30 created a pattern for future data loaders and an option for using pre-loaded dataframes.
This exposes that functionality to be passed in at setup.
Description of changes
Add two optional paramaters to run setup.
Starting Functional Tests:
Run the example notebook and verify that old pattern works without change.
Turn the urllib.requests cell into a raw or markdown cell (this assumes that the dataset has been pulled locally from the first run)
Add something like the following to a first cell or two :
import seismometer.data.loader as dload
import seismometer.configuration
import seismometer
# Load data from files without calling Seismogram
config = seismometer.configuration.ConfigProvider('.')
loader = dload.loader_factory(config)
data = loader.load_data()
# Now pass that loaded frame through
import pandas as pd
import seismometer as sm
sm.run_startup(config_path='.', predictions_frame=data, events_frame=pd.DataFrame(), log_level=1)
Restart and rerun the
Modifying run_startup revealed some state was unintentionally shared between testcases, so tests were updated to more specifically test the intended logger
Overview
30 created a pattern for future data loaders and an option for using pre-loaded dataframes.
This exposes that functionality to be passed in at setup.
Description of changes
Add two optional paramaters to run setup.
Starting Functional Tests:
Run the example notebook and verify that old pattern works without change.
Turn the urllib.requests cell into a raw or markdown cell (this assumes that the dataset has been pulled locally from the first run) Add something like the following to a first cell or two :
Restart and rerun the
Modifying run_startup revealed some state was unintentionally shared between testcases, so tests were updated to more specifically test the intended logger
Author Checklist
changelog/ISSUE.TYPE.rst
files; see changelog/README.md.