JelmerBot · JelmerBot · Dec 17, 2025 · Dec 17, 2025 · Dec 17, 2025
diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml
@@ -40,6 +40,6 @@ jobs:
       run: |
         TAG=${{ github.ref }}
         VERSION=${TAG#refs/tags/v}
-        gh release create -R JelmerBot/plscan -t "Version $VERSION" -n "**Full Changelog**: https://github.com/JelmerBot/plscan/commits/$TAG" "$TAG" dist/*.whl dist/*.tar.gz
+        gh release create -R JelmerBot/fast_plscan -t "Version $VERSION" -n "**Full Changelog**: https://github.com/JelmerBot/fast_plscan/commits/$TAG" "$TAG" dist/*.whl dist/*.tar.gz
       env:
         GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}  
diff --git a/README.md b/README.md
@@ -1,11 +1,13 @@
-[![PyPi version](https://badge.fury.io/py/plscan.svg)](https://badge.fury.io/py/plscan)
-![Conda version](https://anaconda.org/conda-forge/plscan/badges/version.svg)
+[![PyPi version](https://badge.fury.io/py/fast-plscan.svg)](https://badge.fury.io/py/plscan)
+![Conda version](https://anaconda.org/conda-forge/fast-plscan/badges/version.svg)
 [![Repository DOI](https://zenodo.org/badge/xxx.svg)](https://zenodo.org/doi/xxx/zenodo.yyy)
 
-# Persistent Leaf Spatial Clustering for Applications with Noise
+# Persistent Leaves Spatial Clustering for Applications with Noise
 
 This library provides a new clustering algorithm based on HDBSCAN*. The primary
-advantages of PLSCAN over the standard ``hdbscan`` library are:
+advantages of PLSCAN over the
+[``hdbscan``](https://github.com/scikit-learn-contrib/hdbscan) and
+[``fast_hdbscan``](https://github.com/TutteInstitute/fast_hdbscan) libraries are:
 
  - PLSCAN automatically finds the optimal minimum cluster size.
  - PLSCAN can easily use all available cores to speed up computation.
@@ -22,7 +24,7 @@ stable clusters.
 import numpy as np
 import matplotlib.pyplot as plt
 
-from plscan import PLSCAN
+from fast_plscan import PLSCAN
 
 data = np.load("docs/data/data.npy")
 
@@ -186,4 +188,4 @@ When using this work, please cite our (upcoming) preprint:
 
 ## Licensing
 
-The ``plscan`` package has a 3-Clause BSD license.
+The ``fast-plscan`` package has a 3-Clause BSD license.
diff --git a/docs/_paper_figures.ipynb b/docs/_paper_figures.ipynb
@@ -2,19 +2,19 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": null,
    "id": "098248f7",
    "metadata": {},
    "outputs": [],
    "source": [
     "import warnings\n",
     "import numpy as np\n",
     "import pandas as pd\n",
-    "from plscan import PLSCAN\n",
     "from hdbscan import HDBSCAN\n",
     "from itertools import product\n",
     "from collections import defaultdict\n",
     "\n",
+    "from fast_plscan import PLSCAN\n",
     "from lib.plotting import *\n",
     "from lib.drawing import regplot_lowess_ci\n",
     "\n",

diff --git a/docs/_trial_barcodes.ipynb b/docs/_trial_barcodes.ipynb
@@ -40,7 +40,7 @@
     "import matplotlib.pyplot as plt\n",
     "from gtda.homology import VietorisRipsPersistence\n",
     "\n",
-    "from plscan import PLSCAN\n",
+    "from fast_plscan import PLSCAN\n",
     "\n",
     "warnings.filterwarnings(\"ignore\", category=FutureWarning)\n",
     "plt.rcParams[\"figure.dpi\"] = 150\n",
@@ -82,7 +82,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "from plscan._api import compute_cluster_labels\n",
+    "from fast_plscan._api import compute_cluster_labels\n",
     "\n",
     "l, p = compute_cluster_labels(c._leaf_tree, c._condensed_tree, np.array([18]).astype(np.uint32), n)\n",
     "i = np.where(l >= 0)[0]"

diff --git a/docs/_trial_persistence_pruning.ipynb b/docs/_trial_persistence_pruning.ipynb
diff --git a/docs/conf.py b/docs/conf.py
@@ -21,13 +21,13 @@
 
 # -- Project information -----------------------------------------------------
 
-project = "plscan"
+project = "fast_plscan"
 copyright = "2025, Jelmer Bot"
 author = "Jelmer Bot"
 
