim", 0) >= 3 and not self._reduce:
+        if getattr(data, "ndim", 0) >= 3:
             self.logger.warning(
-                "Zarr array has ndim=%s after subset; pass reduce='%s' "
-                "(mean over axes 1..) to build a 1D table column, or subset further.",
+                "Zarr array has ndim=%s; refusing to aggregate or flatten. "
+                "Select 1D (or a single 2D) arrays via selected_keys.",
                 data.ndim,
-                REDUCE_SPATIAL_MEAN,
             )
             return None
         return data
@@ -384,8 +321,7 @@ def _read_compatible_array_paths(self, paths):
             data = np.asarray(data)
             if data.ndim != 1:
                 self.logger.warning(
-                    "Zarr multi-select requires 1D arrays after subset/reduce; "
-                    "'%s' has ndim=%s",
+                    "Zarr multi-select requires 1D arrays; '%s' has ndim=%s",
                     path,
                     data.ndim,
                 )
diff --git a/aidrin/headless/api.py b/aidrin/headless/api.py
index dff6882d..97f97287 100644
--- a/aidrin/headless/api.py
+++ b/aidrin/headless/api.py
@@ -28,6 +28,7 @@
     run_feature_relevance,
     run_hipaa_compliance,
     run_k_anonymity,
+    using_selected_keys,
     run_l_diversity,
     run_multiple_attribute_risk,
     run_null_count_trend,
@@ -354,19 +355,24 @@ def summarize_dataset(
     file_path: str,
     file_type: Optional[str] = None,
     max_features: Optional[int] = None,
+    selected_keys: Optional[List[str]] = None,
 ) -> Dict[str, Any]:
     """Return shape, per-column descriptive stats, and missing counts for a dataset."""
-    from pathlib import Path
     from aidrin.file_handling.file_parser import read_file
+    from .runners import _build_file_info
 
-    path = Path(file_path)
-    ext = f".{file_type}" if file_type else path.suffix.lower()
-    df = read_file((file_path, path.name, ext))
+    file_info = _build_file_info(
+        file_path,
+        file_type,
+        None,
+        selected_keys=_normalize_list(selected_keys),
+    )
+    df = read_file(file_info)
 
     if df is None or isinstance(df, str):
         detail = df if isinstance(df, str) else (
             "Unable to build a table from this file. For multi-array Zarr/HDF5 stores, "
-            "select compatible paths (selected_keys) or reduce multi-dim arrays first."
+            "pass selected_keys with compatible 1D paths."
         )
         raise ValueError(detail)
 
@@ -506,6 +512,31 @@ def run_metric(
     verbose: bool = False,
     strip_visualizations: bool = False,
     **kwargs: Any,
+) -> Dict[str, Any]:
+    with using_selected_keys(_normalize_list(kwargs.get("selected_keys"))):
+        return _run_metric_impl(
+            metric_name,
+            file_path,
+            file_type=file_type,
+            file_name=file_name,
+            save_images=save_images,
+            image_dir=image_dir,
+            verbose=verbose,
+            strip_visualizations=strip_visualizations,
+            **kwargs,
+        )
+
+
+def _run_metric_impl(
+    metric_name: str,
+    file_path: str,
+    file_type: Optional[str] = None,
+    file_name: Optional[str] = None,
+    save_images: bool = True,
+    image_dir: Optional[str] = None,
+    verbose: bool = False,
+    strip_visualizations: bool = False,
+    **kwargs: Any,
 ) -> Dict[str, Any]:
     metric_key = metric_name.strip().lower().replace("-", "_")
     metric = METRIC_REGISTRY.get(metric_key)
@@ -741,6 +772,7 @@ def run_batch_metrics(
         "timestamp_column": config_obj.timestamp_column,
         "batch_column": config_obj.batch_column,
         "target_columns": config_obj.target_columns,
+        "selected_keys": config_obj.selected_keys,
         "save_images": bool(config_obj.save_images) if config_obj.save_images is not None else True,
         "image_dir": config_obj.image_dir,
         "verbose": verbose,
@@ -765,6 +797,7 @@ def run_data_quality(
     file_name: Optional[str] = None,
     verbose: bool = False,
     strip_visualizations: bool = True,
+    selected_keys: Optional[List[str]] = None,
 ) -> Dict[str, Any]:
     """Run fast data quality metrics (completeness, duplicity, outliers).
 
