From 0a0659539aa92e43fdd7926fc2045362aecba9f9 Mon Sep 17 00:00:00 2001
From: Jerry Wang <jerrywang@Jerrys-MacBook-Pro.local>
Date: Sat, 16 May 2026 20:48:23 +0800
Subject: [PATCH] Make kmedoids dependency optional with lazy import

The top-level 'import kmedoids' in core/initial_selection_strategies
made every active-learning strategy depend on the kmedoids package
because Hydra imports the whole module to instantiate any strategy.
This broke run_config.py in environments that hadn't installed
kmedoids yet, even when the user was not selecting k-medoids.

Defer the import to KMedoidsInitialSelection.__init__ (fail-fast for
opted-in users, invisible for everyone else) and move 'kmedoids' from
required dependencies to an optional extra '[kmedoids]' so the default
install no longer pulls it in.
---
 core/initial_selection_strategies.py | 16 +++++++++++++++-
 pyproject.toml                       |  4 +++-
 uv.lock                              |  8 +++++---
 3 files changed, 23 insertions(+), 5 deletions(-)

diff --git a/core/initial_selection_strategies.py b/core/initial_selection_strategies.py
index 074605f..62b505f 100644
--- a/core/initial_selection_strategies.py
+++ b/core/initial_selection_strategies.py
@@ -5,7 +5,6 @@
 import logging
 from abc import ABC, abstractmethod
 
-import kmedoids
 import numpy as np
 from sklearn.cluster import KMeans, MiniBatchKMeans
 from sklearn.metrics import pairwise_distances, pairwise_distances_argmin_min
@@ -142,6 +141,19 @@ def __init__(
         max_iter: int = 100,
         init: str = "build",
     ) -> None:
+        # Imported lazily so that environments without the optional
+        # ``kmedoids`` package can still use the other strategies. Hydra
+        # only constructs this class when it's actually selected, so the
+        # ImportError surfaces immediately and only for users who opted
+        # in to k-medoids.
+        try:
+            import kmedoids  # noqa: F401  (probe install, used in _kmedoids)
+        except ImportError as exc:  # pragma: no cover - exercised only without dep
+            raise ImportError(
+                "KMedoidsInitialSelection requires the 'kmedoids' package. "
+                "Install it with `uv pip install kmedoids` (or `uv sync`)."
+            ) from exc
+
         super().__init__("KMEDOIDS", starting_batch_size=starting_batch_size)
         self.seed = seed
         self.metric = metric
@@ -163,6 +175,8 @@ def select(
         return selected
 
     def _kmedoids(self, embeddings: np.ndarray) -> list[int]:
+        import kmedoids  # already validated in __init__
+
         num_samples = embeddings.shape[0]
         k = min(self.starting_batch_size, num_samples)
         if k == 0:
diff --git a/pyproject.toml b/pyproject.toml
index e89b864..b54eca8 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -26,7 +26,6 @@ classifiers = [
 dependencies = [
     "botorch>=0.16.1",
     "hydra-core>=1.3.2",
-    "kmedoids>=0.5.1",
     "numpy>=1.21.0",
     "pandas>=1.3.0",
     "scikit-learn>=1.6.0",
@@ -42,6 +41,9 @@ cluster = [
     "hydra-submitit-launcher>=1.2.0",
     "submitit>=1.5.3",
 ]
+kmedoids = [
+    "kmedoids>=0.5.1",
+]
 
 [project.urls]
 Homepage = "https://github.com/cellethology/deepdraw"
diff --git a/uv.lock b/uv.lock
index a69009b..6e90384 100644
--- a/uv.lock
+++ b/uv.lock
@@ -392,7 +392,6 @@ source = { editable = "." }
 dependencies = [
     { name = "botorch" },
     { name = "hydra-core" },
-    { name = "kmedoids" },
     { name = "numpy" },
     { name = "pandas" },
     { name = "scikit-learn" },
@@ -408,6 +407,9 @@ cluster = [
     { name = "hydra-submitit-launcher" },
     { name = "submitit" },
 ]
+kmedoids = [
+    { name = "kmedoids" },
+]
 
 [package.dev-dependencies]
 dev = [
@@ -429,7 +431,7 @@ requires-dist = [
     { name = "botorch", specifier = ">=0.16.1" },
     { name = "hydra-core", specifier = ">=1.3.2" },
     { name = "hydra-submitit-launcher", marker = "extra == 'cluster'", specifier = ">=1.2.0" },
-    { name = "kmedoids", specifier = ">=0.5.1" },
+    { name = "kmedoids", marker = "extra == 'kmedoids'", specifier = ">=0.5.1" },
     { name = "numpy", specifier = ">=1.21.0" },
     { name = "openpyxl", marker = "extra == 'analysis'", specifier = ">=3.1.5" },
     { name = "pandas", specifier = ">=1.3.0" },
@@ -438,7 +440,7 @@ requires-dist = [
     { name = "submitit", marker = "extra == 'cluster'", specifier = ">=1.5.3" },
     { name = "tqdm", marker = "extra == 'analysis'", specifier = "<=4.66.5" },
 ]
-provides-extras = ["analysis", "cluster"]
+provides-extras = ["analysis", "cluster", "kmedoids"]
 
 [package.metadata.requires-dev]
 dev = [