 # -- General configuration ---------------------------------------------------
 
-release = get_version("plscan")
+release = get_version("fast_plscan")
 version = ".".join(release.split(".")[:2])
 master_doc = "index"
 templates_path = ["_templates"]
@@ -61,4 +61,4 @@
 # -- Options for HTML output -------------------------------------------------
 
 html_theme = "furo"
-htmlhelp_basename = "plscan_doc"
+htmlhelp_basename = "fast_plscan_doc"
diff --git a/docs/demo_computational_performance.ipynb b/docs/demo_computational_performance.ipynb
@@ -7,17 +7,17 @@
    "source": [
     "# Performance Benchmarks\n",
     "\n",
-    "The ``plscan`` library is a re-implementation of the ``fast_hdbscan`` library,\n",
-    "while retaining most of the features from the classic ``hdbscan`` library. In\n",
-    "this notebook, we compare the computational performance of these libraries.\n",
-    "Since all three implementations use *space trees*, we are particularly\n",
-    "interested in understanding how their performance changes as the number of\n",
-    "dimensions increases."
+    "The ``fast-plscan`` library is a re-implementation of the ``fast-hdbscan``\n",
+    "library, while retaining most of the features from the classic ``hdbscan``\n",
+    "library. In this notebook, we compare the computational performance of these\n",
+    "libraries. Since all three implementations use *space trees*, we are\n",
+    "particularly interested in understanding how their performance changes as the\n",
+    "number of dimensions increases."
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": 1,
    "id": "32686d54",
    "metadata": {},
    "outputs": [],
@@ -31,10 +31,10 @@
     "from tqdm import tqdm\n",
     "from sklearn.datasets import make_blobs\n",
     "\n",
-    "from plscan import PLSCAN\n",
     "from hdbscan import HDBSCAN\n",
-    "from sklearn.cluster import KMeans\n",
     "from fast_hdbscan import HDBSCAN as FastHDBSCAN\n",
+    "from sklearn.cluster import KMeans\n",
+    "from fast_plscan import PLSCAN\n",
     "\n",
     "plt.rcParams[\"figure.dpi\"] = 150\n",
     "plt.rcParams[\"figure.figsize\"] = (2.75, 0.618 * 2.75)"
@@ -50,16 +50,16 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "plscan 0.1.dev28+g9eb67a7\n",
+      "fast-plscan 0.1.dev33+gdfffaf8.d20251217\n",
       "hdbscan 0.8.40\n",
-      "fast_hdbscan 0.2.2\n",
+      "fast-hdbscan 0.2.2\n",
       "scikit-learn 1.7.0\n"
      ]
     }
    ],
    "source": [
     "from importlib.metadata import version\n",
-    "for name in ['plscan', 'hdbscan', 'fast_hdbscan', 'scikit-learn']:\n",
+    "for name in ['fast-plscan', 'hdbscan', 'fast-hdbscan', 'scikit-learn']:\n",
     "    print(name, version(name))"
    ]
   },
@@ -204,8 +204,8 @@
    "metadata": {},
    "source": [
     "Plotting the results, we observe considerable differences between ``hdbscan``\n",
-    "and the ``fast_hdbscan`` and ``plscan`` libraries on low-dimensional data. The\n",
-    "parallel implementations offer substantial improvements in scalability."
+    "and the ``fast-hdbscan`` and ``fast-plscan`` libraries on low-dimensional data.\n",
+    "The parallel implementations offer substantial improvements in scalability."
    ]
   },
   {
@@ -247,7 +247,7 @@
     "At 10 dimensions, the advantage of parallelization remains, but the scaling\n",
     "curves become steeper than that of ``kmeans``, which is largely unaffected by\n",
     "the number of dimensions. Interestingly, the ball-tree implementation in\n",
-    "``plscan`` scales considerably worse than the kd-tree implementation.\n",
+    "``fast-plscan`` scales considerably worse than the kd-tree implementation.\n",
     "\n",
     "At 20 dimensions, the parallel implementations provide little benefit when\n",
     "compared to ``kmeans``, which continues to perform efficiently regardless of\n",
@@ -256,10 +256,10 @@
     "Plotting over the dimensions we see only ``kmeans`` does not need more time with\n",
     "higher dimensional data. The scaling trends are a bit more complicated.\n",
     "``hdbscan`` starts of steepest but appears to level off a bit. Ball-tree\n",
-    "``plscan`` quickly approaches the ``hdbscan`` curve with higher dimensions. The\n",
-    "curves for ``fast_hdbscan`` and kd-tree ``plscan`` remain fairly shallow, with\n",
-    "``plscan`` consistently being slightly quicker. ``kmeans`` is unaffected by the\n",
-    "number of dimensions, achieving quick times in all cases."
+    "``fast-plscan`` quickly approaches the ``hdbscan`` curve with higher dimensions.\n",
+    "The curves for ``fast_hdbscan`` and kd-tree ``fast-plscan`` remain fairly\n",
+    "shallow, with ``fast-plscan`` consistently being slightly quicker. ``kmeans`` is\n",
+    "unaffected by the number of dimensions, achieving quick times in all cases."
    ]
   },
   {

