Handle warnings from third-party libraries

[marfox]

These warnings are duplicate: both logged and printed to stderr. logging.captureWarnings(True) seems to be enabled by default.

See https://docs.python.org/3.7/library/warnings.html#temporarily-suppressing-warnings

General

Update: this has been fixed in 6567781

• [x] Silence Using TensorFlow backend.: it gets printed whenever keras is imported.

Importer

• [x] To hit: python -m soweego importer import TARGET

WARNING: /usr/local/lib/python3.7/site-packages/pymysql/cursors.py:170: Warning: (124, 'InnoDB rebuilding table to add column FTS_DOC_ID')
  result = self._query(query)

Feature extraction

The following warnings are emitted during feature extraction, i.e., linker/workflow#extract_features.

• [x] ~~Silence FutureWarning from recordlinkage:~~ not hitting this anymore with recordlinkage 0.13

WARNING: /Users/focs/.local/share/virtualenvs/soweego-jENbvpMi/lib/python3.7/site-packages/recordlinkage/index.py:155: FutureWarning: the 'labels' keyword is deprecated, use 'codes' instead
  verify_integrity=False)

• [ ] silence the following RunTimeWarning, a recordlinkage known open issue: https://github.com/J535D165/recordlinkage/issues/56 UPDATE: the issue is now claimed to be closed inrecordlinkage 0.13, but we are hitting it anyway

WARNING: /Users/focs/soweego/soweego/linker/feature_extraction.py:122: RuntimeWarning: invalid value encountered in true_divide
  cosine = numpy.divide(ab, numpy.multiply(a, b)).A1

• [x] ~~investigate this, happens when using Naïve Bayes:~~ not hitting this anymore

WARNING: /Users/focs/.local/share/virtualenvs/soweego-jENbvpMi/lib/python3.7/site-packages/recordlinkage/algorithms/nb_sklearn.py:324: RuntimeWarning: divide by zero encountered in log
  self.class_log_prior_ = (np.log(self.class_count_) -

Neural network training

Update: both of these messages were silenced in 3a971d95e4a7e6c1387806a899a01cbb08d7f4c2

• [x] To hit: python -m soweego linker evaluate slp TARGET ENTITY and python -m soweego linker train slp TARGET ENTITY .

WARNING:tensorflow:From /usr/local/lib/python3.7/site-packages/tensorflow/python/framework/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.
Instructions for updating:
Colocations handled automatically by placer.

• [x] to hit: python -m soweego linker train slp TARGET ENTITY

WARNING:tensorflow:From /usr/local/lib/python3.7/site-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.cast instead.

Grid search

• [ ] To hit: python -m soweego linker evaluate svm TARGET ENTITY --nested

WARNING: /usr/local/lib/python3.7/site-packages/sklearn/svm/base.py:196: FutureWarning: The default value of gamma will change from 'auto' to 'scale' in version 0.22 to account better for unscaled features. Set gamma explicitly to 'auto' or 'scale' to avoid this warning.
  "avoid this warning.", FutureWarning)

• [ ] to hit: python -m soweego linker evaluate lsvm TARGET ENTITY --nested

WARNING: /usr/local/lib/python3.7/site-packages/sklearn/svm/base.py:931: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.
  "the number of iterations.", ConvergenceWarning)

MLens

• [x] [MLENS] backend: threading seems to get printed whenever mlens is imported. For instance, with python -m soweego. See http://ml-ensemble.com/docs/config.html for an eventual solution.

[marfox]

See https://github.com/J535D165/recordlinkage/pull/86