Better handle missing files

bin/aip

@@ -37,20 +37,22 @@ from aip.models.alpha import Alpha
 from aip.models.prioritize import New
 from aip.models.prioritize import Consistent
 from aip.models.prioritize import RandomForest
-from pathlib import Path
 from os import makedirs, path, scandir
 from datetime import date, timedelta, datetime
 import sys
 
+
 def validate_and_convert_date(date_str):
+    """
+    Validates a date string in 'YYYY-MM-DD' format and converts it to a date object.
+    """
     try:
         dateobj = datetime.strptime(date_str, '%Y-%m-%d')
         return dateobj.date()
     except ValueError as e:
         print('Invalid date format. It should be YYYY-MM-DD')
         raise e
 
-#project_dir = Path(__file__).resolve().parents[1]
 
 if __name__ == '__main__':
     if len(sys.argv) == 2:
@@ -59,12 +61,10 @@ if __name__ == '__main__':
     else:
         day = date.today()
 
-
     log_fmt = '%(asctime)s - %(name)s - %(levelname)s - %(message)s'
     #logging.basicConfig(level=logging.INFO, format=log_fmt)
     logging.basicConfig(level=logging.DEBUG, format=log_fmt)
 
-
     #Alpha Model
     output_dir = path.join(project_dir, 'data', 'output', 'Alpha')
     if not path.exists(output_dir):

etc/docker/entrypoint.sh

@@ -1,8 +1,16 @@
 #!/bin/bash --login
+
+# Exit immediately if any command exits with a non-zero status
 set -e
 
+# Force the command prompt to display colors
+export force_color_prompt=yes
+
+# Initialize Conda in the current shell session
 source $HOME/miniconda3/etc/profile.d/conda.sh
 
+# Activate the conda environment already created in the docker
 conda activate aip
-export force_color_prompt=yes
+
+# Execute any command passed to the container when run
 exec "$@"

lib/aip/data/access.py

@@ -50,6 +50,14 @@ def _get_honeypot_ips(for_date=None):
     '''
     Filter those honeypots active due date for_date, if there are operation dates in the honeypot file.
     '''
+    logger = logging.getLogger(__name__)
+    # Check if the file exists before attempting to read it
+    honeypot_public_ips = path.join(project_dir, 'data', 'external', 'honeypots_public_ips.csv')
+
+    if not path.exists(honeypot_public_ips):
+        logger.error(f"File 'honeypot_public_ips.csv' does not exist. Raising error.")
+        raise FileNotFoundError("Required file 'honeypots_public_ips.csv' does not exist.")
+
     honeypots = pd.read_csv(path.join(project_dir, 'data', 'external', 'honeypots_public_ips.csv'), comment='#')
     if for_date is not None:
         for_date = pd.to_datetime(for_date)

lib/aip/models/base.py

@@ -25,6 +25,7 @@
 __version__ = "0.0.1"
 
 import pandas as pd
+import logging
 
 from aip.data.access import data_dir
 from aip.utils.autoload import register, models
@@ -36,9 +37,19 @@ class BaseModel():
     Template class for AIP models
     '''
     def __init__(self):
+        # Set up the logger for the class
+        self.logger = logging.getLogger(self.__class__.__name__)
+
         # Model initialization and configuration
         self.blocklist = pd.DataFrame()
-        self.donotblocklist = pd.read_csv(path.join(data_dir, 'external', 'do_not_block_these_ips.csv'))
+        exclude_ips = path.join(data_dir, 'external', 'do_not_block_these_ips.csv')
+
+        if path.exists(exclude_ips):
+            self.donotblocklist = pd.read_csv(exclude_ips)
+        else:
+            # Warning: File 'do_not_block_these_ips.csv' does not exist. Initializing with empty DataFrame.
+            self.logger.warning("File 'do_not_block_these_ips.csv' does not exist. Initializing with empty DataFrame.")
+            self.donotblocklist = pd.DataFrame(columns=['ip'])
 
     def sanitize(self, blocklist=None):
         if blocklist is None: