Home URL Adder

.env-template

@@ -0,0 +1,39 @@
+# ------------------------------
+# Environmental Variables
+# ------------------------------
+LOG_LEVEL=INFO
+LOG_VERBOSE=FALSE
+
+# ------------------------------
+# Databases
+# ------------------------------
+
+# Postgres
+DB_POSTGRES_USER=
+DB_POSTGRES_PASSWORD=
+DB_POSTGRES_NAME=
+DB_POSTGRES_HOST=
+DB_POSTGRES_PORT=
+
+# Clickhouse
+DB_CLICKHOUSE_HOST=
+DB_CLICKHOUSE_PORT=
+DB_CLICKHOUSE_USER=
+DB_CLICKHOUSE_PASSWORD=
+DB_CLICKHOUSE_NAME=
+
+# ------------------------------
+# Optional Vars
+# ------------------------------
+
+# Pyroscope Configuration
+PYROSCOPE_SERVER=
+PYROSCOPE_APPLICATION_NAME=
+PYROSCOPE_API_KEY=
+
+# Sentry Configuration
+SENTRY_DSN=
+
+
+
+

.gitignore

@@ -1,7 +1,5 @@
-.nova
 .DS_Store
 .env
 venv
 __pycache__
-BUILDER.md
-logs*
+logs*

.nova/Configuration.json

@@ -0,0 +1,5 @@
+{
+  "workspace.art_style" : 1,
+  "workspace.color" : 0,
+  "workspace.name" : "Rabbit Run"
+}

app/__init__.py

@@ -1,4 +1,10 @@
-# __init__.py
-# Relative Path: app/__init__.py
-from .utils import logger
-from .processes import preprocess_data
+# app/__init__.py
+from .utils import configure_monitoring, logger
+from dotenv import load_dotenv
+from .database.postgres.connect import test_connection
+
+def startup():
+    logger.info('Starting up...')
+    load_dotenv()
+    configure_monitoring()
+    test_connection()

app/database/__init__.py

@@ -1,7 +1,8 @@
-# __init__.py
-# Relative Path: app/database/__init__.py
-from app.utils import logger
+# app/database/__init__.py
+from ..utils import logger
+
+#from app.utils import logger
 # Import from Postgres directory
-from .postgres import axe_postgres, fetch_unprocessed_rules, mark_rule_as_processed as mark_axe_rule_as_processed
+#from .postgres import axe_postgres, fetch_unprocessed_rules, mark_rule_as_processed as mark_axe_rule_as_processed
 # Import from ClickHouse directory
-from .clickhouse import axe_clickhouse
+#from .clickhouse import axe_clickhouse

app/database/postgres/__init__.py

@@ -1,4 +1,5 @@
-# __init__.py
-# Relative Path: app/database/postgres/__init__.py
+# app/database/postgres/__init__.py
+
+
 from .process_tests import select_rules_data as axe_postgres, mark_rule_as_processed
-from .fetch_unprocessed import fetch_unprocessed_rules
+from .fetch_unprocessed import fetch_unprocessed_rules

app/database/postgres/connect.py

@@ -1,13 +1,14 @@
-# connect.py
-# Relative Path: app/database/postgres/connect.py
+# app/database/postgres/connect.py
+import os
 from dotenv import load_dotenv
 from sqlalchemy import create_engine
 from sqlalchemy.orm import sessionmaker, registry
-import os
-from .. import logger
+from app import logger
+
+DB_USER = os.getenv("DB_POSTGRES_USER")
+if not DB_USER:
+    raise ValueError("Environment variable DB_POSTGRES_USER is not set!")
 
-# load .env variables
-load_dotenv()
 
 # Retrieving environment variables
 DB_USER = os.getenv("DB_POSTGRES_USER")
@@ -20,7 +21,7 @@
 SQLALCHEMY_DATABASE_URL = f"postgresql://{DB_USER}:{DB_PASSWORD}@{DB_HOST}:{DB_PORT}/{DB_NAME}"
 
 engine = create_engine(
-    SQLALCHEMY_DATABASE_URL
+    SQLALCHEMY_DATABASE_URL, future=True
 )
 SessionLocal = sessionmaker(autocommit=False, autoflush=False, bind=engine)
 
@@ -31,20 +32,21 @@
 mapper_registry = registry()
 Base = mapper_registry.generate_base()
 
