v0.9.5

.github/workflows/py.yml

@@ -25,7 +25,7 @@ jobs:
         sudo apt-get update
         sudo apt-get -y upgrade
         sudo apt-get -y install libcurl4-openssl-dev python3-rpy2 libmlpack-dev r-base-dev python3-dev pandoc
-        sudo Rscript -e "install.packages(c('genie', 'Rcpp', 'devtools', 'testthat', 'emstreeR', 'knitr'))"
+        sudo Rscript -e "install.packages(c('genie', 'Rcpp', 'devtools', 'tinytest', 'emstreeR', 'knitr'))"
         sudo Rscript -e "devtools::install_github('gagolews/Rd2rst')"
         python -m pip install --upgrade pip
         if [ -f requirements.txt ]; then pip install -r requirements.txt --upgrade; fi

.github/workflows/r.yml

@@ -16,7 +16,7 @@ jobs:
         sudo apt-get update
         sudo apt-get -y upgrade
         sudo apt-get -y install libcurl4-openssl-dev python3-rpy2 libmlpack-dev r-base-dev
-        sudo Rscript -e "install.packages(c('genie', 'Rcpp', 'devtools', 'testthat', 'emstreeR'))"
+        sudo Rscript -e "install.packages(c('genie', 'Rcpp', 'devtools', 'tinytest', 'emstreeR'))"
     - name: Test and check R
       run: |
         sudo make r-test

.gitignore

@@ -21,4 +21,3 @@ cache-*
 coverage.xml
 .coverage
 htmlcov
-

DESCRIPTION

@@ -2,7 +2,7 @@ Package: genieclust
 Type: Package
 Title: The Genie++ Hierarchical Clustering Algorithm with Noise Points Detection
 Version: 0.9.5
-Date: 2020-09-23
+Date: 2020-12-26
 Authors@R: c(
     person("Marek", "Gagolewski",
         role = c("aut", "cre", "cph"),

Makefile

@@ -60,8 +60,7 @@ r-check: r
 	Rscript -e 'devtools::check(cran=TRUE, remote=TRUE, manual=TRUE)'
 
 r-test: r
-	Rscript -e 'options(width=120); source("devel/testthat.R")'
-
+	Rscript -e 'options(width=120); source("devel/tinytest.R")'
 
 r-build:
 	Rscript -e 'Rcpp::compileAttributes()'

NEWS

@@ -1,10 +1,14 @@
 # `genieclust` Package NEWS
 
 
-## genieclust 0.9.5 (under development)
+## genieclust 0.9.5 (2020-12-26)
 
 -   [Python] `nmslib` is now required.
 
+-   [R] switched to `tinytest` for unit testing.
+
+-   [R] fixed the examples that relied on a now-changed API of `emstreeR`.
+
 
 ## genieclust 0.9.4 (2020-07-31)
 

R/mst.R

@@ -212,9 +212,9 @@ emst_mlpack <- function(X, verbose=FALSE)
     if (requireNamespace("emstreeR", quietly=TRUE)) {
 
         if (!verbose)
-            capture.output({mst <- emstreeR::mlpack_mst(X)})
+            capture.output({mst <- emstreeR:::mlpack_mst(X)})
         else
-            mst <- emstreeR::mlpack_mst(X)
+            mst <- emstreeR:::mlpack_mst(X)
 
         structure(t(mst),
             class="mst",

README.rst

@@ -48,7 +48,7 @@ difficult-to-set `eps` parameter.
 Author and Contributors
 -----------------------
 
-Author and Maintainer: `Marek Gagolewski <https://www.gagolewski.com>`_
+Author and maintainer: `Marek Gagolewski <https://www.gagolewski.com>`_
 
 Contributors of the code from the original R package `genie`:
 `Anna Cena <https://cena.rexamine.com>`_,

devel/sandbox_scanpy.py

@@ -10,17 +10,18 @@
 import scanpy as sc
 sc.set_figure_params(dpi=100)
 ad = sc.datasets.pbmc3k()
+
 sc.pp.filter_genes(ad, min_cells=10)
 ad.layers['counts'] = ad.X.copy()
 sc.pp.normalize_total(ad, target_sum=10000)
 sc.pp.log1p(ad)
 sc.pp.highly_variable_genes(ad, n_top_genes=1000, flavor='seurat_v3', subset=True, layer='counts')
 sc.pp.scale(ad, max_value=8)
 sc.pp.pca(ad)
+
 sc.pp.neighbors(ad)
 
 sc.tl.umap(ad)
-
 sc.tl.louvain(ad, resolution=0.2)
 
 X_hidim = ad.X
@@ -30,7 +31,7 @@
 import genieclust
 import numpy as np
 
-g = genieclust.Genie(n_clusters=3, affinity='cosine')
+g = genieclust.Genie(n_clusters=4, affinity='cosine', M=25, postprocess="all")
 labels = g.fit_predict(X_hidim)
 ad.obs['genie_labels'] = labels.astype(str)
 sc.pl.umap(ad, color='genie_labels')
@@ -43,12 +44,12 @@
 sc.pl.umap(ad, color='louvain')
 
 
-mst = genieclust.internal.mst_from_distance(X_hidim, 'cosine')
+mst = genieclust.internal.mst_from_distance(X_hidim)
 genieclust.plots.plot_segments(mst[1], ad.obsm["X_umap"])
 
 
-X_hidim_std = (X_hidim-X_hidim.mean(axis=0))/(X_hidim.std(axis=0))
-g = genieclust.Genie(n_clusters=3, affinity='cosine', M=5, postprocess="merge")
+X_hidim_std = (X_hidim-X_hidim.mean(axis=0))/(X_hidim.std(axis=0, ddof=1))
+g = genieclust.Genie(n_clusters=3)
 labels = g.fit_predict(X_hidim_std)
 ad.obs['genie_labels_std'] = labels.astype(str)
 sc.pl.umap(ad, color='genie_labels_std')
@@ -63,3 +64,10 @@
 #sc.pl.umap(ad, color='genie_labels')
 genieclust.plots.plot_segments(mst[1], ad.obsm["X_umap"])
 
