Editable pipeline

q2_boots/__init__.py

@@ -12,15 +12,15 @@
                     alpha_average)
 from ._version import get_versions
 from .beta import (beta, get_medoid, beta_collection, per_cell_average, beta_average)
-from .core_metrics import (core_metrics)
+from .diversity_metrics import diversity_metrics
 from . import _version
 
 __all__ = ['_bootstrap_iteration', 'resample',
            'alpha',
            'alpha_collection',
            'beta',
-           'core_metrics_alpha_bootstrap',
-           'core_metrics',
+           'diversity_metrics_alpha_bootstrap',
+           'diversity_metrics',
            'get_medoid',
            'per_cell_average',
            'beta_collection',

q2_boots/diversity_metrics.py

@@ -0,0 +1,140 @@
+# ----------------------------------------------------------------------------
+# Copyright (c) 2023, QIIME 2 development team.
+#
+# Distributed under the terms of the Modified BSD License.
+#
+# The full license is in the file LICENSE, distributed with this software.
+# ----------------------------------------------------------------------------
+
+from q2_diversity._alpha import METRICS as ALPHA_METRICS
+from q2_diversity._beta import METRICS as BETA_METRICS
+
+
+def diversity_metrics(ctx, table, sampling_depth, metadata,
+                      n_jobs=1, phylogeny=None, n=100, alpha_method='median',
+                      beta_method='non-metric-median', with_replacement=True,
+                      random_seed=None, alpha_metrics=["observed_features",
+                                                       "shannon",
+                                                       "pielou_e",
+                                                       "faith_pd"],
+                      beta_metrics=["jaccard", "braycurtis",
+                                    "weighted_unifrac", "unweighted_unifrac"]):
+
+    bootstrap = ctx.get_action('boots', 'resample')
+    pcoa = ctx.get_action('diversity', 'pcoa')
+    emperor_plot = ctx.get_action('emperor', 'plot')
+    alpha_average = ctx.get_action('boots', 'alpha_average')
+    beta_average = ctx.get_action('boots', 'beta_average')
+
+    bootstrapped_tables, = bootstrap(table=table,
+                                     sampling_depth=sampling_depth,
+                                     n=n, with_replacement=with_replacement,
+                                     random_seed=random_seed)
+
+    tables = bootstrapped_tables.values()
+
+    alpha_metric_actions = {'PHYLO': [], 'NONPHYLO': []}
+
+    for metric in alpha_metrics:
+        if metric in ALPHA_METRICS["PHYLO"]["IMPL"] or metric in\
+           ALPHA_METRICS["PHYLO"]["UNIMPL"]:
+            metric = ALPHA_METRICS['NAME_TRANSLATIONS'][metric]
+            alpha_metric_actions['PHYLO'].append(ctx.get_action('diversity_lib',
+                                                                metric))
+        else:
+            if metric in ALPHA_METRICS['NONPHYLO']['IMPL']:
+                metric = ALPHA_METRICS['NAME_TRANSLATIONS'][metric]
+                alpha_metric_actions['NONPHYLO'].append(ctx.get_action('diversity_lib',
+                                                                       metric))
+            else:
+                alpha_metric_actions['NONPHYLO'].append(
+                    ctx.get_action('diversity_lib', 'alpha_passthrough'))
+
+    beta_metric_actions = {'PHYLO': [], 'NONPHYLO': []}
+
+    for metric in beta_metrics:
+        if metric in BETA_METRICS["PHYLO"]["IMPL"] or metric in\
