fix(util): replace string parsing with AST in inspect_without_import #3504

augur/util/inspect_without_import.py

@@ -3,17 +3,59 @@
 #while the config needs the phase names before population
 #The solution is to either make the user define the phase names seperate or to do this
 #Which is to import the .py as text and parse the function names.
-def get_phase_names_without_import():
-    raw_file = open("augur/tasks/start_tasks.py")
-    lines = raw_file.readlines()
-    raw_file.close()
 
-    phase_names = []
+"""
+This module lets us inspect Python files without actually importing them.
+
+It uses Python's Abstract Syntax Tree (AST) to parse source files and extract specific 
+function names. This is much safer and more robust than simple string parsing because 
+it handles things like indentation, whitespace, comments, and decorators correctly.
+"""
+
+import ast
+from pathlib import Path
+from typing import List
 
-    for line in lines:
-        if "def " in line and "_phase(" in line:
-            without_def = line.split()[1]
-            phase_names.append(without_def.split('(')[0])
-
 
+def get_phase_names_without_import() -> List[str]:
+    """
+    Grabs the names of phase functions from start_tasks.py.
+
+    Instead of importing the file (which runs code), we parse it strictly as text 
+    using the AST module. We're looking for any function definition that has 
+    '_phase' somewhere in its name.
+
+    Returns:
+        List[str]: A list of found function names, like 'prelim_phase' or 'secondary_repo_collect_phase'.
+
+    Raises:
+        FileNotFoundError: If we can't locate the start_tasks.py file.
+        SyntaxError: If start_tasks.py has broken Python syntax.
+    """
+    current_file = Path(__file__).resolve()
+    start_tasks_path = current_file.parent.parent / "tasks" / "start_tasks.py"
+
+    try:
+        source_code = start_tasks_path.read_text(encoding='utf-8')
+    except FileNotFoundError as e:
+        raise FileNotFoundError(
+            f"We couldn't find start_tasks.py at {start_tasks_path}"
+        ) from e
+
+    try:
+        tree = ast.parse(source_code, filename=str(start_tasks_path))
+    except SyntaxError as e:
+        raise SyntaxError(
+            f"The file exists, but it contains invalid Python syntax: {e}"
+        ) from e
+
+    phase_names = []
+
+    # We walk through every node in the abstract syntax tree
+    for node in ast.walk(tree):
+        if isinstance(node, ast.FunctionDef):
+            # We want any function that identifies itself as a 'phase'
+            if node.name.endswith('_phase'):
+                phase_names.append(node.name)
+
     return phase_names

pyproject.toml

@@ -151,6 +151,7 @@ addopts = "-ra -s"
 testpaths = [
     "tests/test_classes",
     "tests/test_application/test_cli/test_csv_utils.py",
+    "tests/test_util/test_inspect_without_import.py",
     # "tests/test_routes", # runs, but needs a fixture for connecting to the web interface of Augur
     # "tests/test_metrics",
     # "tests/test_tasks",

tests/test_util/__init__.py

@@ -0,0 +1 @@
+"""Tests for the augur.util package."""