devagent-ai 0.6.0__tar.gz → 0.7.0__tar.gz

{devagent_ai-0.6.0/devagent_ai.egg-info → devagent_ai-0.7.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: devagent-ai
-Version: 0.6.0
+Version: 0.7.0
 Summary: Evidence-driven local autonomous software engineering agent
 Author: Tom Ha
 Maintainer: Tom Ha

{devagent_ai-0.6.0 → devagent_ai-0.7.0}/devagent/__init__.py

@@ -1,3 +1,3 @@
 """DevAgent: evidence-driven local software engineering automation."""
 
-__version__ = "0.6.0"
+__version__ = "0.7.0"

{devagent_ai-0.6.0 → devagent_ai-0.7.0}/devagent/discovery.py

@@ -32,6 +32,8 @@ _LANGUAGE_EXTENSIONS = {
     ".cpp": "c++",
     ".h": "c/c++",
     ".cs": "c#",
+    ".fs": "f#",
+    ".vb": "visual-basic",
     ".rb": "ruby",
     ".php": "php",
 }
@@ -46,7 +48,11 @@ _MANIFESTS = {
     "cargo.toml",
     "go.mod",
     "pom.xml",
+    "mvnw",
     "build.gradle",
+    "build.gradle.kts",
+    "settings.gradle",
+    "settings.gradle.kts",
     "gradlew",
     "cmakelists.txt",
     "makefile",
@@ -56,6 +62,7 @@ _MANIFESTS = {
     ".gitlab-ci.yml",
     "azure-pipelines.yml",
 }
+_PROJECT_SUFFIXES = {".sln", ".slnx", ".csproj", ".fsproj", ".vbproj"}
 
 
 def _fingerprint(path: Path) -> str:
@@ -121,6 +128,65 @@ def _is_safe_generated_untracked_path(path: str) -> bool:
     )
 
 
+def _is_manifest(path: Path) -> bool:
+    return path.name.lower() in _MANIFESTS or path.suffix.lower() in _PROJECT_SUFFIXES
+
+
+def _is_ci_file(path: Path, root: Path) -> bool:
+    relative = path.relative_to(root).as_posix()
+    return relative.startswith(".github/workflows/") or path.name.lower() in {
+        "jenkinsfile",
+        ".gitlab-ci.yml",
+        "azure-pipelines.yml",
+    }
+
+
+def _priority_repository_files(root: Path, limit: int = 20_000) -> list[Path]:
+    """Recover high-value tracked manifests/CI files even after the bounded walk is full.
+
+    A huge monorepo can contain more than the normal 12k inventory budget before a deep
+    component is reached. Git's index is a cheap authoritative directory of tracked paths;
+    we inspect only manifest/project/CI candidates and keep the result bounded.
+    """
+    listed = _git(root, "ls-files")
+    if listed is None:
+        return []
+    priority: list[Path] = []
+    for raw in listed.splitlines():
+        if len(priority) >= limit:
+            break
+        if not raw or "\x00" in raw:
+            continue
+        relative = Path(raw)
+        if any(part in SKIP_DIRECTORIES for part in relative.parts) or is_secret_path(relative):
+            continue
+        candidate = root / relative
+        try:
+            resolved = candidate.resolve(strict=True)
+            resolved_relative = resolved.relative_to(root)
+        except (OSError, ValueError):
+            continue
+        if not resolved.is_file() or is_secret_path(resolved_relative):
+            continue
+        if _is_manifest(candidate) or _is_ci_file(candidate, root):
+            priority.append(candidate)
+    return priority
+
+
+def _manifest_languages(path: Path) -> set[str]:
+    name = path.name.lower()
+    suffix = path.suffix.lower()
+    if name in {"pom.xml", "mvnw", "build.gradle", "build.gradle.kts", "settings.gradle", "settings.gradle.kts", "gradlew"}:
+        return {"java"}
+    if suffix == ".csproj":
+        return {"c#"}
+    if suffix == ".fsproj":
+        return {"f#"}
+    if suffix == ".vbproj":
+        return {"visual-basic"}
+    return set()
+
+
 def _package_capabilities(path: Path, relative: str) -> tuple[list[str], list[Capability]]:
     frameworks: list[str] = []
     capabilities: list[Capability] = []
@@ -149,9 +215,42 @@ def _package_capabilities(path: Path, relative: str) -> tuple[list[str], list[Ca
     return sorted(set(frameworks)), capabilities
 
