Issue #117 gemelli installation in q2 20264 bug fixes

CHANGELOG.md

@@ -1,3 +1,10 @@
+v0.0.13 (2026-03-02)
+
+### Bug fixes
+
+* Update dependency updates
+    * See issue #117
+    * All back-compatible
 
 v0.0.12 (2024-10-01)
 

README.md

@@ -84,7 +84,6 @@ Martino, C. and Shenhav, L. et al. Context-aware dimensionality reduction deconv
 }
 ```
 
-
 ## Citation for RPCA
 
 ```
@@ -108,7 +107,6 @@ Martino, C. et al. A Novel Sparse Compositional Technique Reveals Microbial Pert
 }
 ```
 
-
 ## Citation for Phylogenetic RPCA
 
 ```
@@ -137,20 +135,40 @@ Martino, C. et al. A Novel Sparse Compositional Technique Reveals Microbial Pert
 ## Citation for TEMPTED
 
 ```
-Shi, p. et al. Time-Informed Dimensionality Reduction for Longitudinal Microbiome Studies. bioRxiv, (2023)
+Shi, p. et al. TEMPTED: time-informed dimensionality reduction for longitudinal microbiome studies. Genome Biology, (2024)
 ```
 
 ```
-@ARTICLE{Shi2023,
+@ARTICLE{Shi2024,
   author = {Shi, Pixu and Martino, Cameron and Han, Rungang and Janssen,
             Stefan and Buck, Gregory and Serrano, Myrna and Owzar, Kouros and
             Knight, Rob and Shenhav, Liat and Zhang, Anru R},
-  title = {{Time-Informed} Dimensionality Reduction for Longitudinal
-           Microbiome Studies},
-  year =  {2023},
-  doi = {10.1101/2023.07.26.550749},
-  URL = {https://www.biorxiv.org/content/10.1101/2023.07.26.550749v1},
-  journal = {bioRxiv},
+  title = {TEMPTED: time-informed dimensionality reduction for longitudinal microbiome studies},
+  year =  {2024},
+  doi = {10.1186/s13059-024-03453-x},
+  URL = {https://link.springer.com/article/10.1186/s13059-024-03453-x},
+  journal = {Genome Biology},
+}
+```
+
+## Citation for Joint-RPCA
+
+```
+Cordazzo Vargas, b. et al. Joint-RPCA: Domain-Aware Multi-Omics Integration for Systems Microbiology. Molecular Systems Biology, (2026)
+```
+
+```
+@ARTICLE{CordazzoVargas2026,
+  author = {Cordazzo Vargas, Bianca and Martino, Cameron and Dilmore, Amanda Hazel and Metcalf,
+            Jessica and Burcham, Zachary and Lahti, Leo and Bektanov, Aituar and Borman, Tuomas and
+            Salomaa, Veikko and Niiranen, Teemu and Havulinna, Aki and Gregor, Rachel and Eyal, Stav
+            and Meijler, Michael and Mizrahi, Itzhak and Song, Se and Bartko, Andrew and Dorrestein, Pieter
+            and Morton, James and McDonald, Daniel and Knight, Rob and Shenhav, Liat},
+  title = {Joint-RPCA: Domain-Aware Multi-Omics Integration for Systems Microbiology},
+  year =  {2026},
+  doi = {},
+  URL = {},
+  journal = {Molecular Systems Biology},
 }
 ```
 

gemelli/__init__.py

@@ -6,4 +6,4 @@
 # The full license is in the file COPYING.txt, distributed with this software.
 # ----------------------------------------------------------------------------
 
-__version__ = "0.0.12"
+__version__ = "0.0.13"

gemelli/factorization.py

@@ -183,7 +183,7 @@ def _fit(self):
             raise ValueError('Input data is should be type numpy.ndarray')
         # ensure the data contains missing values.
         # other methods would be better in the case of fully dense data
-        n_entries = np.product(tensor.shape)
+        n_entries = np.prod(tensor.shape)
         if self.check_dense:
             if (np.count_nonzero(tensor) == n_entries and
                     np.count_nonzero(~np.isnan(tensor)) == n_entries):

gemelli/preprocessing.py

@@ -21,7 +21,14 @@
 from inspect import getfullargspec
 from gemelli._defaults import DEFAULT_MTD
 from skbio.stats.composition import clr
-from skbio.diversity._util import _vectorize_counts_and_tree
+try:
+    # scikit-bio newer versions
+    from skbio.diversity._util import (
+        vectorize_counts_and_tree as _vectorize_counts_and_tree,
+    )
+except ImportError:
+    # scikit-bio older versions
+    from skbio.diversity._util import _vectorize_counts_and_tree
 from bp import parse_newick, to_skbio_treenode
 from scipy.sparse.linalg import svds
 # import QIIME2 if in a Q2env otherwise set type to str
@@ -30,8 +37,6 @@
 except ImportError:
     # python does not check but technically this is the type
     NewickFormat = str
-
-
 VALID_TAXONOMY_COLUMN_NAMES = ('taxon', 'taxonomy')
 
