add app to generate target data for flusight

pyproject.toml

@@ -22,6 +22,7 @@ dependencies = [
 
 [project.optional-dependencies]
 dev = [
+    "freezegun",
     "pytest",
     "pip-tools"
 ]

requirements/requirements.txt

@@ -14,6 +14,8 @@ click==8.1.7
     # via
     #   hub-dashboard-predtimechart (pyproject.toml)
     #   pip-tools
+freezegun==1.5.1
+    # via hub-dashboard-predtimechart (pyproject.toml)
 iniconfig==2.0.0
     # via pytest
 jsonschema==4.23.0
@@ -45,7 +47,9 @@ pyproject-hooks==1.1.0
 pytest==8.3.2
     # via hub-dashboard-predtimechart (pyproject.toml)
 python-dateutil==2.9.0.post0
-    # via pandas
+    # via
+    #   freezegun
+    #   pandas
 pytz==2024.1
     # via pandas
 pyyaml==6.0.2

src/hub_predtimechart/app/generate_target_json_files_FluSight.py

@@ -0,0 +1,78 @@
+from datetime import date, datetime, timedelta
+import json
+import re
+from pathlib import Path
+
+import click
+import polars as pl
+import structlog
+
+from hub_predtimechart.generate_target_data import target_data_for_FluSight
+from hub_predtimechart.app.generate_json_files import json_file_name
+from hub_predtimechart.util.logs import setup_logging
+
+
+setup_logging()
+logger = structlog.get_logger()
+
+
+@click.command()
+@click.argument('hub_dir', type=click.Path(file_okay=False, exists=True))
+@click.argument('target_out_dir', type=click.Path(file_okay=False, exists=True))
+def main(hub_dir, target_out_dir):
+    '''
+    Generates the target data json files used by https://github.com/reichlab/predtimechart to
+    visualize a hub's forecasts.
+
+    HUB_DIR: (input) a directory Path of a https://hubverse.io hub to generate target data json files from
+
+    TARGET_OUT_DIR: (output) a directory Path to output the viz target data json files to
+    \f
+    :param hub_dir: (input) a directory Path of a https://hubverse.io hub to generate target data json files from
+    :param target_out_dir: (output) a directory Path to output the viz target data json files to
+
+    '''
+    logger.info(f'main({hub_dir=}, {target_out_dir=}): entered')
+
+    hub_dir = Path(hub_dir)
+    target_out_dir = Path(target_out_dir)
+
+    # load the target data csv file from the hub repo
+    # for now, file path for target data is hard coded
+    target_data_df = pl.read_csv(hub_dir / 'target-data/target-hospital-admissions.csv')
+
+    # for each location,
+    # - generate target data file contents
+    # - save as json
+    json_files = []
+    for loc in target_data_df['location'].unique():
+        task_ids_tuple = (loc,)
+        location_data_dict = target_data_for_FluSight(target_data_df, task_ids_tuple)
+        file_name = json_file_name('wk inc flu hosp', task_ids_tuple, reference_date_from_today())
+        json_files.append(target_out_dir / file_name)
+        with open(target_out_dir / file_name, 'w') as fp:
+            json.dump(location_data_dict, fp, indent=4)
+
+    logger.info(f'main(): done: {len(json_files)} JSON files generated: {[str(_) for _ in json_files]}. ')
+
+
+
+#
+# _generate_json_files() and helpers
+#
+def reference_date_from_today(now: datetime = date.today()) -> datetime:
+    # Calculate the days until the next Saturday
+    days_to_saturday = 5 - now.weekday()
+    if days_to_saturday < 0:
+        days_to_saturday += 7
+
+    # Add the calculated days to the given date
+    return now + timedelta(days=days_to_saturday)
+
+
+#
+# main()
+#
+
+if __name__ == '__main__':
+    main()

tests/hub_predtimechart/test_generate_target_data.py

@@ -1,10 +1,13 @@
 import json
+from datetime import date
 from pathlib import Path
 
+from freezegun import freeze_time
 import polars as pl
 
 from hub_predtimechart.app.generate_json_files import json_file_name
 from hub_predtimechart.generate_target_data import target_data_for_FluSight
+from hub_predtimechart.app.generate_target_json_files_FluSight import reference_date_from_today
 
 
 def test_generate_target_data_for_FluSight():
@@ -20,3 +23,19 @@ def test_generate_target_data_for_FluSight():
         model_output_file = hub_dir / 'model-output/Flusight-baseline/2022-10-22-Flusight-baseline.csv'
         act_data = target_data_for_FluSight(target_data_df, task_ids_tuple)
         assert act_data == exp_data
+
+
+@freeze_time("2024-10-24")
+def test_reference_date_from_today():
+    # Test dates are Sunday, Thursday, and Saturday.
+    # For all of these, the expected reference date is the same: the Saturday
+    exp_reference_date = date.fromisoformat("2024-10-26")
+    for reference_date_str in ["2024-10-20", "2024-10-24", "2024-10-26"]:
+        act_reference_date = reference_date_from_today(date.fromisoformat(reference_date_str))
+        assert act_reference_date == exp_reference_date
+
+    # test that now is used if no parameter passed
+    # in this case, we expect that reference_date_from_today uses now,
+    # which is set to "2024-10-24" via the freeze_time decorator.
+    act_reference_date = reference_date_from_today()
+    assert act_reference_date == exp_reference_date