🔥 Remove deprecated kedro_mlflow.framework.context and cancel depreca…

…tion of kedro_mlflow.io.metrics.MlflowMetricsDataSet for upcoming 0.8.0 release

CHANGELOG.md

@@ -2,15 +2,23 @@
 
 ## [Unreleased]
 
-### Fixed
+### Added
 
-- :bug: ``KedroMlflowConfig.setup()`` methods now sets the experiment globally to ensure all runs are launched under the experiment specified in the configuraiton even in interactive mode ([#256](https://github.com/Galileo-Galilei/kedro-mlflow/issues/256)).
+- :memo: Format code blocks in documentation with ``blacken-docs``
+- :construction_worker: Enforce the use of ``black`` and ``isort`` in the CI to enforce style guidelines for developers
 
 ### Changed
 
 - :sparkles: :boom: The ``KedroPipelineModel`` custom mlflow model now accepts any kedro `Pipeline` as input (provided they have a single DataFrame input and a single output because this is an mlflow limitation) instead of only ``PipelineML`` objects. This simplifies the API for user who want to customise the model logging ([#171](https://github.com/Galileo-Galilei/kedro-mlflow/issues/171)). `KedroPipelineModel.__init__` argument `pipeline_ml` is renamed `pipeline` to reflect this change.
-- :memo: Format code blocks in documentation with ``blacken-docs``
-- :construction_worker: Enforce the use of ``black`` and ``isort`` in the CI to enforce style guidelines
+- :wastebasket: `kedro_mlflow.io.metrics.MlflowMetricsDataSet` is no longer deprecated because there is no alternative for now to log many metrics at the same time.
+
+### Fixed
+
+- :bug: ``KedroMlflowConfig.setup()`` methods now sets the experiment globally to ensure all runs are launched under the experiment specified in the configuraiton even in interactive mode ([#256](https://github.com/Galileo-Galilei/kedro-mlflow/issues/256)).
+
+### Removed
+
+- :fire: :boom: ``KedroMlflowConfig`` and `get_mlflow_config` were deprecated since ``0.7.3`` and are now removed from ``kedro_mlflow.framework.context``. Direct import must now use `kedro_mlflow.config`.
 
 ## [0.7.6] - 2021-10-08
 

docs/source/04_experimentation_tracking/05_version_metrics.md

@@ -9,7 +9,7 @@ MLflow defines a metric as "a (key, value) pair, where the value is numeric". Ea
 `kedro-mlflow` introduces 3 ``AbstractDataSet`` to manage metrics:
 - ``MlflowMetricDataSet`` which can log a float as a metric
 - ``MlflowMetricHistoryDataSet`` which can log the evolution over time of a given metric, e.g. a list or a dict of float.
-- ``MlflowMetricsDataSet`` (**DEPRECATED**). It is a wrapper around a dictionary with metrics which is returned by node and log metrics in MLflow.
+- ``MlflowMetricsDataSet``. It is a wrapper around a dictionary with metrics which is returned by node and log metrics in MLflow.
 
 ### Saving a single float as a metric with ``MlflowMetricDataSet``
 
@@ -20,9 +20,11 @@ The ``MlflowMetricDataSet`` is an ``AbstractDataSet`` which enable to save or lo
 ```python
 from kedro_mlflow.io.metrics import MlflowMetricDataSet
 
-metric_ds=MlflowMetricDataSet(key="my_metric")
+metric_ds = MlflowMetricDataSet(key="my_metric")
 with mlflow.start_run():
-    metric_ds.save(0.3) # create a "my_metric=0.3" value in the "metric" field in mlflow UI
+    metric_ds.save(
+        0.3
+    )  # create a "my_metric=0.3" value in the "metric" field in mlflow UI
 ```
 
 ```eval_rst
@@ -34,24 +36,28 @@ with mlflow.start_run():
 ```python
 from kedro_mlflow.io.metrics import MlflowMetricDataSet
 
-metric_ds=MlflowMetricDataSet(key="my_metric", run_id="123456789")
+metric_ds = MlflowMetricDataSet(key="my_metric", run_id="123456789")
 with mlflow.start_run():
-    metric_ds.save(0.3) # create a "my_metric=0.3" value in the "metric" field of the run 123456789
+    metric_ds.save(
+        0.3
+    )  # create a "my_metric=0.3" value in the "metric" field of the run 123456789
 ```
 
 It is also possible to pass ``load_args`` and ``save_args`` to control which step should be logged (in case you have logged several step for the same metric.) ``save_args`` accepts a ``mode`` key which can be set to ``overwrite`` (mlflow default) or ``append``. In append mode, if no step is specified, saving the metric will "bump" the last existing step to create a linear history. **This is very useful if you have a monitoring pipeline which calculates a metric frequently to check the performance of a deployed model.**
 
 ```python
 from kedro_mlflow.io.metrics import MlflowMetricDataSet
 
-metric_ds=MlflowMetricDataSet(key="my_metric", load_args={"step": 1}, save_args={"mode": "append"})
+metric_ds = MlflowMetricDataSet(
+    key="my_metric", load_args={"step": 1}, save_args={"mode": "append"}
+)
 
 with mlflow.start_run():
-    metric_ds.save(0) # step 0 stored for "my_metric"
-    metric_ds.save(0.1) # step 1 stored for "my_metric"
-    metric_ds.save(0.2) # step 2 stored for "my_metric"
+    metric_ds.save(0)  # step 0 stored for "my_metric"
+    metric_ds.save(0.1)  # step 1 stored for "my_metric"
+    metric_ds.save(0.2)  # step 2 stored for "my_metric"
 
