Merge pull request #106 from gibsramen/quiet

Suppression of warnings and (optionally) output

Author: fedarko

scripts/songbird

@@ -1,15 +1,22 @@
 #!/usr/bin/env python3
 import os
-import tensorflow as tf
-from tensorboard.plugins.hparams import api as hp
 import pandas as pd
 import numpy as np
 from biom import load_table
 import click
 from skbio.stats.composition import clr, clr_inv
 from songbird.multinomial import MultRegression
-from songbird.util import read_metadata, match_and_filter, split_training
+from songbird.util import (read_metadata, match_and_filter, split_training,
+                           silence_output)
 from songbird.parameter_info import DESCS, DEFAULTS
+import warnings
+
+with warnings.catch_warnings():
+    warnings.filterwarnings("ignore", category=FutureWarning)
+    from tensorboard.plugins.hparams import api as hp
+    import tensorflow as tf
+
+warnings.filterwarnings("ignore", category=RuntimeWarning)
 
 
 @click.group()
@@ -106,6 +113,12 @@ def songbird():
     help=DESCS["random-seed"],
     type=int,
 )
+@click.option(
+    "--silent/--no-silent",
+    default=DEFAULTS["silent"],
+    show_default=True,
+    help=DESCS["silent"],
+)
 def multinomial(
     input_biom,
     metadata_file,
@@ -123,7 +136,11 @@ def multinomial(
     summary_interval,
     summary_dir,
     random_seed,
+    silent,
 ):
+    if silent:
+        silence_output()
+
     # load metadata and tables
     metadata = read_metadata(metadata_file)
     table = load_table(input_biom)
@@ -146,6 +163,7 @@ def multinomial(
                'learning_rate': learning_rate,
                'min_sample_count': min_sample_count,
                'min_feature_count': min_feature_count,
+               'silent': silent,
                }
     if random_seed is not None:
         hparams.update({
@@ -181,6 +199,7 @@ def multinomial(
             epochs=epochs,
             summary_interval=summary_interval,
             checkpoint_interval=checkpoint_interval,
+            silent=silent,
         )
 
         summary_writer = tf.contrib.summary.create_file_writer(summary_dir)

scripts/test_songbird_cli.py

@@ -22,6 +22,40 @@ def setUp(self) -> None:
     def tearDown(self) -> None:
         shutil.rmtree(self.path)
 
+    def test_cli_silent(self):
+        runner = CliRunner()
+        test_args = ['--input-biom', 'data/redsea/redsea.biom',
+                     '--metadata-file', 'data/redsea/redsea_metadata.txt',
+                     '--formula',
+                     'Depth+Temperature+Salinity+Oxygen+Fluorescence'
+                     '+Nitrate',
+                     '--epochs', '100',
+                     '--differential-prior', '0.5',
+                     '--summary-interval', '1',
+                     '--summary-dir', self.path,
+                     '--silent']
+
+        result = runner.invoke(songbird.multinomial, test_args)
+
+        assert result.output == ""
+
+    def test_cli_not_silent(self):
+        runner = CliRunner()
+        test_args = ['--input-biom', 'data/redsea/redsea.biom',
+                     '--metadata-file', 'data/redsea/redsea_metadata.txt',
+                     '--formula',
+                     'Depth+Temperature+Salinity+Oxygen+Fluorescence'
+                     '+Nitrate',
+                     '--epochs', '100',
+                     '--differential-prior', '0.5',
+                     '--summary-interval', '1',
+                     '--summary-dir', self.path,
+                     '--no-silent']
+
+        result = runner.invoke(songbird.multinomial, test_args)
+
+        assert result.output != ""
+
     def test_cli_no_seed_set(self):
         runner = CliRunner()
         test_args = ['--input-biom', 'data/redsea/redsea.biom',

songbird/__init__.py

@@ -1 +1,3 @@
+import warnings
+warnings.filterwarnings("ignore", category=FutureWarning)
 __version__ = "1.0.1"

songbird/multinomial.py

@@ -1,5 +1,5 @@
-import tensorflow as tf
 import os
+import tensorflow as tf
 from tensorflow.contrib.distributions import Multinomial, Normal
 from tqdm import tqdm
 import time
@@ -135,7 +135,8 @@ def __call__(self, session, trainX, trainY, testX, testY):
             self.writer = None
         tf.global_variables_initializer().run()
 
-    def fit(self, epochs=10, summary_interval=100, checkpoint_interval=3600):
+    def fit(self, epochs=10, summary_interval=100, checkpoint_interval=3600,
+            silent=False):
         """ Fits the model.
 
         Parameters
@@ -146,6 +147,10 @@ def fit(self, epochs=10, summary_interval=100, checkpoint_interval=3600):
            Number of seconds until a summary is recorded
         checkpoint_interval : int
            Number of seconds until a checkpoint is recorded
+        silent : bool
+           Flag denoting whether to suppress progress bar and TensorFlow
+           warnings. If True, show neither progress bar nor warnings. If False
+           (default), show both.
 