 
+def _java_maven_command(path: Path, root: Path, component: str) -> tuple[str, ...]:
+    wrapper = path.parent / "mvnw"
+    if wrapper.is_file():
+        executable = "./mvnw" if component == "." else f"./{component}/mvnw"
+    else:
+        executable = "mvn"
+    return (executable, "test") if component == "." else (executable, "-f", path.relative_to(root).as_posix(), "test")
+
+
+def _dotnet_capabilities(path: Path, root: Path, component: str) -> list[Capability]:
+    relative = path.relative_to(root).as_posix()
+    suffix = path.suffix.lower()
+    capabilities: list[Capability] = []
+    if suffix in {".sln", ".slnx"}:
+        capabilities.append(Capability("build", ("dotnet", "build", relative), relative, component, True))
+        return capabilities
+    if suffix not in {".csproj", ".fsproj", ".vbproj"}:
+        return capabilities
+
+    text = path.read_text(encoding="utf-8", errors="ignore").lower()
+    assets = path.parent / "obj" / "project.assets.json"
+    no_restore = ("--no-restore",) if assets.is_file() else ()
+    capabilities.append(
+        Capability("build", ("dotnet", "build", relative, *no_restore), relative, component, True)
+    )
+    if "<istestproject>true" in text or "microsoft.net.test.sdk" in text:
+        capabilities.append(
+            Capability("test", ("dotnet", "test", relative, *no_restore), relative, component, False)
+        )
+    return capabilities
+
+
 def _manifest_capabilities(path: Path, root: Path) -> tuple[list[str], list[Capability]]:
     relative = path.relative_to(root).as_posix()
     name = path.name.lower()
+    suffix = path.suffix.lower()
     frameworks: list[str] = []
     capabilities: list[Capability] = []
     component = path.parent.relative_to(root).as_posix() or "."
@@ -180,12 +279,16 @@ def _manifest_capabilities(path: Path, root: Path) -> tuple[list[str], list[Capa
         command = ("go", "test", "./...") if component == "." else ("go", "-C", component, "test", "./...")
         capabilities.append(Capability("test", command, relative, component))
     elif name == "pom.xml":
-        command = ("mvn", "test") if component == "." else ("mvn", "-f", relative, "test")
-        capabilities.append(Capability("test", command, relative, component))
-    elif name in {"build.gradle", "gradlew"}:
+        frameworks.append("maven")
+        capabilities.append(Capability("test", _java_maven_command(path, root, component), relative, component))
+    elif name in {"build.gradle", "build.gradle.kts", "gradlew"}:
+        frameworks.append("gradle")
         executable = "./gradlew" if component == "." and (path.parent / "gradlew").exists() else (f"./{component}/gradlew" if (path.parent / "gradlew").exists() else "gradle")
         command = (executable, "test") if component == "." else (executable, "-p", component, "test")
         capabilities.append(Capability("test", command, relative, component))
+    elif suffix in _PROJECT_SUFFIXES:
+        frameworks.append("dotnet")
+        capabilities.extend(_dotnet_capabilities(path, root, component))
     elif name == "cmakelists.txt":
         build_dir = "build" if component == "." else f"{component}/build"
         capabilities.append(Capability("test", ("ctest", "--test-dir", build_dir), relative, component))
@@ -201,7 +304,7 @@ def _ci_capabilities(path: Path, root: Path) -> list[Capability]:
     """Extract simple argv commands from CI run/command lines as high-value evidence."""
     relative = path.relative_to(root).as_posix()
     commands: list[Capability] = []
-    allowed = {"python", "python3", "pytest", "npm", "pnpm", "yarn", "go", "cargo", "mvn", "gradle", "./gradlew", "make", "ctest", "dotnet", "ruff", "mypy"}
+    allowed = {"python", "python3", "pytest", "npm", "pnpm", "yarn", "go", "cargo", "mvn", "gradle", "./gradlew", "./mvnw", "make", "ctest", "dotnet", "ruff", "mypy"}
     for raw_line in path.read_text(encoding="utf-8", errors="ignore").splitlines():
         stripped = raw_line.strip()
         value = ""
@@ -414,14 +517,17 @@ def _probe_python_test_capability(
 