@@ -777,6 +810,7 @@ def run_data_quality(
             file_type=file_type,
             file_name=file_name,
             metrics=["completeness", "duplicity", "outliers"],
+            selected_keys=selected_keys or [],
             save_images=False,
         ),
         verbose=verbose,
diff --git a/aidrin/headless/cli.py b/aidrin/headless/cli.py
index e4671c88..aaf37b29 100644
--- a/aidrin/headless/cli.py
+++ b/aidrin/headless/cli.py
@@ -269,6 +269,7 @@ def _build_run_kwargs(args: argparse.Namespace) -> dict:
         "timestamp_column": getattr(args, "timestamp_column", None),
         "batch_column": getattr(args, "batch_column", None),
         "target_columns": _parse_list(getattr(args, "target_columns", None)),
+        "selected_keys": _parse_list(getattr(args, "selected_keys", None)),
         "rules": parsed_rules,
         "rules_json": rules_json,
         "rules_file": rules_file,
@@ -287,6 +288,12 @@ def _build_run_kwargs(args: argparse.Namespace) -> dict:
 
 def _configure_common_run_args(parser: argparse.ArgumentParser) -> None:
     parser.add_argument("--file-type", dest="file_type", default=None, help="Input file type override")
+    parser.add_argument(
+        "--selected-keys",
+        dest="selected_keys",
+        default=None,
+        help="Comma-separated HDF5/Zarr array paths to read (same idea as web selected_keys)",
+    )
     parser.add_argument("--save-images", dest="save_images", action="store_true", help="Save visualizations to disk")
     parser.add_argument("--no-save-images", dest="save_images", action="store_false", help="Do not save visualizations")
     parser.set_defaults(save_images=True)
@@ -298,6 +305,13 @@ def _configure_common_run_args(parser: argparse.ArgumentParser) -> None:
 
 def _configure_minimal_run_args(parser: argparse.ArgumentParser) -> None:
     """Lightweight args for top-level metric shortcuts."""
+    parser.add_argument("--file-type", dest="file_type", default=None, help="Input file type override")
+    parser.add_argument(
+        "--selected-keys",
+        dest="selected_keys",
+        default=None,
+        help="Comma-separated HDF5/Zarr array paths to read",
+    )
     parser.add_argument("-v", "--verbose", action="store_true", help="Show progress output")
 
 
@@ -530,13 +544,30 @@ def main() -> None:
     dq_parser = subparsers.add_parser("data-quality", help="Run fast data quality metrics (completeness, duplicity, outliers)")
     dq_parser.add_argument("file_path")
     dq_parser.add_argument("--file-type", dest="file_type", default=None)
+    dq_parser.add_argument(
+        "--selected-keys",
+        dest="selected_keys",
+        default=None,
+        help="Comma-separated HDF5/Zarr array paths to read",
+    )
     dq_parser.add_argument("-v", "--verbose", action="store_true", help="Show progress output")
     dq_parser.add_argument("--detail", action="store_true", help="Output full per-feature JSON instead of summary")
 
     # Dataset summary command
     summarize_parser = subparsers.add_parser("summarize", help="Describe numerical and categorical features of a dataset")
     summarize_parser.add_argument("file_path", help="Path to the dataset")
-    summarize_parser.add_argument("--file-type", dest="file_type", default=None, help="File type override (csv, parquet, xlsx, hdf5, json, npz)")
+    summarize_parser.add_argument(
+        "--file-type",
+        dest="file_type",
+        default=None,
+        help="File type override (csv, parquet, xlsx, hdf5, json, npz, zarr)",
+    )
+    summarize_parser.add_argument(
+        "--selected-keys",
+        dest="selected_keys",
+        default=None,
+        help="Comma-separated HDF5/Zarr array paths to read",
+    )
     summarize_parser.add_argument(
         "--max-features", dest="max_features", type=int, default=None,
         help="Limit stats to N features (split evenly between numerical and categorical)"
@@ -669,6 +700,7 @@ def main() -> None:
                 args.file_path,
                 file_type=args.file_type,
                 max_features=args.max_features,
+                selected_keys=_parse_list(getattr(args, "selected_keys", None)),
             )
             if args.human_readable:
                 _print_summary_table(result, args.file_path)
@@ -682,6 +714,7 @@ def main() -> None:
                 file_type=args.file_type,
                 verbose=args.verbose,
                 strip_visualizations=True,
+                selected_keys=_parse_list(getattr(args, "selected_keys", None)),
             )
             if args.detail:
                 _dump_result(_round_floats(result))
diff --git a/aidrin/headless/config.py b/aidrin/headless/config.py
index 0e675599..61767d6e 100644
--- a/aidrin/headless/config.py
+++ b/aidrin/headless/config.py
@@ -39,6 +39,7 @@ class HeadlessConfig:
     timestamp_column: Optional[str] = None
     batch_column: Optional[str] = None
     target_columns: Optional[List[str]] = field(default_factory=list)
+    selected_keys: Optional[List[str]] = field(default_factory=list)
     save_images: Optional[bool] = None
     image_dir: Optional[str] = None
 
