From 1968e8ebcf3e02ccc44213bffde508318f154245 Mon Sep 17 00:00:00 2001
From: Sunnyhaze <mxch1122@126.com>
Date: Sun, 13 Jul 2025 17:11:40 +0800
Subject: [PATCH] [storage] add error logging when don't call step before first
 run.

---
 dataflow/utils/storage.py           | 7 +++++++
 test/small_functions/translation.py | 8 +++++---
 2 files changed, 12 insertions(+), 3 deletions(-)

diff --git a/dataflow/utils/storage.py b/dataflow/utils/storage.py
index ed5e6595..7514aabc 100644
--- a/dataflow/utils/storage.py
+++ b/dataflow/utils/storage.py
@@ -40,6 +40,9 @@ def __init__(self,
         self.logger = get_logger()
 
     def _get_cache_file_path(self, step) -> str:
+        if step == -1:
+            self.logger.error("You must call storage.step() before reading or writing data. Please call storage.step() first for each operator step.")  
+            raise ValueError("You must call storage.step() before reading or writing data. Please call storage.step() first for each operator step.")
         if step == 0:
             # If it's the first step, use the first entry file name
             return os.path.join(self.first_entry_file_name)
@@ -56,6 +59,10 @@ def reset(self):
     
     def _load_local_file(self, file_path: str, file_type: str) -> pd.DataFrame:
         """Load data from local file based on file type."""
+        # check if file exists
+        if not os.path.exists(file_path):
+            raise FileNotFoundError(f"File {file_path} does not exist. Please check the path.")
+        # Load file based on type
         try:
             if file_type == "json":
                 return pd.read_json(file_path)
diff --git a/test/small_functions/translation.py b/test/small_functions/translation.py
index 0ffa2b31..ea773db0 100644
--- a/test/small_functions/translation.py
+++ b/test/small_functions/translation.py
@@ -5,7 +5,7 @@
 class GPT_generator():
     def __init__(self):
         self.storage = FileStorage(
-            first_entry_file_name="../../dataflow/example/GeneralTextPipeline/translation.jsonl",
+            first_entry_file_name="../../dataflow/example/GeneralTextPipeline/translation.jsonl1",
             cache_path="./cache",
             file_name_prefix="translation",
             cache_type="jsonl",
@@ -16,13 +16,15 @@ def __init__(self):
                 model_name="gpt-4o",
                 max_workers=2
         )
-        self.prompt_generator = PromptedGenerator(llm_serving = self.llm_serving)        
+        self.prompt_generator = PromptedGenerator(
+            llm_serving = self.llm_serving,
+            system_prompt = "Please translate to Chinese.",
+        )        
 
     def forward(self):
         # Initial filters
         self.prompt_generator.run(
             storage = self.storage.step(),
-            system_prompt = "Please translate to Chinese.",
             input_key = "raw_content",
         )