From 1084faaf0be7852522163f1d8e9d55e0adc47fcc Mon Sep 17 00:00:00 2001
From: AI Christianson <ai.christianson@christianson.ai>
Date: Thu, 13 Feb 2025 09:45:32 -0500
Subject: [PATCH] normalize/dedupe files for programmer tool

---
 ra_aid/tools/programmer.py      | 18 +++++++--------
 tests/ra_aid/test_programmer.py | 40 ++++++++++++++++++++++++++++++++-
 2 files changed, 47 insertions(+), 11 deletions(-)

diff --git a/ra_aid/tools/programmer.py b/ra_aid/tools/programmer.py
index 4d9df17..1902b28 100644
--- a/ra_aid/tools/programmer.py
+++ b/ra_aid/tools/programmer.py
@@ -1,4 +1,5 @@
 import os
+import os
 import sys
 from pathlib import Path
 from typing import Dict, List, Union
@@ -77,13 +78,6 @@ def run_programming_task(
 
     Returns: { "output": stdout+stderr, "return_code": 0 if success, "success": True/False }
     """
-    # Get related files if no specific files provided
-    file_paths = (
-        list(_global_memory["related_files"].values())
-        if "related_files" in _global_memory
-        else []
-    )
-
     # Build command
     aider_exe = get_aider_executable()
     command = [
@@ -96,6 +90,13 @@ def run_programming_task(
         "--no-check-update",
     ]
 
+    # Get combined list of files (explicit + related) with normalized paths
+    # and deduplicated using set operations
+    files_to_use = list({os.path.abspath(f) for f in (files or [])} | {
+        os.path.abspath(f) for f in _global_memory["related_files"].values() 
+        if "related_files" in _global_memory
+    })
+
     # Add config file if specified
     if "config" in _global_memory and _global_memory["config"].get("aider_config"):
         command.extend(["--config", _global_memory["config"]["aider_config"]])
@@ -112,9 +113,6 @@ def run_programming_task(
     command.append("-m")
 
     command.append(instructions)
-
-    # Add files to command
-    files_to_use = file_paths + (files or [])
     if files_to_use:
         command.extend(files_to_use)
 
diff --git a/tests/ra_aid/test_programmer.py b/tests/ra_aid/test_programmer.py
index 4540cbc..bef87e1 100644
--- a/tests/ra_aid/test_programmer.py
+++ b/tests/ra_aid/test_programmer.py
@@ -1,5 +1,6 @@
 import pytest
 from pathlib import Path
+from langchain_core.tools import Tool
 
 from ra_aid.tools.programmer import parse_aider_flags, run_programming_task, get_aider_executable
 
@@ -88,7 +89,7 @@ def test_aider_config_flag(mocker):
         "ra_aid.tools.programmer.run_interactive_command", return_value=(b"", 0)
     )
 
-    run_programming_task("test instruction")
+    run_programming_task.invoke({"instructions": "test instruction"})
 
     args = mock_run.call_args[0][0]  # Get the first positional arg (command list)
     assert "--config" in args
@@ -96,6 +97,43 @@ def test_aider_config_flag(mocker):
     assert args[config_index + 1] == "/path/to/config.yml"
 
 
+def test_path_normalization_and_deduplication(mocker, tmp_path):
+    """Test path normalization and deduplication in run_programming_task."""
+    # Create a temporary test file
+    test_file = tmp_path / "test.py"
+    test_file.write_text("")
+    new_file = tmp_path / "new.py"
+    
+    # Mock dependencies
+    mocker.patch("ra_aid.tools.programmer._global_memory", {"related_files": {}})
+    mocker.patch("ra_aid.tools.programmer.get_aider_executable", return_value="/path/to/aider")
+    mock_run = mocker.patch("ra_aid.tools.programmer.run_interactive_command", return_value=(b"", 0))
+
+    # Test duplicate paths
+    run_programming_task.invoke({
+        "instructions": "test instruction",
+        "files": [str(test_file), str(test_file)]  # Same path twice
+    })
+    
+    # Get the command list passed to run_interactive_command
+    cmd_args = mock_run.call_args[0][0]
+    # Count occurrences of test_file path in command
+    test_file_count = sum(1 for arg in cmd_args if arg == str(test_file))
+    assert test_file_count == 1, "Expected exactly one instance of test_file path"
+
+    # Test mixed paths
+    run_programming_task.invoke({
+        "instructions": "test instruction",
+        "files": [str(test_file), str(new_file)]  # Two different paths
+    })
+    
+    # Get the command list from the second call
+    cmd_args = mock_run.call_args[0][0]
+    # Verify both paths are present exactly once
+    assert sum(1 for arg in cmd_args if arg == str(test_file)) == 1, "Expected one instance of test_file"
+    assert sum(1 for arg in cmd_args if arg == str(new_file)) == 1, "Expected one instance of new_file"
+
+
 def test_get_aider_executable(mocker):
     """Test the get_aider_executable function."""
     mock_sys = mocker.patch("ra_aid.tools.programmer.sys")