From 528aaf1957a2ba24bb176576fabec484b3debbea Mon Sep 17 00:00:00 2001 From: Ed_ Date: Fri, 27 Feb 2026 22:23:50 -0500 Subject: [PATCH] feat(mma): Finalize Phase 1 with AST-based outline and improved tiered selection --- aggregate.py | 15 ++--- outline_tool.py | 109 ++++++++++++----------------------- tests/test_tiered_context.py | 27 +++++++++ 3 files changed, 73 insertions(+), 78 deletions(-) diff --git a/aggregate.py b/aggregate.py index 6473875..538e5a9 100644 --- a/aggregate.py +++ b/aggregate.py @@ -204,7 +204,7 @@ def build_discussion_text(history: list[str]) -> str: def build_tier1_context(file_items: list[dict], screenshot_base_dir: Path, screenshots: list[str], history: list[str]) -> str: """ Tier 1 Context: Strategic/Orchestration. - Full content for core conductor files, summaries for others. + Full content for core conductor files and files with tier=1, summaries for others. """ core_files = {"product.md", "tech-stack.md", "workflow.md", "tracks.md"} @@ -217,13 +217,14 @@ def build_tier1_context(file_items: list[dict], screenshot_base_dir: Path, scree path = item.get("path") name = path.name if path else "" - if name in core_files: + if name in core_files or item.get("tier") == 1: # Include in full sections.append("### `" + (item.get("entry") or str(path)) + "`\n\n" + f"```{path.suffix.lstrip('.') if path.suffix else 'text'}\n{item.get('content', '')}\n```") else: # Summarize - sections.append(summarize.summarise_file(path, item.get("content", ""))) + sections.append("### `" + (item.get("entry") or str(path)) + "`\n\n" + + summarize.summarise_file(path, item.get("content", ""))) parts.append("## Files (Tier 1 - Mixed)\n\n" + "\n\n---\n\n".join(sections)) @@ -247,7 +248,7 @@ def build_tier2_context(file_items: list[dict], screenshot_base_dir: Path, scree def build_tier3_context(file_items: list[dict], screenshot_base_dir: Path, screenshots: list[str], history: list[str], focus_files: list[str]) -> str: """ Tier 3 Context: Execution/Worker. - Full content for focus_files, summaries for others. + Full content for focus_files and files with tier=3, summaries/skeletons for others. """ parts = [] @@ -265,7 +266,7 @@ def build_tier3_context(file_items: list[dict], screenshot_base_dir: Path, scree is_focus = True break - if is_focus: + if is_focus or item.get("tier") == 3: sections.append("### `" + (entry or path_str) + "`\n\n" + f"```{path.suffix.lstrip('.') if path and path.suffix else 'text'}\n{item.get('content', '')}\n```") else: @@ -277,9 +278,9 @@ def build_tier3_context(file_items: list[dict], screenshot_base_dir: Path, scree sections.append(f"### `{entry or path_str}` (AST Skeleton)\n\n```python\n{skeleton}\n```") except Exception as e: # Fallback to summary if AST parsing fails - sections.append(summarize.summarise_file(path, content)) + sections.append(f"### `{entry or path_str}`\n\n" + summarize.summarise_file(path, content)) else: - sections.append(summarize.summarise_file(path, content)) + sections.append(f"### `{entry or path_str}`\n\n" + summarize.summarise_file(path, content)) parts.append("## Files (Tier 3 - Focused)\n\n" + "\n\n---\n\n".join(sections)) diff --git a/outline_tool.py b/outline_tool.py index 0198b06..d3e87f3 100644 --- a/outline_tool.py +++ b/outline_tool.py @@ -1,88 +1,55 @@ -import tree_sitter -import tree_sitter_python +import ast from pathlib import Path class CodeOutliner: def __init__(self): - self.language = tree_sitter.Language(tree_sitter_python.language()) - self.parser = tree_sitter.Parser(self.language) + pass def outline(self, code: str) -> str: - tree = self.parser.parse(bytes(code, "utf8")) - lines = code.splitlines() - + code = code.lstrip(chr(0xFEFF)) + try: + tree = ast.parse(code) + except SyntaxError as e: + return f"ERROR parsing code: {e}" + output = [] def get_docstring(node): - # In Python, docstring is usually the first expression statement in a block - body = node.child_by_field_name("body") - if body and body.type == "block": - for child in body.children: - if child.type == "comment": - continue - if child.type == "expression_statement": - expr = child.children[0] - if expr.type == "string": - doc = code[expr.start_byte:expr.end_byte].strip() - # Strip quotes - if doc.startswith(('"""', "'''")): - doc = doc[3:-3] - elif doc.startswith(('"', "'")): - doc = doc[1:-1] - return doc.splitlines()[0] if doc else "" - break + doc = ast.get_docstring(node) + if doc: + return