-    my_metric=metric_ds.load() # value=0.1 (step number 1)
+    my_metric = metric_ds.load()  # value=0.1 (step number 1)
 ```
 
 Since it is an ``AbstractDataSet``, it can be used with the YAML API in your ``catalog.yml``, e.g. :
@@ -78,48 +84,55 @@ It enables logging either:
 ```python
 from kedro_mlflow.io.metrics import MlflowMetricHistoryDataSet
 
-metric_history_ds=MlflowMetricDataSet(key="my_metric", save_args={"mode": "list"})
+metric_history_ds = MlflowMetricDataSet(key="my_metric", save_args={"mode": "list"})
 
 with mlflow.start_run():
-    metric_history_ds.save([0.1,0.2,0.3]) # will be logged with incremental steps
+    metric_history_ds.save([0.1, 0.2, 0.3])  # will be logged with incremental steps
 ```
   - a dict of {step: value} as a metric:
 
 ```python
 from kedro_mlflow.io.metrics import MlflowMetricHistoryDataSet
 
-metric_history_ds=MlflowMetricDataSet(key="my_metric", save_args={"mode": "dict"})
+metric_history_ds = MlflowMetricDataSet(key="my_metric", save_args={"mode": "dict"})
 
 with mlflow.start_run():
-    metric_history_ds.save({0: 0.1, 1: 0.2, 2: 0.3}) # will be logged with incremental steps
+    metric_history_ds.save(
+        {0: 0.1, 1: 0.2, 2: 0.3}
+    )  # will be logged with incremental steps
 ```
 
   - a list of dict [{log_metric_arg: value}] as a metric, e.g:
 
 ```python
 from kedro_mlflow.io.metrics import MlflowMetricHistoryDataSet
 
-metric_history_ds=MlflowMetricDataSet(key="my_metric", save_args={"mode": "history"})
+metric_history_ds = MlflowMetricDataSet(key="my_metric", save_args={"mode": "history"})
 
 with mlflow.start_run():
     metric_history_ds.save(
         [
             {"step": 0, "value": 0.1, "timestamp": 1345545},
             {"step": 1, "value": 0.2, "timestamp": 1345546},
             {"step": 2, "value": 0.3, "timestamp": 1345547},
-        ])
+        ]
+    )
 ```
 
 You can combine the different mode for save and load, e.g:
 
 ```python
 from kedro_mlflow.io.metrics import MlflowMetricHistoryDataSet
 
-metric_history_ds=MlflowMetricDataSet(key="my_metric", save_args={"mode": "dict"}, save_args={"mode": "list"})
+metric_history_ds = MlflowMetricDataSet(
+    key="my_metric", save_args={"mode": "dict"}, save_args={"mode": "list"}
+)
 
 with mlflow.start_run():
-    metric_history_ds.save({0: 0.1, 1: 0.2, 2: 0.3}) # will be logged with incremental steps
-metric_history_ds.load() # return [0.1,0.2,0.3]
+    metric_history_ds.save(
+        {0: 0.1, 1: 0.2, 2: 0.3}
+    )  # will be logged with incremental steps
+metric_history_ds.load()  # return [0.1,0.2,0.3]
 ```
 
 As usual, since it is an ``AbstractDataSet``, it can be used with the YAML API in your ``catalog.yml``, and in this case, the ``key`` argument is optional:
@@ -137,11 +150,6 @@ my_model_metric:
 
 ### Saving several metrics with their entire history with ``MlflowMetricsDataSet``
 
-```eval_rst
-.. warning:: This class is deprecated and will be removed soon. Use ``MlflowMetricDataSet`` or ``MlflowMetricHistoryDataSet`` instead.
-```
-
-
 Since it is an ``AbstractDataSet``, it can be used with the YAML API. You can define it in your ``catalog.yml`` as:
 
 ```yaml
@@ -167,7 +175,7 @@ Let assume that you have node which doesn't have any inputs and returns dictiona
 def metrics_node() -> Dict[str, Union[float, List[float]]]:
     return {
         "metric1": {"value": 1.1, "step": 1},
-        "metric2": [{"value": 1.1, "step": 1}, {"value": 1.2, "step": 2}]
+        "metric2": [{"value": 1.1, "step": 1}, {"value": 1.2, "step": 2}],
     }
 ```
 
@@ -199,10 +207,12 @@ As any entry in the catalog, the metrics data set must be defined in a Kedro pip
 
 ```python
 def create_pipeline() -> Pipeline:
-    return Pipeline(node(
-        func=metrics_node,
-        inputs=None,
-        outputs="my_model_metrics",
-        name="log_metrics",
-    ))
+    return Pipeline(
+        node(
+            func=metrics_node,
+            inputs=None,
+            outputs="my_model_metrics",
+            name="log_metrics",
+        )
+    )
 ```