+           BETA_METRICS["PHYLO"]["UNIMPL"]:
+            if metric in ('unweighted_unifrac', 'weighted_unifrac'):
+                metric = BETA_METRICS['NAME_TRANSLATIONS'][metric]
+                beta_metric_actions['PHYLO'].append(ctx.get_action('diversity_lib',
+                                                                   metric))
+            else:
+                # handle unimplemented unifracs
+                beta_metric_actions['PHYLO'].append(
+                    ctx.get_action('diversity_lib', 'beta_phylogenetic_passthrough'))
+        else:
+
+            if metric in BETA_METRICS['NONPHYLO']['IMPL']:
+                metric = BETA_METRICS['NAME_TRANSLATIONS'][metric]
+                beta_metric_actions['NONPHYLO'].append(ctx.get_action('diversity_lib',
+                                                                      metric))
+            else:
+                beta_metric_actions['NONPHYLO'].append(
+                    ctx.get_action('diversity_lib', 'beta_passthrough'))
+
+    if phylogeny is None and (len(alpha_metrics['PHYLO']) > 0 or
+                              len(beta_metrics['PHYLO']) > 0):
+        print("WARNING: NO PHYLOGENY PROVIDED, PHYLOGENETIC METRICS WILL NOT BE RUN")
+
+    alphas = {}
+    for m in (alpha_metrics['NONPHYLO']):
+        alpha = alpha_representative(m, tables, alpha_method)
+        res, = alpha_average(data=alpha, average_method=alpha_method)
+        alphas[m.__name__] = res
+    if phylogeny is not None:
+        for m in (alpha_metrics['PHYLO']):
+            alpha = alpha_representative(m, tables, alpha_method, phylogeny=phylogeny)
+            res, = alpha_average(data=alpha, average_method=alpha_method)
+            alphas[m.__name__] = res
+
+    dms = {}
+    for m in (beta_metrics['NONPHYLO']):
+        beta_results = beta_representative(m, tables, beta_method)
+        res, = beta_average(data=beta_results, representative=beta_method)
+        dms[m.__name__] = res
+    if phylogeny is not None:
+        for m in (beta_metrics['PHYLO']):
+            beta_results = beta_representative(m, tables, beta_method, phylogeny,
+                                               n_threads=n_jobs)
+            beta_results, = beta_average(data=beta_results,
+                                         representative=beta_method)
+            dms[m.__name__] = beta_results
+
+    pcoas = {}
+    for key, dm in dms.items():
+        pcoa_results, = pcoa(distance_matrix=dm)
+        pcoas[key] = pcoa_results
+
+    visualizations = {}
+    for key, pcoa in pcoas.items():
+        visualizations[key] = (emperor_plot(pcoa=pcoa, metadata=metadata)[0])
+
+    return bootstrapped_tables, alphas, dms, pcoas, visualizations
+
+
+def beta_representative(func, tables, method,
+                        phylogeny=None, n_threads=1):
+    metric = []
+    if phylogeny is None:
+        for table in tables:
+            metric.append(func(table=table)[0])
+    else:
+        for table in tables:
+            metric.append(func(table=table, phylogeny=phylogeny,
+                               threads=n_threads)[0])
+
+    return metric
+
+
+def alpha_representative(func, tables, method, phylogeny=None):
+
+    alpha = []
+    if phylogeny is None:
+        for table in tables:
+            alpha.append(func(table=table)[0])
+    else:
+        for table in tables:
+            alpha.append(func(table=table, phylogeny=phylogeny)[0])
+
+    return alpha