 def discover_repository(root: Path | str, *, probe_capabilities: bool = True) -> RepositoryModel:
     root_path = Path(root).expanduser().resolve()
-    files = _walk(root_path)
-    ci_files = [
-        path
-        for path in files
-        if path.relative_to(root_path).as_posix().startswith(".github/workflows/")
-        or path.name.lower() in {"jenkinsfile", ".gitlab-ci.yml", "azure-pipelines.yml"}
-    ]
-    manifests = [path for path in files if path.name.lower() in _MANIFESTS or path.suffix.lower() in {".sln", ".csproj"}]
+    walked_files = _walk(root_path)
+    files = list(walked_files)
+    seen = {path.relative_to(root_path).as_posix() for path in files}
+    for path in _priority_repository_files(root_path):
+        relative = path.relative_to(root_path).as_posix()
+        if relative not in seen:
+            files.append(path)
+            seen.add(relative)
+
+    ci_files = [path for path in files if _is_ci_file(path, root_path)]
+    manifests = [path for path in files if _is_manifest(path)]
     component_paths = {root_path}
     for manifest in manifests:
         if manifest.parent != root_path and manifest not in ci_files:
@@ -434,6 +540,7 @@ def discover_repository(root: Path | str, *, probe_capabilities: bool = True) ->
         owned_files = [path for path in files if path == component_path or component_path in path.parents]
         language_counts = Counter(_LANGUAGE_EXTENSIONS[path.suffix.lower()] for path in owned_files if path.suffix.lower() in _LANGUAGE_EXTENSIONS)
         component_manifests = [path for path in manifests if path.parent == component_path]
+        inferred_languages = set().union(*(_manifest_languages(path) for path in component_manifests)) if component_manifests else set()
         frameworks: list[str] = []
         capabilities: list[Capability] = []
         for manifest in component_manifests:
@@ -456,9 +563,13 @@ def discover_repository(root: Path | str, *, probe_capabilities: bool = True) ->
             if probed is not None:
                 capabilities.append(probed)
                 frameworks.append("pytest" if "pytest" in probed.command else "unittest")
+        ordered_languages = [name for name, _ in language_counts.most_common()]
+        for language in sorted(inferred_languages):
+            if language not in ordered_languages:
+                ordered_languages.append(language)
         component = Component(
             path=relative_component,
-            languages=[name for name, _ in language_counts.most_common()],
+            languages=ordered_languages,
             frameworks=sorted(set(frameworks)),
             manifests=[path.relative_to(root_path).as_posix() for path in component_manifests],
             test_locations=_test_locations(owned_files, root_path, component_path),
@@ -501,7 +612,7 @@ def discover_repository(root: Path | str, *, probe_capabilities: bool = True) ->
         git_branch=_git(root_path, "branch", "--show-current"),
         git_head=_git(root_path, "rev-parse", "HEAD"),
         dirty_files=sorted(tracked_dirty | untracked_dirty),
-        inventory_file_count=len(files),
+        inventory_file_count=len(walked_files),
         capability_diagnostics=capability_diagnostics,
     )
 