@@ -68,6 +69,7 @@ def from_dict(cls, data: Dict[str, Any]) -> "HeadlessConfig":
             "file-name": "file_name",
             "image-dir": "image_dir",
             "save-images": "save_images",
+            "selected-keys": "selected_keys",
         }
 
         normalized: Dict[str, Any] = {}
@@ -90,6 +92,7 @@ def from_dict(cls, data: Dict[str, Any]) -> "HeadlessConfig":
             "required_columns",
             "duplicate_columns",
             "target_columns",
+            "selected_keys",
         ):
             if key in normalized:
                 normalized[key] = _normalize_list(normalized[key])
diff --git a/aidrin/headless/runners.py b/aidrin/headless/runners.py
index 2ff9d160..a3bf4782 100644
--- a/aidrin/headless/runners.py
+++ b/aidrin/headless/runners.py
@@ -1,7 +1,9 @@
 import os
 import json
+from contextlib import contextmanager
+from contextvars import ContextVar
 from pathlib import Path
-from typing import Any, Dict, List, Optional
+from typing import Any, Dict, Iterator, List, Optional
 
 from aidrin.file_handling.file_parser import read_file
 from aidrin.structured_data_metrics.add_noise import return_noisy_stats
@@ -48,6 +50,22 @@
 _EXCEL_TYPES = {".xls", ".xlsx", ".xlsm", ".xlsb"}
 _EXCEL_KEY = ".xls, .xlsb, .xlsx, .xlsm"
 
+# Optional HDF5/Zarr path selection for the current headless call (CLI/API/batch).
+_selected_keys_ctx: ContextVar[Optional[List[str]]] = ContextVar(
+    "aidrin_selected_keys", default=None
+)
+
+
+@contextmanager
+def using_selected_keys(selected_keys: Optional[List[str]] = None) -> Iterator[None]:
+    """Apply ``selected_keys`` to all ``_build_file_info`` calls in this context."""
+    keys = [str(k).strip() for k in (selected_keys or []) if str(k).strip()] or None
+    token = _selected_keys_ctx.set(keys)
+    try:
+        yield
+    finally:
+        _selected_keys_ctx.reset(token)
+
 
 class NullTask:
     def update_state(self, *args: Any, **kwargs: Any) -> None:
@@ -68,10 +86,16 @@ def _normalize_file_type(file_type: Optional[str], file_path: str) -> Optional[s
 
 
 def _build_file_info(
-    file_path: str, file_type: Optional[str], file_name: Optional[str]
+    file_path: str,
+    file_type: Optional[str],
+    file_name: Optional[str],
+    selected_keys: Optional[List[str]] = None,
 ) -> tuple:
     normalized_type = _normalize_file_type(file_type, file_path)
     final_name = file_name or os.path.basename(file_path)
+    keys = selected_keys if selected_keys is not None else _selected_keys_ctx.get()
+    if keys:
+        return (file_path, final_name, normalized_type, list(keys))
     return (file_path, final_name, normalized_type)
 
 
diff --git a/docs/source/cli_usage.rst b/docs/source/cli_usage.rst
index 040aa1e1..22fc7db7 100644
--- a/docs/source/cli_usage.rst
+++ b/docs/source/cli_usage.rst
@@ -20,6 +20,14 @@ Quick Start
    # Run a batch of metrics from a YAML config
    aidrin batch /path/to/my_project/batch_config.yaml
 
+   # HDF5 / Zarr: pick compatible 1D arrays (comma-separated paths)
+   aidrin run completeness /path/to/store.zarr --selected-keys age,income
+   aidrin summarize /path/to/file.h5 --selected-keys S1/X,S1/Y
+
+Install Zarr support with ``pip install 'aidrin[zarr]'`` (Python >= 3.11). Local web upload
+does not accept ``.zarr`` directories; use the CLI or library. Multi-dimensional grids are not
+auto-flattened — select 1D (or a single 2D) arrays only.
+
 ----
 