+postgres_conn = SessionLocal
 
 def test_connection():
     connection = None
     try:
         connection = engine.connect()
-        logger.debug("Connected to 🐘")
+        print("Connected to 🐘")
     except Exception as e:
-        logger.error(f"Unable to connect to PostgreSQL: {str(e)}")
+        print(f"Unable to connect to PostgreSQL: {str(e)}")
     finally:
         # Ensure the connection object is not None before trying to close it
         if connection:
             connection.close()
-            logger.debug("🐘 Connection closed")
-
+            print("🐘 Connection closed")
 
 
-test_connection()
+if __name__ == "__main__":
+    test_connection()

app/database/postgres/queries/clothe_domains.sql

@@ -0,0 +1,6 @@
+-- app/database/postgres/queries/clothe_domains.sql
+-- Creates url entry
+UPDATE targets.domains
+SET home_url = :home_url
+WHERE id = :domain_id
+RETURNING id;

app/database/postgres/queries/get_naked_domains.sql

@@ -0,0 +1,10 @@
+-- app/database/postgres/queries/get_naked_domains.sql
+SELECT
+  id AS "domain_id",
+  "domain"
+FROM targets.domains d
+WHERE (home_url IS NULL OR home_url = '')
+  AND active = TRUE
+  AND "valid" = TRUE
+LIMIT 1;
+

app/database/postgres/queries/get_rule_data.sql

@@ -0,0 +1 @@
+-- app/database/postgres/queries/get_rule_data.sql

app/database/postgres/queries/get_unprocessed_rules.sql

@@ -0,0 +1,7 @@
+-- app/database/postgres/queries/get_unprocessed_rules.sql
+
+SELECT id as rule_id
+FROM axe.rules
+WHERE imported = false
+ORDER BY id
+LIMIT %s OFFSET %s

app/database/postgres/queries/upsert_url.sql

@@ -0,0 +1,5 @@
+-- app/database/postgres/queries/upsert_url.sql
+INSERT INTO targets.urls (url, domain_id)
+VALUES (:home_url, :domain_id)
+ON CONFLICT (url) DO UPDATE SET url = :home_url, domain_id = :domain_id
+RETURNING id;

app/database/postgres/run_query.py

@@ -0,0 +1,37 @@
+# app/database/postgres/run_query.py
+import os
+import re
+from app.database.postgres.connect import postgres_conn as conn
+from app import logger
+from sqlalchemy.exc import SQLAlchemyError
+from sqlalchemy import text
+
+QUERIES_DIRECTORY = os.path.join(os.path.dirname(__file__), "queries")
+
+def run_query(query_name, vars=None):
+    query_file = os.path.join(QUERIES_DIRECTORY, f"{query_name}.sql")
+
+    with open(query_file) as file:
+        sql_content = file.read()
+
+    logger.info(f"Running query: {query_name}")
+
+    session = conn()
+
+    try:
+        result = session.execute(text(sql_content), vars)
+        logger.debug(f'Formatted SQL to Run:\n %s', sql_content)
+
+        session.commit()
+
+        rows = result.fetchall()
+        logger.debug(f"Result rows: {rows}")
+
+        return rows
+
+    except SQLAlchemyError as e:
+        session.rollback()
+        logger.error(f"Error while running query {query_name}: {str(e)}")
+        return None
+    finally:
+        session.close()

app/main.py

@@ -1,24 +1 @@
 # app/main.py
-import time
-import sys
-import os
-from .utils import logger
-sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), '..')))
-from app.processes import execute_axes
-from app.database import fetch_unprocessed_rules, mark_axe_rule_as_processed
-
-def yeet_axes():
-    while True:
-        rules_to_process = fetch_unprocessed_rules()
-        if rules_to_process:
-            # When there are rule_ids to process, process them.
-            for rule_id in rules_to_process:
-                execute_axes(rule_id)  # Inserts into ClickHouse
-                mark_axe_rule_as_processed(rule_id)  # Marks as processed in Postgres
-        else:
-            # When there are no more rule_ids to process, sleep for 10 seconds before checking again.
-            time.sleep(10)
-
-
-if __name__ == "__main__":
-    yeet_axes()

app/processes/__init__.py

@@ -1,5 +1,25 @@
-# __init__.py
-# Relative Path: app/processes/__init__.py
+# app/processes/__init__.py
+import time
+from multiprocessing import Process
+from app import logger
 