{devagent_ai-0.6.0 → devagent_ai-0.7.0}/devagent/orchestrator.py

@@ -99,7 +99,58 @@ _WRITE_ACTION = {
     "required": ["tool", "arguments"],
     "additionalProperties": False,
 }
-_ACTION = {"anyOf": [_REPLACE_ACTION, _COUNTED_REPLACE_ACTION, _WRITE_ACTION]}
+_DELETE_ACTION = {
+    "type": "object",
+    "properties": {
+        "tool": {"type": "string", "const": "delete_file"},
+        "arguments": {
+            "type": "object",
+            "properties": {"path": _NON_EMPTY_STRING},
+            "required": ["path"],
+            "additionalProperties": False,
+        },
+    },
+    "required": ["tool", "arguments"],
+    "additionalProperties": False,
+}
+_MOVE_ACTION = {
+    "type": "object",
+    "properties": {
+        "tool": {"type": "string", "const": "move_file"},
+        "arguments": {
+            "type": "object",
+            "properties": {"source": _NON_EMPTY_STRING, "destination": _NON_EMPTY_STRING},
+            "required": ["source", "destination"],
+            "additionalProperties": False,
+        },
+    },
+    "required": ["tool", "arguments"],
+    "additionalProperties": False,
+}
+_RENAME_ACTION = {
+    "type": "object",
+    "properties": {
+        "tool": {"type": "string", "const": "rename_file"},
+        "arguments": {
+            "type": "object",
+            "properties": {"source": _NON_EMPTY_STRING, "destination": _NON_EMPTY_STRING},
+            "required": ["source", "destination"],
+            "additionalProperties": False,
+        },
+    },
+    "required": ["tool", "arguments"],
+    "additionalProperties": False,
+}
+_ACTION = {
+    "anyOf": [
+        _REPLACE_ACTION,
+        _COUNTED_REPLACE_ACTION,
+        _WRITE_ACTION,
+        _DELETE_ACTION,
+        _MOVE_ACTION,
+        _RENAME_ACTION,
+    ]
+}
 
 UNDERSTANDING_SCHEMA = {
     "type": "object",
@@ -310,12 +361,35 @@ def _execute_actions(workspace: Workspace, response: dict[str, Any], allowed_pat
     if not isinstance(actions, list) or not actions:
         raise OrchestrationError("Implementation requires at least one structured action")
     changed: list[str] = []
+    allowed_tools = {"replace_text", "write_file", "delete_file", "move_file", "rename_file"}
     for action in actions:
         if not isinstance(action, dict) or set(action) != {"tool", "arguments"} or not isinstance(action["arguments"], dict):
             raise OrchestrationError("Every action must contain only tool and arguments")
         tool, arguments = action["tool"], action["arguments"]
-        if tool not in {"replace_text", "write_file"}:
+        if tool not in allowed_tools:
             raise OrchestrationError(f"Implementation tool is not allowed: {tool}")
+
+        if tool in {"move_file", "rename_file"}:
+            if set(arguments) != {"source", "destination"}:
+                raise OrchestrationError(f"{tool} requires only string source and destination")
+            source = arguments.get("source")
+            destination = arguments.get("destination")
+            if (
+                not isinstance(source, str)
+                or not isinstance(destination, str)
+                or source not in allowed_paths
+                or destination not in allowed_paths
+            ):
+                raise OrchestrationError(
+                    f"Structural source/destination was not inspected/planned: {source} -> {destination}"
+                )
+            if tool == "move_file":
+                workspace.move_file(source, destination)
+            else:
+                workspace.rename_file(source, destination)
+            changed.extend((source, destination))
+            continue
+
         path = arguments.get("path")
         if not isinstance(path, str) or path not in allowed_paths:
             raise OrchestrationError(f"Action path was not inspected/planned: {path}")
@@ -329,12 +403,16 @@ def _execute_actions(workspace: Workspace, response: dict[str, Any], allowed_pat
             if isinstance(count, bool) or not isinstance(count, int) or count < 1:
                 raise OrchestrationError("replace_text count must be an integer greater than zero")
             workspace.replace_text(path, arguments["old"], arguments["new"], count)
-        else:
+        elif tool == "write_file":
             if set(arguments) != {"path", "content"} or not isinstance(arguments.get("content"), str):
                 raise OrchestrationError("write_file requires only string path and content")
             workspace.write_file(path, arguments["content"])
+        else:
+            if set(arguments) != {"path"}:
+                raise OrchestrationError("delete_file requires only string path")
+            workspace.delete_file(path)
         changed.append(path)
-    return changed
+    return list(dict.fromkeys(changed))
 