+
+from sklearn.decomposition import PCA
+g = genieclust.Genie(n_clusters=3)
+labels = g.fit_predict(PCA(20).fit_transform(X_hidim))
+ad.obs['genie_labels'] = labels.astype(str)
+sc.pl.umap(ad, color='genie_labels')
+

devel/sphinx/index.rst

@@ -1,6 +1,8 @@
 `genieclust`: Fast and Robust Hierarchical Clustering with Noise Point Detection
 ================================================================================
 
+
+
     **Genie finds meaningful clusters and is fast even on large data sets.**
 
     -- by `Marek Gagolewski <https://www.gagolewski.com/>`_

devel/sphinx/rapi/comparing_partitions.rst

@@ -87,4 +87,3 @@ Examples
    adjusted_mi_score(y_true, y_pred)
    normalized_accuracy(y_true, y_pred)
    pair_sets_index(y_true, y_pred)
-

devel/sphinx/rapi/gclust.rst

@@ -147,4 +147,3 @@ Examples
 
    # Fast for low-dimensional Euclidean spaces:
    if (require("emstreeR")) h <- gclust(emst_mlpack(X))
-

devel/sphinx/rapi/inequity.rst

@@ -18,9 +18,9 @@ Usage
 Arguments
 ~~~~~~~~~
 
-+-------+---------------------------------------+
-| ``x`` | numeric vector of non-negative values |
-+-------+---------------------------------------+
+===== =====================================
+``x`` numeric vector of non-negative values
+===== =====================================
 
 Details
 ~~~~~~~
@@ -72,4 +72,3 @@ Examples
    bonferroni_index(c(0, 0, 10, 0, 0))
    bonferroni_index(c(7, 0, 3, 0, 0))
    bonferroni_index(c(6, 0, 3, 1, 0))
-

devel/sphinx/rapi/mst.rst

@@ -92,4 +92,3 @@ Examples
    data("iris")
    X <- iris[1:4]
    tree <- mst(X)
-

devel/tinytest.R

@@ -0,0 +1,4 @@
+set.seed(123)
+library("tinytest")
+library("genieclust")
+run_test_dir("devel/tinytest/")

devel/testthat/test-compare-partitions.R → devel/tinytest/test-compare-partitions.R

@@ -1,6 +1,5 @@
-library("testthat")
+library("tinytest")
 library("genieclust")
-context("compare-partitions")
 
 # More thorough tests are performed by pytest
 

devel/testthat/test-gclust.R → devel/tinytest/test-gclust.R

@@ -1,6 +1,5 @@
-library("testthat")
+library("tinytest")
 library("genieclust")
-context("gclust")
 
 
 verbose <- FALSE
@@ -20,7 +19,7 @@ if (require("genie")) {
     #for (k in 2:20) expect_equal(adjusted_rand_score(cutree(h1, k), cutree(h2, k)), 1.0)
 }
 
-if (require("emstreeR")) {
+if (suppressMessages(require("emstreeR"))) {
     print(system.time(h2 <- gclust(emst_mlpack(X, verbose=verbose), verbose=verbose)))
     #for (k in 2:20) expect_equal(adjusted_rand_score(cutree(h1, k), cutree(h2, k)), 1.0)
 }

devel/testthat/test-inequity.R → devel/tinytest/test-inequity.R

@@ -1,6 +1,6 @@
-library("testthat")
+library("tinytest")
 library("genieclust")
-context("inequity")
+
 
 # More thorough tests are performed by pytest
 

devel/testthat/test-mst.R → devel/tinytest/test-mst.R

@@ -1,6 +1,6 @@
-library("testthat")
+library("tinytest")
 library("genieclust")
-context("mst")
+
 
 
 set.seed(123)

docs/_sources/index.rst.txt

@@ -1,6 +1,8 @@
 `genieclust`: Fast and Robust Hierarchical Clustering with Noise Point Detection
 ================================================================================
 
+
+
     **Genie finds meaningful clusters and is fast even on large data sets.**
 
     -- by `Marek Gagolewski <https://www.gagolewski.com/>`_

docs/_sources/rapi/comparing_partitions.rst.txt

@@ -87,4 +87,3 @@ Examples
    adjusted_mi_score(y_true, y_pred)
    normalized_accuracy(y_true, y_pred)
    pair_sets_index(y_true, y_pred)
-

docs/_sources/rapi/gclust.rst.txt

@@ -147,4 +147,3 @@ Examples
 
    # Fast for low-dimensional Euclidean spaces:
    if (require("emstreeR")) h <- gclust(emst_mlpack(X))
-

docs/_sources/rapi/inequity.rst.txt

@@ -18,9 +18,9 @@ Usage
 Arguments
 ~~~~~~~~~
 
-+-------+---------------------------------------+
-| ``x`` | numeric vector of non-negative values |
-+-------+---------------------------------------+
+===== =====================================
+``x`` numeric vector of non-negative values
+===== =====================================
 
 Details
 ~~~~~~~
@@ -72,4 +72,3 @@ Examples
    bonferroni_index(c(0, 0, 10, 0, 0))
    bonferroni_index(c(7, 0, 3, 0, 0))
    bonferroni_index(c(6, 0, 3, 1, 0))
-

docs/_sources/rapi/mst.rst.txt

@@ -92,4 +92,3 @@ Examples
    data("iris")
    X <- iris[1:4]
    tree <- mst(X)
-