Load Metadata and State of Health Data

Metadata and state of health are key pieces of information to understand how a station is faring. In this example we examine how to obtain the metadata and health sensor data from the DataWindow.

Running the example

The first step is to load RedVox data into a DataWindow.

from redvox.common.data_window import DataWindow

# Input Directory
input_dir = "path/to/redvox/data/dw_1648830257000498_2.pkl.lz4"

# Load data window from report
dw = DataWindow.deserialize(input_dir)

Let’s take a look at the metadata for each Station.

from redvox.common.data_window import DataWindow

# Input Directory
input_dir = "path/to/redvox/data/dw_1648830257000498_2.pkl.lz4"

# Load data window from report
dw = DataWindow.deserialize(input_dir)

for station in dw.stations():
        # Print some metadata
        print(f"Station {station.id()} Metadata:\n"
              f"station description: {station.metadata().station_description}\n"
              f"is station private: {station.metadata().is_private}\n"
              f"station make: {station.metadata().make}\n"
              f"station model: {station.metadata().model}\n"
              f"station os: {station.metadata().os}\n"
              f"station os version: {station.metadata().os_version}\n"
              f"redvox app version:  {station.metadata().app_version}\n"
              f"redvox packet duration in seconds: {station.metadata().packet_duration_s}\n")

This is just a subset of the available metadata. For a complete list, visit the Station Metadata Documentation.

Let’s take a look at state of health sensor. In this example, we will extract the power state, battery charge % and internal temperature. The complete list of available state of health can be found in the Health Sensor Documentation .

from redvox.common.data_window import DataWindow
import matplotlib.pyplot as plt

# Input Directory
input_dir = "path/to/redvox/data/dw_1648830257000498_2.pkl.lz4"

# Load data window from report
dw = DataWindow.deserialize(input_dir)

for station in dw.stations():
        # Check that there is state of health data
        if station.has_health_sensor():

            # Get some state of health data - example of what is available
            power_state = station.health_sensor().get_power_state_data()
            battery_charge_percentage = station.health_sensor().get_battery_charge_remaining_data()
            internal_temp_c = station.health_sensor().get_internal_temp_c_data()

            # The channels share the same timestamps
            health_time_micros = station.health_sensor().data_timestamps() - station.health_sensor().first_data_timestamp()
            health_time_s = health_time_micros*1E-6

We can also plot the state of health using the Matplotlib library.

from redvox.common.data_window import DataWindow
import matplotlib.pyplot as plt

# Input Directory
input_dir = "path/to/redvox/data/dw_1648830257000498_2.pkl.lz4"

# Load data window from report
dw = DataWindow.deserialize(input_dir)

for station in dw.stations():
        # Check that there is state of health data
        if station.has_health_sensor():

            # Get some state of health data - example of what is available
            power_state = station.health_sensor().get_power_state_data()
            battery_charge_percentage = station.health_sensor().get_battery_charge_remaining_data()
            internal_temp_c = station.health_sensor().get_internal_temp_c_data()

            # The channels share the same timestamps
            health_time_micros = station.health_sensor().data_timestamps() - station.health_sensor().first_data_timestamp()
            health_time_s = health_time_micros*1E-6

            # Plot the state of health data - one subplot per channel
            fig, ax = plt.subplots(nrows=3, ncols=1, sharex='col')

            ax[0].plot(health_time_s, power_state)
            ax[1].plot(health_time_s, battery_charge_percentage)
            ax[2].plot(health_time_s, internal_temp_c)

            # Set labels and subplot title
            ax[0].set_ylabel('Power State')
            ax[1].set_ylabel('Batt. %')
            ax[2].set_ylabel('Int. temp. [C]')
            ax[2].set_xlabel(f"Seconds from {int(dw.start_date()*1E-6)} Unix epoch UTC")
            plt.suptitle(f"RedVox Station ID {station.id()}")

plt.show()

Example output

After running the above code snippet, the following graph should appear:

For a more complete example on how to load the metadata and health sensor data, visit Github.

In the next section, we will take a look at plotting audio waveforms.