diff --git a/docs/demo_parameter_sensitivity.ipynb b/docs/demo_parameter_sensitivity.ipynb
@@ -8,8 +8,8 @@
     "# Parameter sensitivity analysis\n",
     "\n",
     "[The cluster selection strategies notebook](./demo_selection_strategies.ipynb)\n",
-    "demonstrates ``plscan`` is less sensitive to the ``min_samples`` parameter ($k$)\n",
-    "than ``hdbscan``. This notebook runs a more comprehensive parameter sensitivity\n",
+    "demonstrates PLSCAN is less sensitive to the ``min_samples`` parameter ($k$)\n",
+    "than HDBSCAN*. This notebook runs a more comprehensive parameter sensitivity\n",
     "analysis to determine whether that pattern holds on other datasets. This\n",
     "parameter sensitivity analysis tells us how the clustering quality changes as a\n",
     "result of changing $k$. The resulting value is independent of $k$ itself.\n",
@@ -52,7 +52,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 18,
+   "execution_count": 1,
    "id": "ef96a020",
    "metadata": {},
    "outputs": [],
@@ -72,8 +72,8 @@
     "from umap import UMAP\n",
     "from lensed_umap import embed_graph\n",
     "\n",
-    "from plscan import PLSCAN\n",
     "from hdbscan import HDBSCAN\n",
+    "from fast_plscan import PLSCAN\n",
     "from sklearn.metrics import adjusted_rand_score, homogeneity_score, completeness_score\n",
     "\n",
     "from lib.plotting import sns, plt, mpl, lighten, frame_off\n",
@@ -240,7 +240,7 @@
    "source": [
     "Then, we create functions that evaluate the algorithms. These functions return\n",
     "one or more records identifying the dataset, algorithm configuration, and\n",
-    "resulting quality scores. For `plscan`, we compute scores for its top-$n$\n",
+    "resulting quality scores. For PLSCAN, we compute scores for its top-$n$\n",
     "layers:"
    ]
   },

diff --git a/docs/demo_selection_strategies.ipynb b/docs/demo_selection_strategies.ipynb
diff --git a/docs/index.rst b/docs/index.rst
@@ -24,10 +24,10 @@
    :maxdepth: 1
    :hidden:
 
-   _autosummary/plscan
-   _autosummary/plscan.plots
-   _autosummary/plscan._api
-   _autosummary/plscan._helpers
+   _autosummary/fast_plscan
+   _autosummary/fast_plscan.plots
+   _autosummary/fast_plscan._api
+   _autosummary/fast_plscan._helpers
 
 .. toctree::
    :caption: Development
@@ -42,7 +42,9 @@ Persistent Leaves Spatial Clustering of Applications with Noise
 ===============================================================
 
 This library provides a new clustering algorithm based on HDBSCAN*. The primary
-advantages of PLSCAN over the standard ``hdbscan`` library are:
+advantages of PLSCAN over the `hdbscan
+<https://github.com/scikit-learn-contrib/hdbscan>`_ and `fast_hdbscan
+<https://github.com/TutteInstitute/fast_hdbscan>`_ libraries are:
 
  - PLSCAN automatically finds the optimal minimum cluster size.
  - PLSCAN can easily use all available cores to speed up computation;
@@ -60,7 +62,7 @@ stable clusters.
     import numpy as np
     import matplotlib.pyplot as plt
 
-    from plscan import PLSCAN
+    from fast_plscan import PLSCAN
 
     data = np.load("docs/data/data.npy")
 
@@ -144,11 +146,11 @@ When using this work, please cite our (upcoming) preprint:
 Licensing
 ---------
 
-The ``plscan`` package has a 3-Clause BSD license.
+The ``fast-plscan`` package has a 3-Clause BSD license.
 
-.. |PyPI version| image:: https://badge.fury.io/py/plscan.svg
-   :target: https://badge.fury.io/py/plscan
-.. |Conda version| image:: https://anaconda.org/conda-forge/plscan/badges/version.svg
-   :target: https://anaconda.org/conda-forge/plscan
+.. |PyPI version| image:: https://badge.fury.io/py/fast-plscan.svg
+   :target: https://badge.fury.io/py/fast-plscan
+.. |Conda version| image:: https://anaconda.org/conda-forge/fast-plscan/badges/version.svg
+   :target: https://anaconda.org/conda-forge/fast-plscan
 .. |DOI badge| image:: https://zenodo.org/badge/xxx.svg
    :target: https://zenodo.org/doi/xxx/zenodo.yyy
diff --git a/docs/reference_plscan.rst b/docs/reference_plscan.rst
@@ -6,5 +6,5 @@ The public API
    :toctree: _autosummary
    :recursive:
 