 
@@ -364,7 +369,7 @@ def tensor_rclr(T, branch_lengths=None):
                           + [1] + conditions_index[:-1])
         # transpose to flatten
         T = T.transpose(forward_T)
-        M = T.reshape(np.product(T.shape[:len(T.shape) - 1]),
+        M = T.reshape(np.prod(T.shape[:len(T.shape) - 1]),
                       T.shape[-1])
         with np.errstate(divide='ignore', invalid='ignore'):
             M_tensor_rclr = matrix_rclr(M, branch_lengths=branch_lengths)

gemelli/q2/_visualizer.py

@@ -9,13 +9,20 @@
 from urllib.parse import quote
 from skbio import DistanceMatrix
 from gemelli.utils import qc_rarefaction as _qc_rarefaction
-import pkg_resources
 import seaborn as sns
 import itertools
 import matplotlib.pyplot as plt
-
-TEMPLATES = pkg_resources.resource_filename('gemelli', 'q2')
-
+# fix old and new dependency chnages
+try:
+    from importlib.resources import files
+    TEMPLATES = files("gemelli").joinpath("q2")
+except ImportError:
+    try:
+        from importlib_resources import files
+        TEMPLATES = files("gemelli").joinpath("q2")
+    except ImportError:
+        import pkg_resources
+        TEMPLATES = pkg_resources.resource_filename("gemelli", "q2")
 
 def qc_rarefy(output_dir: str,
               table: biom.Table,

gemelli/q2/tests/test_rpca_method.py

@@ -83,6 +83,11 @@ def create_pos_cntrl_test_table(feature_prefix='',
 class Testqc(unittest.TestCase):
 
     def setUp(self):
+        # The positive/negative control tables are built with Subsample(),
+        # which draws from the global numpy RNG. Seed it so the controls are
+        # deterministic and the significance checks below are reproducible on
+        # CI (otherwise a borderline random draw can flip the p-value).
+        np.random.seed(0)
         btneg, btpos = create_pos_cntrl_test_table()
         self.btneg = Artifact.import_data("FeatureTable[Frequency]",
                                           btneg)

gemelli/rpca.py

@@ -931,10 +931,18 @@ def transform(ordination: OrdinationResults,
                          ' the features in the ordination.'
                          ' Either set subset_tables to True or'
                          ' match the tables to the ordination.')
-    ordination.samples = transform_helper(Udf,
-                                          Vdf,
-                                          s_eig,
-                                          rclr_table_df)
+    projected_samples = transform_helper(Udf,
+                                         Vdf,
+                                         s_eig,
+                                         rclr_table_df)
+    ordination.samples = projected_samples
+    # In skbio >= 0.7, sample_ids is stored as a separate attribute at
+    # construction and is not re-derived when .samples is reassigned, so it
+    # must be updated explicitly to match the projected samples. Older skbio
+    # versions derive the ids from the .samples index and have no such
+    # attribute, so this is skipped there (back-compatible).
+    if hasattr(ordination, 'sample_ids'):
+        ordination.sample_ids = list(projected_samples.index)
     return ordination
 
 

gemelli/tempted.py

@@ -780,8 +780,9 @@ def tempted_helper(individual_id_tables,
         individual_loadings.iloc[:, s] = a_hat
         feature_loadings.iloc[:, s] = b_hat
         state_loadings.iloc[:, s] = phi_hat.flatten()
-        lambda_coeff[s] = lambda_coeff_
-        rsquared[s] = 1 - resid / (y.size * y.var())
+        lambda_coeff[s] = np.asarray(lambda_coeff_).squeeze().item()
+        rsquared_ = 1 - resid / (y.size * y.var())
+        rsquared[s] = np.asarray(rsquared_).squeeze()
         # update data
         for i, (individual_id, m) in enumerate(tables_update.items()):
             temp = tipos[i]

setup.py

@@ -110,6 +110,8 @@ def run(self):
           'biom-format',
           'h5py',
           'iow',
+          'matplotlib',
+          'seaborn',
           'tax2tree'],
       classifiers=classifiers,
       entry_points={'qiime2.plugins': q2cmds,