Graph of measured values (Metrics API) about sacredboard HOT 3 CLOSED

Data can be collected using the Metrics API:
https://github.com/chovanecm/sacred/blob/feature/data-logger/docs/collected_information.rst#metrics-api

Sacredboard shall display a chart for each associated metric that gets measured during the execution of the experiment run.
Metrics are mostly qualitative variables of the adaptive model, such as the value of the loss function, accuracy of correctly classified items, etcetera
Instead of running the training process for a fixed number of iterations, researchers may opt to engage techniques that measure the metrics after each (or e.g. after each 20th) training iteration to decide whether the training process should continue or if it should be terminated because of a low expectation of future progress.

A visual chart of a metric shall be a 2D graph with the training step or elapsed time on the x-axis and the measured scalar value on the y-axis.
Sacredboard shall only display one chart per metric per run, there should be only one curve in the graph. A run can have one to a few metrics.

from sacredboard.

Work in progress: https://github.com/chovanecm/sacredboard/tree/feature/plot-metrics

from sacredboard.

Scenario:

1. The user indicates he wants to see the progress of a specific metric or metrics.
2. The system responds with a list of available metrics
3. The user selects the metrics he wants to plot
4. The system plots the selected metrics with x-axis as the step and y-axis as value.

Alternate scenarios:

1. In step 3., the user chooses to set the x-axis as time instead of step number.
   1.1 The system draws the selected metrics with date-based x axis.

from sacredboard.