-   plscan
-   plscan.plots
+   fast_plscan
+   fast_plscan.plots
diff --git a/docs/using_basic_api.ipynb b/docs/using_basic_api.ipynb
@@ -7,10 +7,10 @@
    "source": [
     "# Basic usage\n",
     "\n",
-    "You can use ``plscan`` as a drop in replacement for ``hdbscan`` for basic usage\n",
-    "of ``hdbscan``. That is, ``plscan`` mostly uses the same interface as\n",
-    "``hdbscan`` -- so whenever you are using the parts of ``hdbscan`` that\n",
-    "``plscan`` supports then you can simply swap them."
+    "You can use PLSCAN as a drop in replacement for HDBSCAN for basic usage. That\n",
+    "is, ``fast-plscan`` mostly uses the same interface as ``hdbscan`` -- so whenever\n",
+    "you are using the parts of ``hdbscan`` that ``fast-plscan`` supports then you\n",
+    "can simply swap them."
    ]
   },
   {
@@ -67,10 +67,10 @@
    "id": "9fcf9ff8",
    "metadata": {},
    "source": [
-    "To use ``plscan`` you can apply it exactly as you would ``hdbscan``. In this\n",
-    "case, no parameter values are needed to find a usable clustering because PLSCAN\n",
-    "will automatically find the minimum cluster size that produces \"optimal\" leaf\n",
-    "clusters."
+    "To use ``fast-plscan`` you can apply it exactly as you would ``hdbscan``. In\n",
+    "this case, no parameter values are needed to find a usable clustering because\n",
+    "PLSCAN will automatically find the minimum cluster size that produces \"optimal\"\n",
+    "leaf clusters."
    ]
   },
   {
@@ -91,7 +91,7 @@
     }
    ],
    "source": [
-    "from plscan import PLSCAN\n",
+    "from fast_plscan import PLSCAN\n",
     "\n",
     "labels = PLSCAN().fit_predict(data)\n",
     "\n",
@@ -106,7 +106,7 @@
    "id": "b337de5c",
    "metadata": {},
    "source": [
-    "And that's all there is to it in terms of getting started with ``plscan``."
+    "And that's all there is to it in terms of getting started with PLSCAN."
    ]
   }
  ],

diff --git a/docs/using_bi_persistences.ipynb b/docs/using_bi_persistences.ipynb
diff --git a/docs/using_exploration_plots.ipynb b/docs/using_exploration_plots.ipynb
@@ -7,7 +7,7 @@
    "source": [
     "# Exploration plots\n",
     "\n",
-    "``plscan`` contains several attributes for exploring and plotting its the cluster\n",
+    "PLSCAN contains several attributes for exploring and plotting its the cluster\n",
     "hierarchy."
    ]
   },
@@ -20,7 +20,7 @@
    "source": [
     "import numpy as np\n",
     "import matplotlib.pyplot as plt\n",
-    "from plscan import PLSCAN\n",
+    "from fast_plscan import PLSCAN\n",
     "\n",
     "plt.rcParams[\"figure.dpi\"] = 150\n",
     "plt.rcParams[\"figure.figsize\"] = (2.75, 0.618 * 2.75)\n",
@@ -35,9 +35,9 @@
    "source": [
     "### Condensed tree\n",
     "\n",
-    "Like ``hdbscan``, ``plscan`` has a condensed tree showing the cluster hierarchy\n",
-    "along data point distances. Unlike ``hdbscan``, ``plscan`` uses distances rather\n",
-    "than density estimates to plot the tree."
+    "Like HDBSCAN, PLSCAN has a condensed tree showing the cluster hierarchy along\n",
+    "data point distances. Unlike HDBSCAN, PLSCAN uses distances rather than density\n",
+    "estimates to plot the tree."
    ]
   },
   {
@@ -103,8 +103,9 @@
    "source": [
     "### Leaf tree\n",
     "\n",
-    "``plscan`` does not extract clusters from the condensed tree directly. Instead, it\n",
-    "first creates a cluster hierarchy describing which clusters exist at each minimum cluster size value. The `leaf_tree_` attribute contains this hierarchy:"
+    "PLSCAN does not extract clusters from the condensed tree directly. Instead, it\n",
+    "first creates a cluster hierarchy describing which clusters exist at each\n",
+    "minimum cluster size value. The `leaf_tree_` attribute contains this hierarchy:"
    ]
   },
   {
@@ -170,7 +171,7 @@
    "source": [
     "### Persistence trace\n",
     "\n",
-    "``plscan`` extracts clusters from the leaf tree by computing each leaf cluster's\n",
+    "PLSCAN extracts clusters from the leaf tree by computing each leaf cluster's\n",
     "persistence and finding the minimum cluster size with the highest leaf-cluster\n",
     "persistence sum. The `persistence_trace_` attribute can plot this minimum\n",
     "cluster size -- persistence trace."