 
 def _diff(root: Path, workspace: Workspace, max_chars: int = 40_000) -> str:

{devagent_ai-0.6.0 → devagent_ai-0.7.0}/devagent/workspace.py

@@ -146,6 +146,57 @@ class Workspace:
         self.modified_paths.add(self._relative(target))
         self.artifacts.record("text_replaced", path=path, count=count, revision=self.revision)
 
+    def _reject_structural_symlinks(self, path: str) -> None:
+        lexical = self.root
+        for part in Path(path).parts:
+            lexical = lexical / part
+            if lexical.is_symlink():
+                raise SafetyError(f"Structural file operations do not follow symlinks: {path}")
+
+    def _structural_file(self, path: str) -> Path:
+        self._reject_structural_symlinks(path)
+        target = self._writable(path)
+        if not target.exists() or not target.is_file():
+            raise SafetyError(f"Structural operation requires an existing regular file: {path}")
+        return target
+
+    def delete_file(self, path: str) -> None:
+        """Delete one planned regular file after preserving a run-local backup."""
+        target = self._structural_file(path)
+        relative = self._relative(target)
+        self.artifacts.backup(target, relative_to=self.root)
+        target.unlink()
+        self.revision += 1
+        self.modified_paths.add(relative)
+        self.artifacts.record("file_deleted", path=relative, revision=self.revision)
+
+    def move_file(self, source: str, destination: str, *, operation: str = "move") -> None:
+        """Move/rename one planned regular file without overwriting existing content."""
+        source_target = self._structural_file(source)
+        self._reject_structural_symlinks(destination)
+        destination_target = self._writable(destination)
+        source_relative = self._relative(source_target)
+        destination_relative = self._relative(destination_target)
+        if source_relative == destination_relative:
+            raise SafetyError("Structural source and destination must be different")
+        if destination_target.exists():
+            raise SafetyError(f"Refusing to overwrite structural destination: {destination_relative}")
+        self.artifacts.backup(source_target, relative_to=self.root)
+        destination_target.parent.mkdir(parents=True, exist_ok=True)
+        source_target.replace(destination_target)
+        self.revision += 1
+        self.modified_paths.update({source_relative, destination_relative})
+        self.artifacts.record(
+            "file_moved",
+            source=source_relative,
+            destination=destination_relative,
+            operation=operation,
+            revision=self.revision,
+        )
+
+    def rename_file(self, source: str, destination: str) -> None:
+        self.move_file(source, destination, operation="rename")
+
     def _validate_python_requirement_file(self, target: Path) -> None:
         if target.stat().st_size > 1_000_000:
             raise SafetyError(f"Dependency requirement file is too large: {self._relative(target)}")
@@ -258,9 +309,6 @@ class Workspace:
             for key, value in os.environ.items()
             if key in {"PATH", "LANG", "LC_ALL", "TERM", "VIRTUAL_ENV", "PYTHONPATH", "SYSTEMROOT", "WINDIR"}
         }
-        # Keep HOME sandboxed so package-manager and cloud credentials are not inherited.
-        # Rustup-managed cargo/rustc binaries still need the installed toolchain directory,
-        # which is safe to expose separately from CARGO_HOME and registry credentials.
         rustup_home = os.environ.get("RUSTUP_HOME")
         if rustup_home:
             environment["RUSTUP_HOME"] = rustup_home
@@ -284,9 +332,6 @@ class Workspace:
             }
         )
         if dependency_install and Path(argv[0]).name.lower() == "yarn":
-            # Yarn Classic understands ignore-scripts; Yarn Modern uses enableScripts
-            # and rejects the old --ignore-scripts CLI flag. Supplying both environment
-            # settings is safe: each generation consumes the setting it understands.
             environment["YARN_IGNORE_SCRIPTS"] = "true"
             environment["YARN_ENABLE_SCRIPTS"] = "false"
         try:

