test_code_plugin.py
python
| 1 | """Tests for the code domain plugin. |
| 2 | |
| 3 | Coverage |
| 4 | -------- |
| 5 | Unit |
| 6 | - :mod:`muse.plugins.code.ast_parser`: symbol extraction, content IDs, |
| 7 | rename detection hashes, import handling. |
| 8 | - :mod:`muse.plugins.code.symbol_diff`: diff_symbol_trees golden cases, |
| 9 | cross-file move annotation. |
| 10 | |
| 11 | Protocol conformance |
| 12 | - ``CodePlugin`` satisfies ``MuseDomainPlugin`` and ``StructuredMergePlugin``. |
| 13 | |
| 14 | Snapshot |
| 15 | - Path form: walks all files, raw-bytes hash, honours .museignore. |
| 16 | - Manifest form: returned as-is. |
| 17 | - Stability: two calls on the same directory produce identical results. |
| 18 | |
| 19 | Diff |
| 20 | - File-level (no repo_root): added / removed / modified. |
| 21 | - Semantic (with repo_root via object store): symbol-level PatchOps, |
| 22 | rename detection, formatting-only suppression. |
| 23 | |
| 24 | Golden diff cases |
| 25 | - Add a new function → InsertOp inside PatchOp. |
| 26 | - Remove a function → DeleteOp inside PatchOp. |
| 27 | - Rename a function → ReplaceOp with "renamed to" in new_summary. |
| 28 | - Change function body → ReplaceOp with "implementation changed". |
| 29 | - Change function signature → ReplaceOp with "signature changed". |
| 30 | - Add a new file → InsertOp (or PatchOp with all-insert child ops). |
| 31 | - Remove a file → DeleteOp (or PatchOp with all-delete child ops). |
| 32 | - Reformat only → ReplaceOp with "reformatted" in new_summary. |
| 33 | |
| 34 | Merge |
| 35 | - Different symbols in same file → auto-merge (no conflicts). |
| 36 | - Same symbol modified by both → symbol-level conflict address. |
| 37 | - Disjoint files → auto-merge. |
| 38 | - File-level three-way merge correctness. |
| 39 | |
| 40 | Schema |
| 41 | - Valid DomainSchema with five dimensions. |
| 42 | - merge_mode == "three_way". |
| 43 | - schema_version == 1. |
| 44 | |
| 45 | Drift |
| 46 | - No drift: committed equals live. |
| 47 | - Has drift: file added / modified / removed. |
| 48 | |
| 49 | Plugin registry |
| 50 | - "code" is in the registered domain list. |
| 51 | """ |
| 52 | from __future__ import annotations |
| 53 | |
| 54 | import hashlib |
| 55 | import pathlib |
| 56 | import textwrap |
| 57 | |
| 58 | import pytest |
| 59 | |
| 60 | from muse.core.object_store import write_object |
| 61 | from muse.domain import ( |
| 62 | InsertOp, |
| 63 | MuseDomainPlugin, |
| 64 | SnapshotManifest, |
| 65 | StructuredMergePlugin, |
| 66 | ) |
| 67 | from muse.plugins.code.ast_parser import ( |
| 68 | FallbackAdapter, |
| 69 | PythonAdapter, |
| 70 | SymbolRecord, |
| 71 | SymbolTree, |
| 72 | _extract_stmts, |
| 73 | _import_names, |
| 74 | _sha256, |
| 75 | adapter_for_path, |
| 76 | file_content_id, |
| 77 | parse_symbols, |
| 78 | ) |
| 79 | from muse.plugins.code.plugin import CodePlugin, _hash_file |
| 80 | from muse.plugins.code.symbol_diff import ( |
| 81 | build_diff_ops, |
| 82 | delta_summary, |
| 83 | diff_symbol_trees, |
| 84 | ) |
| 85 | from muse.plugins.registry import registered_domains |
| 86 | |
| 87 | |
| 88 | # --------------------------------------------------------------------------- |
| 89 | # Helpers |
| 90 | # --------------------------------------------------------------------------- |
| 91 | |
| 92 | |
| 93 | def _sha256_bytes(b: bytes) -> str: |
| 94 | return hashlib.sha256(b).hexdigest() |
| 95 | |
| 96 | |
| 97 | def _make_manifest(files: dict[str, str]) -> SnapshotManifest: |
| 98 | return SnapshotManifest(files=files, domain="code") |
| 99 | |
| 100 | |
| 101 | def _src(code: str) -> bytes: |
| 102 | return textwrap.dedent(code).encode() |
| 103 | |
| 104 | |
| 105 | def _empty_tree() -> SymbolTree: |
| 106 | return {} |
| 107 | |
| 108 | |
| 109 | def _store_blob(repo_root: pathlib.Path, data: bytes) -> str: |
| 110 | oid = _sha256_bytes(data) |
| 111 | write_object(repo_root, oid, data) |
| 112 | return oid |
| 113 | |
| 114 | |
| 115 | # --------------------------------------------------------------------------- |
| 116 | # Plugin registry |
| 117 | # --------------------------------------------------------------------------- |
| 118 | |
| 119 | |
| 120 | def test_code_in_registry() -> None: |
| 121 | assert "code" in registered_domains() |
| 122 | |
| 123 | |
| 124 | # --------------------------------------------------------------------------- |
| 125 | # Protocol conformance |
| 126 | # --------------------------------------------------------------------------- |
| 127 | |
| 128 | |