 Sample Dataset
@@ -207,6 +215,9 @@ Results are printed as JSON to stdout. Redirect to a file to save:
 
    target-column: approved
 
+For HDF5/Zarr multi-array stores, add ``selected-keys`` (list or comma-separated string)
+with compatible 1D paths, for example ``selected-keys: [age, income]``.
+
 **Example** — fairness analysis on the sample dataset:
 
 .. code-block:: yaml
diff --git a/tests/unit/test_cli.py b/tests/unit/test_cli.py
index 28d9f9f4..7f55d2e0 100644
--- a/tests/unit/test_cli.py
+++ b/tests/unit/test_cli.py
@@ -633,5 +633,65 @@ def test_no_command_exits_nonzero(self):
         self.assertNotEqual(code, 0)
 
 
+# ===========================================================================
+# selected-keys (HDF5 / Zarr)
+# ===========================================================================
+
+
+class TestSelectedKeysCLI(unittest.TestCase):
+    """CLI --selected-keys for multi-array Zarr stores."""
+
+    @classmethod
+    def setUpClass(cls):
+        pytest = __import__("pytest")
+        zarr = pytest.importorskip("zarr")
+        cls._tmpdir = tempfile.TemporaryDirectory()
+        store = os.path.join(cls._tmpdir.name, "pick.zarr")
+        root = zarr.open_group(store, mode="w")
+        for name, length in (("age", 20), ("income", 20), ("meta", 1)):
+            arr = root.create_array(name, shape=(length,), dtype="f8")
+            arr[:] = np.arange(length, dtype=np.float64)
+        cls.store = store
+
+    @classmethod
+    def tearDownClass(cls):
+        cls._tmpdir.cleanup()
+
+    def test_run_completeness_with_selected_keys(self):
+        stdout, stderr, code = _run_cli(
+            "run",
+            "completeness",
+            self.store,
+            "--selected-keys",
+            "age,income",
+        )
+        self.assertEqual(code, 0, msg=stderr)
+        payload = json.loads(stdout)
+        scores = payload.get("Completeness scores", {})
+        self.assertIn("age", scores)
+        self.assertIn("income", scores)
+        self.assertNotIn("meta", scores)
+
+    def test_summarize_with_selected_keys(self):
+        stdout, stderr, code = _run_cli(
+            "summarize",
+            self.store,
+            "--selected-keys",
+            "age",
+        )
+        self.assertEqual(code, 0, msg=stderr)
+        payload = json.loads(stdout)
+        self.assertEqual(payload["shape"]["columns"], 1)
+        self.assertIn("age", payload.get("numerical", {}))
+
+    def test_build_file_info_includes_selected_keys(self):
+        from aidrin.headless.runners import _build_file_info, using_selected_keys
+
+        with using_selected_keys(["S1/X", "S1/Y"]):
+            info = _build_file_info(self.store, None, None)
+        self.assertEqual(info[2], ".zarr")
+        self.assertEqual(info[3], ["S1/X", "S1/Y"])
+
+
 if __name__ == "__main__":
     unittest.main()
diff --git a/tests/unit/test_structured_readers.py b/tests/unit/test_structured_readers.py
index f3aa0b3c..ab72e594 100644
--- a/tests/unit/test_structured_readers.py
+++ b/tests/unit/test_structured_readers.py
@@ -188,9 +188,8 @@ def test_zarr_empty_store(tmp_path, logger):
 
 
 def _write_multidim_store(path):
-    """3D array shaped like (time, lat, lon) for subset/reduce tests."""
+    """3D array shaped like (time, lat, lon) — not tabular without aggregation."""
     root = zarr.open_group(str(path), mode="w")
-    # time=5, lat=4, lon=3 — values = time index for easy spatial-mean checks
     data = np.zeros((5, 4, 3), dtype=np.float64)
     for t in range(5):
         data[t, :, :] = float(t)
@@ -198,52 +197,18 @@ def _write_multidim_store(path):
     arr[:] = data
 
 
-def test_zarr_multidim_requires_reduce(tmp_path, logger):
+def test_zarr_multidim_refused(tmp_path, logger):
+    """ndim >= 3 must not be averaged or flattened for metrics."""
     store = tmp_path / "grid.zarr"
     _write_multidim_store(store)
     df = zarrReader(str(store), logger, selected_keys=["tmax_grid"]).read()
     assert df is None
 
