Dev/sklearn (#119)

* Links, docstrings and error fixes

* Added logging regression to sklearn integration

* Added all list

* Small fixes

* Refactor, added more visuals

* refactor

* Added classifier summary

* Dropped try catch clauses

* Added visuals to classifier

* Changed visuals

* Minor fixes

* Added clustering

* Updated dependencies

* moved private method down

* k matches kmeans api

* Added log csv to api

* Log csv, not html, fix in getting pred proba

* Added docstring, changed interface

* fix

* Refactor - log meta-data separately

* Refactor - log meta-data separately, added docstrings

* docs fixes

* disable module too long

* fix

* added model name option

* print exception msg

* added explicit parameter html or csv, made str formatting consistent

* docstring fix

* added number of rows option for logging csv

* fix

* Changed log_test_score to log_scores

docs/conf.py

@@ -44,7 +44,8 @@
                         'optuna',
                         'scikitplot',
                         'scikitplot.metrics',
-                        'xgboost']
+                        'xgboost',
+                        'yellowbrick']
 
 # -- Project information -----------------------------------------------------
 

docs/index.rst

@@ -55,14 +55,15 @@ And the best thing is you can extend it yourself or... tell us to do it for you
    api.chart <user_guide/api/chart>
    api.explainers <user_guide/api/explainers>
    api.html <user_guide/api/html>
-   api.table  <user_guide/api/table>
+   api.table <user_guide/api/table>
    api.utils <user_guide/api/utils>
    api.video <user_guide/api/video>
    hpo.utils <user_guide/hpo/utils>
    bots.telegram_bot <user_guide/bots/telegram_bot>
    monitoring.keras <user_guide/monitoring/keras>
    monitoring.lightgbm <user_guide/monitoring/lightgbm>
    monitoring.xgboost <user_guide/monitoring/xgboost>
+   monitoring.sklearn <user_guide/monitoring/sklearn>
    monitoring.fastai <user_guide/monitoring/fastai>
    monitoring.metrics <user_guide/monitoring/metrics>
    monitoring.fairness <user_guide/monitoring/fairness>

docs/user_guide/monitoring/sklearn.rst

@@ -0,0 +1,6 @@
+Scikit-learn
+============
+
+.. automodule:: neptunecontrib.monitoring.sklearn
+    :members:
+    :show-inheritance:

neptunecontrib/api/__init__.py

@@ -18,7 +18,7 @@
 from neptunecontrib.api.chart import log_chart
 from neptunecontrib.api.explainers import log_explainer, log_local_explanations, log_global_explanations
 from neptunecontrib.api.html import log_html
-from neptunecontrib.api.table import log_table
+from neptunecontrib.api.table import log_table, log_csv
 from neptunecontrib.api.utils import (
     concat_experiments_on_channel,
     extract_project_progress_info,
@@ -39,6 +39,7 @@
     'log_audio',
     'log_video',
     'log_table',
+    'log_csv',
     'log_html',
     'log_chart',
     'log_pickle',

neptunecontrib/api/html.py

@@ -29,11 +29,10 @@ def log_html(name, html, experiment=None):
     Args:
         name (:obj:`str`):
             | Name of the chart (without extension) that will be used as a part of artifact's destination.
-        html_body (:obj:`str`):
+        html (:obj:`str`):
             | HTML string that is logged and rendered as HTML.
         experiment (:obj:`neptune.experiments.Experiment`, optional, default is ``None``):
-            | For advanced users only. Pass Neptune
-              `Experiment <https://docs.neptune.ai/neptune-client/docs/experiment.html#neptune.experiments.Experiment>`_
+            | For advanced users only. Pass Neptune ``Experiment``
               object if you want to control to which experiment data is logged.
             | If ``None``, log to currently active, and most recent experiment.
 

neptunecontrib/api/table.py

@@ -13,13 +13,14 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 #
-
 import neptune
 
 __all__ = [
     'log_table',
+    'log_csv',
 ]
 
+
 def log_table(name, table, experiment=None):
     """Logs pandas dataframe to neptune.
 
@@ -62,13 +63,53 @@ def log_table(name, table, experiment=None):
      """
     _exp = experiment if experiment else neptune
 
-    _exp.log_artifact(export_pandas_dataframe(table), "tables/" + name + '.html')
+    _exp.log_artifact(export_pandas_dataframe(table, 'html'), 'tables/{}.html'.format(name))
+
+
+def log_csv(name, table, experiment=None):
+    """Logs pandas dataframe to neptune as csv file.
 
+    Pandas dataframe is converted to csv fie and logged to Neptune as an artifact with path csv/{name}.csv
+
+    Args:
+        name (:obj:`str`):
+            | Name of the file (without extension) that will be used as a part of csv's destination.
+        table (:obj:`pandas.Dataframe`):
+            | DataFrame table
+        experiment (:obj:`neptune.experiments.Experiment`, optional, default is ``None``):
+            | Neptune Experiment object if you want to control to which experiment you log the data.
+            | If ``None``, log to currently active, and most recent experiment.
+
+    Examples:
+        Create or load dataframe:
 
-def export_pandas_dataframe(table):
+        .. code:: python3
+
+            import pandas as pd
+            iris_df = pd.read_csv('https://raw.githubusercontent.com/mwaskom/seaborn-data/master/iris.csv', nrows=100)
+
+        Log it to Neptune:
+
+        .. code:: python3
+
+             from neptunecontrib.api import log_csv
+             log_csv('pandas_df', iris_df)
+     """
+    _exp = experiment if experiment else neptune
+
+    _exp.log_artifact(export_pandas_dataframe(table, 'csv'), 'csv/{}.csv'.format(name))
+
+
+def export_pandas_dataframe(table, target_type):
     from io import StringIO
 
-    buffer = StringIO(table.to_html())
+    if target_type == 'csv':
+        buffer = StringIO(table.to_csv())
+    elif target_type == 'html':
+        buffer = StringIO(table.to_html())
+    else:
+        ValueError('Unsupported format: {}'.format(target_type))
+
     buffer.seek(0)
 
     return buffer