{devagent_ai-0.6.0 → devagent_ai-0.7.0/devagent_ai.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: devagent-ai
-Version: 0.6.0
+Version: 0.7.0
 Summary: Evidence-driven local autonomous software engineering agent
 Author: Tom Ha
 Maintainer: Tom Ha

{devagent_ai-0.6.0 → devagent_ai-0.7.0}/devagent_ai.egg-info/SOURCES.txt

@@ -51,6 +51,8 @@ tests/test_evaluation_harness.py
 tests/test_evaluation_matrix.py
 tests/test_evaluation_regression_evidence.py
 tests/test_functional_qualification.py
+tests/test_huge_monorepo_v070.py
+tests/test_migration_e2e_v070.py
 tests/test_model_routing.py
 tests/test_multilang_technical_review.py
 tests/test_multistack_devagent_e2e.py
@@ -64,7 +66,10 @@ tests/test_retrieval.py
 tests/test_runtime_sandbox.py
 tests/test_safety_workspace.py
 tests/test_source_control_publish.py
+tests/test_structural_devagent_e2e_v070.py
+tests/test_structural_operations.py
 tests/test_structured_provider_contract.py
 tests/test_tasking_state.py
+tests/test_v070_engineering_breadth.py
 tests/test_workspace_environment.py
 tests/test_worktree.py

{devagent_ai-0.6.0 → devagent_ai-0.7.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "devagent-ai"
-version = "0.6.0"
+version = "0.7.0"
 description = "Evidence-driven local autonomous software engineering agent"
 readme = "README.md"
 requires-python = ">=3.10"

devagent_ai-0.7.0/tests/test_huge_monorepo_v070.py

@@ -0,0 +1,44 @@
+from __future__ import annotations
+
+import os
+import subprocess
+from pathlib import Path
+
+import pytest
+
+from devagent.discovery import discover_repository
+
+
+def _production_only() -> None:
+    if os.getenv("DEVAGENT_PRODUCTION_QUALIFICATION") != "1":
+        pytest.skip("huge-monorepo stress runs only through the production qualification gate")
+
+
+def test_deep_dotnet_manifest_is_recovered_beyond_12000_file_walk_frontier(tmp_path: Path) -> None:
+    _production_only()
+    filler = tmp_path / "aaa"
+    filler.mkdir()
+    for index in range(12_025):
+        (filler / f"f{index:05d}.txt").write_text("", encoding="utf-8")
+
+    deep = tmp_path / "zzz" / "service"
+    deep.mkdir(parents=True)
+    (deep / "Service.csproj").write_text(
+        '<Project Sdk="Microsoft.NET.Sdk"><PropertyGroup><TargetFramework>net8.0</TargetFramework>'
+        '</PropertyGroup></Project>\n',
+        encoding="utf-8",
+    )
+    (deep / "Service.cs").write_text(
+        "namespace Deep; public static class Service { public static int Value => 1; }\n",
+        encoding="utf-8",
+    )
+    subprocess.run(["git", "init", "-q"], cwd=tmp_path, check=True)
+    subprocess.run(["git", "add", "."], cwd=tmp_path, check=True)
+
+    repository = discover_repository(tmp_path, probe_capabilities=False)
+
+    assert repository.inventory_file_count == 12_000
+    component = next(item for item in repository.components if item.path == "zzz/service")
+    assert "zzz/service/Service.csproj" in component.manifests
+    assert "c#" in component.languages
+    assert any(item.command == ("dotnet", "build", "zzz/service/Service.csproj") for item in component.capabilities)