doc.splitlines()[0] return None def walk(node, indent=0): - node_type = node.type - name = None + if isinstance(node, ast.ClassDef): + start_line = node.lineno + end_line = getattr(node, "end_lineno", start_line) + output.append(f"{' ' * indent}[Class] {node.name} (Lines {start_line}-{end_line})") + doc = get_docstring(node) + if doc: + output.append(f"{' ' * (indent + 1)}\"\"\"{doc}\"\"\"") + for item in node.body: + walk(item, indent + 1) - if node_type == "class_definition": - name_node = node.child_by_field_name("name") - if name_node: - name = code[name_node.start_byte:name_node.end_byte] - start_line = node.start_point.row + 1 - end_line = node.end_point.row + 1 - output.append(f"{' ' * indent}[Class] {name} (Lines {start_line}-{end_line})") - doc = get_docstring(node) - if doc: - output.append(f"{' ' * (indent + 1)}\"\"\"{doc}\"\"\"") + elif isinstance(node, (ast.FunctionDef, ast.AsyncFunctionDef)): + start_line = node.lineno + end_line = getattr(node, "end_lineno", start_line) + prefix = "[Async Func]" if isinstance(node, ast.AsyncFunctionDef) else "[Func]" + + # Check if it's a method + # We can check the indent or the parent, but in AST walk we know if we are inside a ClassDef + # Let's use a simpler heuristic for the outline: if indent > 0, it's likely a method. + if indent > 0: + prefix = "[Method]" + + output.append(f"{' ' * indent}{prefix} {node.name} (Lines {start_line}-{end_line})") + doc = get_docstring(node) + if doc: + output.append(f"{' ' * (indent + 1)}\"\"\"{doc}\"\"\"") + + for node in tree.body: + walk(node) - elif node_type in ("function_definition", "async_function_definition"): - name_node = node.child_by_field_name("name") - if name_node: - name = code[name_node.start_byte:name_node.end_byte] - start_line = node.start_point.row + 1 - end_line = node.end_point.row + 1 - prefix = "[Async Func]" if node_type == "async_function_definition" else "[Func]" - - # Check if it's a method (parent is a class body) - parent = node.parent - while parent and parent.type != "class_definition": - if parent.type == "module": - break - parent = parent.parent - - if parent and parent.type == "class_definition": - prefix = "[Method]" - - output.append(f"{' ' * indent}{prefix} {name} (Lines {start_line}-{end_line})") - doc = get_docstring(node) - if doc: - output.append(f"{' ' * (indent + 1)}\"\"\"{doc}\"\"\"") - - for child in node.children: - # Don't recurse into function bodies for outlining functions, - # but we DO want to recurse into classes to find methods. - if node_type == "class_definition": - if child.type == "block": - walk(child, indent + 1) - elif node_type == "module": - walk(child, indent) - elif node_type == "block": - walk(child, indent) - - walk(tree.root_node) return "\n".join(output) def get_outline(path: Path, code: str) -> str: diff --git a/tests/test_tiered_context.py b/tests/test_tiered_context.py index 630af1e..04cc611 100644 --- a/tests/test_tiered_context.py +++ b/tests/test_tiered_context.py @@ -107,3 +107,30 @@ def test_build_files_section_with_dicts(tmp_path): result = build_files_section(tmp_path, files_config) assert "content1" in result assert "file1.txt" in result + +def test_tiered_context_by_tier_field(): + file_items = [ + {"path": Path("tier1_file.txt"), "entry": "tier1_file.txt", "content": "Full Tier 1 Content\nLine 2", "tier": 1}, + {"path": Path("tier3_file.txt"), "entry": "tier3_file.txt", "content": "Full Tier 3 Content\nLine 2\nLine 3\nLine 4\nLine 5\nLine 6\nLine 7\nLine 8\nLine 9\nLine 10", "tier": 3}, + {"path": Path("other.txt"), "entry": "other.txt", "content": "Other Content\nLine 2\nLine 3\nLine 4\nLine 5\nLine 6\nLine 7\nLine 8\nLine 9\nLine 10", "tier": None} + ] + + # Test Tier 1 Context + result_t1 = build_tier1_context(file_items, Path("."), [], []) + assert "Full Tier 1 Content" in result_t1 + assert "Line 2" in result_t1 # In full + + # tier3_file.txt should be summarized + assert "tier3_file.txt" in result_t1 + assert "preview:" in result_t1 + assert "Line 9" not in result_t1 # Only first 8 lines in preview + + # Test Tier 3 Context + result_t3 = build_tier3_context(file_items, Path("."), [], [], focus_files=[]) + assert "Full Tier 3 Content" in result_t3 + assert "Line 10" in result_t3 # In full + + # tier1_file.txt should be summarized + assert "tier1_file.txt" in result_t3 + assert "preview:" in result_t3 + assert "Full Tier 1 Content" in result_t3 # It's short, so it's in preview