 
-def test_zarr_spatial_mean_reduce(tmp_path, logger):
-    store = tmp_path / "grid.zarr"
-    _write_multidim_store(store)
-    df = zarrReader(
-        str(store),
-        logger,
-        selected_keys=["tmax_grid"],
-        reduce="spatial_mean",
-    ).read()
-    assert df is not None
-    assert list(df.columns) == ["tmax_grid"]
-    assert len(df) == 5
-    assert list(df["tmax_grid"]) == [0.0, 1.0, 2.0, 3.0, 4.0]
-
-
-def test_zarr_subset_then_spatial_mean(tmp_path, logger):
-    store = tmp_path / "grid.zarr"
-    _write_multidim_store(store)
-    # First 3 time steps, subset of lat/lon
-    df = zarrReader(
-        str(store),
-        logger,
-        selected_keys=["tmax_grid"],
-        subset={0: slice(0, 3), 1: slice(0, 2), 2: slice(0, 2)},
-        reduce="spatial_mean",
-    ).read()
-    assert df is not None
-    assert len(df) == 3
-    assert list(df["tmax_grid"]) == [0.0, 1.0, 2.0]
-
-
-def test_zarr_read_file_passes_subset_reduce(tmp_path):
-    store = tmp_path / "grid.zarr"
-    _write_multidim_store(store)
-    df = read_file(
-        (str(store), "grid.zarr", ".zarr", ["tmax_grid"]),
-        reduce="spatial_mean",
-        subset={0: slice(1, 4)},
-    )
+def test_zarr_read_file_selected_keys(tmp_path):
+    store = tmp_path / "pick.zarr"
+    _write_grouped_hierarchical_store(store)
+    df = read_file((str(store), "pick.zarr", ".zarr", ["S1/X", "S1/Y"]))
     assert df is not None
-    assert len(df) == 3
-    assert list(df["tmax_grid"]) == [1.0, 2.0, 3.0]
+    assert list(df.columns) == ["S1/X", "S1/Y"]
+    assert len(df) == 10

From 66879907db1b4481d62cd8e696496b1060c735b6 Mon Sep 17 00:00:00 2001
From: Dhanushkumar Jv <kumardhanush509@gmail.com>
Date: Tue, 11 Aug 2026 01:17:04 +0530
Subject: [PATCH 5/5] Accept selected_keys on RemoteExecutor summarize and
 data-quality

---
 aidrin/compute/executor.py | 36 +++++++++++++++++++++---------------
 1 file changed, 21 insertions(+), 15 deletions(-)

diff --git a/aidrin/compute/executor.py b/aidrin/compute/executor.py
index d892ac23..0d6c69ad 100644
--- a/aidrin/compute/executor.py
+++ b/aidrin/compute/executor.py
@@ -15,7 +15,7 @@
 """
 
 from dataclasses import asdict, is_dataclass
-from typing import Any, Callable, Dict, Optional
+from typing import Any, Callable, Dict, List, Optional
 
 from aidrin.compute import client
 from aidrin.compute.profiles import RemoteTarget
@@ -138,11 +138,16 @@ def summarize_dataset(
         file_path: str,
         file_type: Optional[str] = None,
         max_features: Optional[int] = None,
+        selected_keys: Optional[List[str]] = None,
     ) -> Dict[str, Any]:
-        return self._call(
-            "summarize",
-            {"file_path": file_path, "file_type": file_type, "max_features": max_features},
-        )
+        payload: Dict[str, Any] = {
+            "file_path": file_path,
+            "file_type": file_type,
+            "max_features": max_features,
+        }
+        if selected_keys:
+            payload["selected_keys"] = selected_keys
+        return self._call("summarize", payload)
 
     def run_data_quality(
         self,
@@ -151,17 +156,18 @@ def run_data_quality(
         file_name: Optional[str] = None,
         verbose: bool = False,
         strip_visualizations: bool = True,
+        selected_keys: Optional[List[str]] = None,
     ) -> Dict[str, Any]:
-        return self._call(
-            "data_quality",
-            {
-                "file_path": file_path,
-                "file_type": file_type,
-                "file_name": file_name,
-                "verbose": verbose,
-                "strip_visualizations": strip_visualizations,
-            },
-        )
+        payload: Dict[str, Any] = {
+            "file_path": file_path,
+            "file_type": file_type,
+            "file_name": file_name,
+            "verbose": verbose,
+            "strip_visualizations": strip_visualizations,
+        }
+        if selected_keys:
+            payload["selected_keys"] = selected_keys
+        return self._call("data_quality", payload)
 
     def run_batch_metrics(
         self,