         Returns
         -------
@@ -165,7 +170,12 @@ def fit(self, epochs=10, summary_interval=100, checkpoint_interval=3600):
         cv = []
         iter_n = []
 
-        for i in tqdm(range(0, num_iter)):
+        if silent:
+            iter_range = range(0, num_iter)
+        else:
+            iter_range = tqdm(range(0, num_iter))
+
+        for i in iter_range:
             now = time.time()
 
             if now - last_summary_time > summary_interval:

songbird/parameter_info.py

@@ -36,6 +36,11 @@
         "for it to be included in the analysis."
     ),
     "summary-interval": "Number of seconds before storing a summary.",
+    "silent": (
+        'Flag denoting whether to hide progress bar and TensorFlow warnings. '
+        'If False, progress bar and warnings are displayed. If True, neither '
+        'progress bar nor warnings are displayed.'
+    ),
     # FYI: The following parameters are exclusive to the non-Q2 Songbird script
     "checkpoint-interval": 'Number of seconds before a saving a checkpoint.',
     "summary-dir": (
@@ -65,4 +70,5 @@
     "summary-interval": 10,
     "summary-dir": "summarydir",
     "random-seed": 0,
+    "silent": False,
 }

songbird/q2/_method.py

@@ -1,12 +1,12 @@
-import biom
 import skbio
 import qiime2
 import pandas as pd
 import numpy as np
+import biom
 import tensorflow as tf
 from skbio import OrdinationResults
 from songbird.multinomial import MultRegression
-from songbird.util import match_and_filter, split_training
+from songbird.util import match_and_filter, split_training, silence_output
 from songbird.parameter_info import DEFAULTS
 from qiime2.plugin import Metadata
 
@@ -27,10 +27,14 @@ def multinomial(table: biom.Table,
                 min_feature_count: int = DEFAULTS["min-feature-count"],
                 summary_interval: int = DEFAULTS["summary-interval"],
                 random_seed: int = DEFAULTS["random-seed"],
+                silent: bool = DEFAULTS["silent"],
                 ) -> (
                     pd.DataFrame, qiime2.Metadata, skbio.OrdinationResults
                 ):
 
+    if silent:
+        silence_output()
+
     # load metadata and tables
     metadata = metadata.to_dataframe()
     # match them
@@ -53,14 +57,16 @@ def multinomial(table: biom.Table,
                            beta_mean=differential_prior,
                            batch_size=batch_size,
                            save_path=None)
+
     with tf.Graph().as_default(), tf.Session() as session:
         tf.set_random_seed(random_seed)
         model(session, trainX, trainY, testX, testY)
 
         loss, cv, its = model.fit(
             epochs=epochs,
             summary_interval=summary_interval,
-            checkpoint_interval=None)
+            checkpoint_interval=None,
+            silent=silent)
 
     md_ids = np.array(design.columns)
     obs_ids = table.ids(axis='observation')

songbird/q2/plugin_setup.py

@@ -10,7 +10,7 @@
 import qiime2.sdk
 from songbird import __version__
 
-from qiime2.plugin import (Str, Properties, Int, Float,  Metadata)
+from qiime2.plugin import (Str, Properties, Int, Float,  Metadata, Bool)
 from q2_types.feature_table import FeatureTable, Frequency
 from q2_types.ordination import PCoAResults
 from q2_types.sample_data import SampleData
@@ -51,6 +51,7 @@
         'min_feature_count': Int,
         'summary_interval': Float,
         'random_seed': Int,
+        'silent': Bool,
     },
     outputs=[
         ('differentials', FeatureData[Differential]),
@@ -81,6 +82,9 @@
         "min_feature_count": DESCS["min-feature-count"],
         "summary_interval": DESCS["summary-interval"],
         "random_seed": DESCS["random-seed"],
+        "silent": (DESCS["silent"] + " (Only has an impact when using this "
+                   "command with the --verbose option or through the Qiime2 "
+                   "Artifact API)"),
     },
     name='Multinomial regression',
     description=("Performs multinomial regression and calculates "

songbird/q2/tests/test_method.py

@@ -1,6 +1,8 @@
 import qiime2
 import unittest
 import numpy as np
+import contextlib
+import io
 import tensorflow as tf
 from songbird.q2._method import multinomial
 from songbird.util import random_multinomial_model
@@ -27,6 +29,21 @@ def setUp(self):
 
         self.table, self.md, self.beta = res
 
+    def _silent_helper(self, silent):
+        tf.set_random_seed(0)
+        md = self.md
+
+        md.name = 'sampleid'
+        md = qiime2.Metadata(md)
+
+        f = io.StringIO()
+        with contextlib.redirect_stderr(f):
+            res_beta, res_stats, res_biplot = multinomial(
+                table=self.table, metadata=md,
+                min_sample_count=0, min_feature_count=0,
+                formula="X", epochs=1000, silent=silent)
+        return f
+
     def test_fit(self):
         tf.set_random_seed(0)
         md = self.md
@@ -51,6 +68,14 @@ def test_fit(self):
         npt.assert_allclose(exp_beta, res_beta.T, atol=0.6, rtol=0.6)
         self.assertGreater(len(res_stats.to_dataframe().index), 1)
 
+    def test_silent(self):
+        f = self._silent_helper(silent=True)
+        assert f.getvalue() == ""
+
+    def test_not_silent(self):
+        f = self._silent_helper(silent=False)
+        assert f.getvalue() != ""
+
     def test_fit_consistency(self):
         md = self.md
 

songbird/util.py

@@ -1,3 +1,5 @@
+import os
+import tensorflow as tf
 import numpy as np
 import pandas as pd
 from sklearn.utils import check_random_state
@@ -195,3 +197,15 @@ def split_training(dense_table, metadata, design, training_column=None,
     testY = dense_table.loc[~train_idx].values
 
     return trainX, testX, trainY, testY
+
+
+def silence_output():
+    # suppress profiling messages & compilation warnings
+    # taken from:
+    # https://stackoverflow.com/questions/47068709/your-cpu-supports-
+    #     instructions-that-this-tensorflow-binary-was-not-compiled-to-u
+    os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
+
+    # suppress deprecation warnings
+    # taken from https://github.com/tensorflow/tensorflow/issues/27023
+    tf.compat.v1.logging.set_verbosity(tf.compat.v1.logging.ERROR)