-from .axe import get_axes, execute_axes
-from .preprocess_tests import preprocess_data
+# Process Imports
+from .naked_urls import find_nakies
+
+def process_loop(process_func, sleep_time):
+    while True:
+        if not process_func():  # If there is no data to process
+            time.sleep(sleep_time)  # Wait for the specified amount of time
+
+def start_processes():
+    logger.info('Starting processes...')
+
+    # Functions with their sleep times
+    processes = [
+        (find_nakies, 30)
+        #(fix_axe, 60)
+    ]
+
+    for process_func, sleep_time in processes:
+        process = Process(target=process_loop, args=(process_func, sleep_time))
+        process.start()

app/processes/naked_urls.py

@@ -0,0 +1,61 @@
+from app import logger
+from app.database.postgres.run_query import run_query
+import requests
+
+def find_nakies():
+    logger.info('Starting to find naked domains...')
+    query_name = "get_naked_domains"
+    result = run_query(query_name)
+
+    logger.debug('SQL result: %s', result)
+
+    if result:
+        try:
+            domain_id, domain = result[0]
+        except IndexError:
+            logger.error('No rows were returned from the SQL query.')
+            return False  # No data to process
+
+        home_url = get_home_url(domain)
+        record_home_url(domain_id, home_url)
+
+        if home_url != "BADDIE":
+            upsert_url(domain_id, home_url)
+
+        if home_url == "BADDIE":
+            logger.debug(f'We got a BADDIE for %s', domain)
+        else:
+            logger.debug(f'%s\'s home url is: %s', domain, home_url)
+
+        return True  # There is data to process
+
+    else:
+        logger.info('No naked domains found.')
+        return False  # No data to process
+
+
+def get_home_url(domain):
+    logger.debug(f'Getting home url for %s', domain)
+
+    try:
+        response = requests.get(f'http://{domain}', timeout=5, allow_redirects=True)
+
+        if response.status_code == 200:
+            return response.url
+        else:
+            return "BADDIE"
+    except requests.exceptions.RequestException as e:
+        logger.error(f"Error while getting home URL for {domain}: {str(e)}")
+        return "BADDIE"
+
+def record_home_url(domain_id, home_url):
+    logger.debug('Fixing home_url for domain_id: %s', domain_id)
+    query_name = "clothe_domains"
+    variables = {"domain_id": domain_id, "home_url": home_url}
+    result = run_query(query_name, variables)
+
+def upsert_url(domain_id, home_url):
+    logger.debug('Upserting url for domain_id: %s', domain_id)
+    query_name = "upsert_url"
+    variables = {"domain_id": domain_id, "home_url": home_url}
+    result = run_query(query_name, variables)

app/utils/__init__.py

@@ -1,3 +1,2 @@
-# __init__.py
-# Relative Path: app/utils/__init__.py
-from .logger import logger
+# app/utils/__init__.py
+from .monitoring import which_extras as configure_monitoring, logger

app/utils/monitoring/__init__.py

@@ -0,0 +1,18 @@
+# app/utils/monitoring/__init__.py
+import os
+from .sentry import configure_sentry
+from .pyroscope import configure_pyroscope
+from .logging import logger
+
+def which_extras():
+    # Check if SENTRY_DSN environment variable has a value
+    if os.getenv("SENTRY_DSN"):
+        configure_sentry()
+        logger.info('Sentry Configured')
+
+    # Check if PYROSCOPE_API_KEY environment variable has a value
+    if os.getenv("PYROSCOPE_API_KEY"):
+        configure_pyroscope()
+        logger.info('Pyroscope Configured')
+
+

app/utils/logger/logger.py → app/utils/monitoring/logging.py

@@ -1,4 +1,4 @@
-# app/utils/logger/logger.py
+# app/utils/monitoring/logger.py
 import logging
 import os
 import time
@@ -8,7 +8,8 @@
 # Logger Name and Level
 LOGGER_NAME = "LoggyMcLogFace"
 LOG_LEVEL = os.environ.get("LOG_LEVEL", "INFO").upper()
-LOG_VERBOSE = True if LOG_LEVEL == "DEBUG" else os.environ.get("LOG_VERBOSE", "False").lower() == "true"
+LOG_VERBOSE = os.environ.get("LOG_VERBOSE", "False").lower() == "true"
+
 
 if LOG_VERBOSE:
     FMT_STREAM = "%(asctime)s.%(msecs)03d %(levelname)-8s [%(filename)s:%(funcName)s:%(lineno)d] %(message)s"
@@ -86,3 +87,4 @@ def log_exception(exc_type, exc_value, exc_traceback):
     logger.warning("This is a warning message")
     logger.error("This is an error message")
     logger.critical("This is a critical message")
+