Merge pull request #108 from openclimatefix/fix-batch-system-process

Fix batch system process

pvoutput/prcoess.py

@@ -1,9 +1,12 @@
 """ Function to process data """
+import logging
 from io import StringIO
 
 import numpy as np
 import pandas as pd
 
+logger = logging.getLogger(__name__)
+
 
 def process_system_status(pv_system_status_text, date) -> pd.DataFrame:
     """
@@ -17,17 +20,20 @@ def process_system_status(pv_system_status_text, date) -> pd.DataFrame:
     Returns: dataframe of data
     """
 
-    # get system id
-    system_id = int(pv_system_status_text.split(";")[0])
-    pv_system_status_text = ";".join(pv_system_status_text.split(";")[1:])
-
     # See https://pvoutput.org/help/data_services.html#data-services-get-system-status
     columns = [
         "cumulative_energy_gen_Wh",
         "instantaneous_power_gen_W",
         "temperature_C",
         "voltage",
     ]
+    if pv_system_status_text == "no status found":
+        logger.debug("Text was empty so return empty dataframe")
+        return pd.DataFrame(columns=columns + ["system_id", "datetime"])
+
+    # get system id
+    system_id = int(pv_system_status_text.split(";")[0])
+    pv_system_status_text = ";".join(pv_system_status_text.split(";")[1:])
 
     try:
         one_pv_system_status = pd.read_csv(

pvoutput/pvoutput.py

@@ -334,7 +334,7 @@ def get_system_status(
 
         except NoStatusFound:
             _LOG.info(f"system_id {all_pv_system_id}: No status found for date %s", date)
-            pv_system_status_text = ""
+            pv_system_status_text = "no status found"
 
         # each pv system is on a new line
         pv_systems_status_text = pv_system_status_text.split("\n")

tests/test_process.py

@@ -27,6 +27,14 @@ def test_process_system_status_2():
     assert (one_status["system_id"] == 1234).all()
 
 
+def test_process_system_status_none():
+
+    one_status = process_system_status(
+        pv_system_status_text="no status found", date=date(2022, 1, 1)
+    )
+    assert len(one_status) == 0
+
+
 def test_process_system_status_less_columns_two_data_points():
     # this has all missing data values
     pv_system_status_text = "1234;07:45,21,255;" "07:45,22,256"