q2_boots/plugin_setup.py

@@ -7,7 +7,7 @@
 # ----------------------------------------------------------------------------
 
 from qiime2.plugin import (Plugin, Int, Range, Collection,
-                           Citations, Str, Choices, Bool, Float,
+                           Citations, Str, Choices, Bool, Float, List,
                            Metadata, Visualization)
 
 from q2_types.feature_table import (
@@ -280,7 +280,7 @@
 )
 
 plugin.pipelines.register_function(
-    function=q2_boots.core_metrics,
+    function=q2_boots.diversity_metrics,
     inputs={
         'table': FeatureTable[Frequency | RelativeFrequency |
                               PresenceAbsence],
@@ -296,7 +296,15 @@
                                      'non-metric-median',
                                      'medoid'),
         'with_replacement': Bool,
-        'random_seed': Int
+        'random_seed': Int,
+        'alpha_metrics' : List[Str % Choices(alpha_metrics['NONPHYLO']['IMPL'] |
+                                             alpha_metrics['NONPHYLO']['UNIMPL'] |
+                                             alpha_metrics['PHYLO']['IMPL'] |
+                                             alpha_metrics['PHYLO']['UNIMPL'])],
+        'beta_metrics': List[Str % Choices(beta_metrics['NONPHYLO']['IMPL'] |
+                                           beta_metrics['NONPHYLO']['UNIMPL'] |
+                                           beta_metrics['PHYLO']['IMPL'] |
+                                           beta_metrics['PHYLO']['UNIMPL'])],
     },
     outputs=[
         ('rarefied_table', Collection[FeatureTable[Frequency]]),
@@ -314,6 +322,6 @@
     output_descriptions={
 
     },
-    name='Core Metrics',
+    name='Diversity Metrics',
     description='Bootstrapped Core Metrics'
 )

q2_boots/tests/test_core_metrics.py → q2_boots/tests/test_diversity_metrics.py

@@ -16,13 +16,13 @@
 from io import StringIO
 
 
-class TestCoreMetrics(TestPluginBase):
+class TestDiversityMetrics(TestPluginBase):
 
     package = 'q2_boots'
 
     def setUp(self):
         super().setUp()
-        self.core_metrics = self.plugin.pipelines['core_metrics']
+        self.diversity_metrics = self.plugin.pipelines['diversity_metrics']
 
     def test_basic(self):
         t = Table(data=np.array([[1000, 1920, 3451], [4536, 1552, 6521],
@@ -35,7 +35,7 @@ def test_basic(self):
                                 columns=['blank'])
         metadata.index.name = 'sample-id'
         metadata = Metadata(metadata)
-        output = self.core_metrics(table=t,
+        output = self.diversity_metrics(table=t,
                                    sampling_depth=500,
                                    metadata=metadata,
                                    n_jobs=1,
@@ -66,7 +66,7 @@ def test_phylogeny(self):
                   sample_ids=['S1', 'S2', 'S3'],
                   observation_ids=['O1', 'O2', 'O3'])
         t = Artifact.import_data('FeatureTable[Frequency]', t)
-        output = self.core_metrics(table=t,
+        output = self.diversity_metrics(table=t,
                                    sampling_depth=500,
                                    metadata=metadata,
                                    n_jobs=1,
@@ -79,3 +79,37 @@ def test_phylogeny(self):
         for table in output[0].values():
             table = Artifact.view(table, pd.DataFrame)
             self.assertEqual(table.shape, (3, 3))
+
+    def test_custom_pipeline(self):
+        with StringIO('(O1:0.3, O2:0.2, O3:0.1, O4:0.2)root;') as f:
+            phylogeny = skbio.read(f, format='newick', into=skbio.TreeNode)
+
+        phylogeny = Artifact.import_data(
+            'Phylogeny[Rooted]',
+            phylogeny
+        )
+        metadata = pd.DataFrame(['not', 'of', 'interest'],
+                                index=['S1', 'S2', 'S3'],
+                                columns=['blank'])
+        metadata.index.name = 'sample-id'
+        metadata = Metadata(metadata)
+        t = Table(data=np.array([[1000, 1920, 3451], [2536, 1552, 1521],
+                                 [1634, 1634, 6721]]),
+                  sample_ids=['S1', 'S2', 'S3'],
+                  observation_ids=['O1', 'O2', 'O3'])
+        t = Artifact.import_data('FeatureTable[Frequency]', t)
+        output = self.diversity_metrics(table=t,
+                                   sampling_depth=500,
+                                   metadata=metadata,
+                                   n_jobs=1,
+                                   n=10,
+                                   phylogeny=phylogeny,
+                                   with_replacement=True,
+                                   alpha_metrics=["pielou_e", "faith_pd"],
+                                   beta_metrics=["jaccard", "unweighted_unifrac"]
+                                   )
+        self.assertEqual(len(output[0]), 10)
+
+        for table in output[0].values():
+            table = Artifact.view(table, pd.DataFrame)
+            self.assertEqual(table.shape, (3, 3))