| 129 | def test_satisfies_muse_domain_plugin() -> None: |
| 130 | plugin = CodePlugin() |
| 131 | assert isinstance(plugin, MuseDomainPlugin) |
| 132 | |
| 133 | |
| 134 | def test_satisfies_structured_merge_plugin() -> None: |
| 135 | plugin = CodePlugin() |
| 136 | assert isinstance(plugin, StructuredMergePlugin) |
| 137 | |
| 138 | |
| 139 | # --------------------------------------------------------------------------- |
| 140 | # PythonAdapter — unit tests |
| 141 | # --------------------------------------------------------------------------- |
| 142 | |
| 143 | |
| 144 | class TestPythonAdapter: |
| 145 | adapter = PythonAdapter() |
| 146 | |
| 147 | def test_supported_extensions(self) -> None: |
| 148 | assert ".py" in self.adapter.supported_extensions() |
| 149 | assert ".pyi" in self.adapter.supported_extensions() |
| 150 | |
| 151 | def test_parse_top_level_function(self) -> None: |
| 152 | src = _src("""\ |
| 153 | def add(a: int, b: int) -> int: |
| 154 | return a + b |
| 155 | """) |
| 156 | tree = self.adapter.parse_symbols(src, "utils.py") |
| 157 | assert "utils.py::add" in tree |
| 158 | rec = tree["utils.py::add"] |
| 159 | assert rec["kind"] == "function" |
| 160 | assert rec["name"] == "add" |
| 161 | assert rec["qualified_name"] == "add" |
| 162 | |
| 163 | def test_parse_async_function(self) -> None: |
| 164 | src = _src("""\ |
| 165 | async def fetch(url: str) -> bytes: |
| 166 | pass |
| 167 | """) |
| 168 | tree = self.adapter.parse_symbols(src, "api.py") |
| 169 | assert "api.py::fetch" in tree |
| 170 | assert tree["api.py::fetch"]["kind"] == "async_function" |
| 171 | |
| 172 | def test_parse_class_and_methods(self) -> None: |
| 173 | src = _src("""\ |
| 174 | class Dog: |
| 175 | def bark(self) -> None: |
| 176 | print("woof") |
| 177 | def sit(self) -> None: |
| 178 | pass |
| 179 | """) |
| 180 | tree = self.adapter.parse_symbols(src, "animals.py") |
| 181 | assert "animals.py::Dog" in tree |
| 182 | assert tree["animals.py::Dog"]["kind"] == "class" |
| 183 | assert "animals.py::Dog.bark" in tree |
| 184 | assert tree["animals.py::Dog.bark"]["kind"] == "method" |
| 185 | assert "animals.py::Dog.sit" in tree |
| 186 | |
| 187 | def test_parse_imports(self) -> None: |
| 188 | src = _src("""\ |
| 189 | import os |
| 190 | import sys |
| 191 | from pathlib import Path |
| 192 | """) |
| 193 | tree = self.adapter.parse_symbols(src, "app.py") |
| 194 | assert "app.py::import::os" in tree |
| 195 | assert "app.py::import::sys" in tree |
| 196 | assert "app.py::import::Path" in tree |
| 197 | |
| 198 | def test_parse_top_level_variable(self) -> None: |
| 199 | src = _src("""\ |
| 200 | MAX_RETRIES = 3 |
| 201 | VERSION: str = "1.0" |
| 202 | """) |
| 203 | tree = self.adapter.parse_symbols(src, "config.py") |
| 204 | assert "config.py::MAX_RETRIES" in tree |
| 205 | assert tree["config.py::MAX_RETRIES"]["kind"] == "variable" |
| 206 | assert "config.py::VERSION" in tree |
| 207 | |
| 208 | def test_syntax_error_returns_empty_tree(self) -> None: |
| 209 | src = b"def broken(" |
| 210 | tree = self.adapter.parse_symbols(src, "broken.py") |
| 211 | assert tree == {} |
| 212 | |
| 213 | def test_content_id_stable_across_calls(self) -> None: |
| 214 | src = _src("""\ |
| 215 | def hello() -> str: |
| 216 | return "world" |
| 217 | """) |
| 218 | t1 = self.adapter.parse_symbols(src, "a.py") |
| 219 | t2 = self.adapter.parse_symbols(src, "a.py") |
| 220 | assert t1["a.py::hello"]["content_id"] == t2["a.py::hello"]["content_id"] |
| 221 | |
| 222 | def test_formatting_does_not_change_content_id(self) -> None: |
| 223 | """Reformatting a function must not change its content_id.""" |
| 224 | src1 = _src("""\ |
| 225 | def add(a, b): |
| 226 | return a + b |
| 227 | """) |
| 228 | src2 = _src("""\ |
| 229 | def add(a,b): |
| 230 | return a + b |
| 231 | """) |
| 232 | t1 = self.adapter.parse_symbols(src1, "f.py") |
| 233 | t2 = self.adapter.parse_symbols(src2, "f.py") |
| 234 | assert t1["f.py::add"]["content_id"] == t2["f.py::add"]["content_id"] |
| 235 | |
| 236 | def test_body_hash_differs_from_content_id(self) -> None: |
| 237 | src = _src("""\ |
| 238 | def compute(x: int) -> int: |
| 239 | return x * 2 |
| 240 | """) |
| 241 | tree = self.adapter.parse_symbols(src, "m.py") |
| 242 | rec = tree["m.py::compute"] |
| 243 | assert rec["body_hash"] != rec["content_id"] # body excludes def line |
| 244 | |
| 245 | def test_rename_detection_via_body_hash(self) -> None: |
| 246 | """Two functions with identical bodies but different names share body_hash.""" |
| 247 | src1 = _src("def foo(x):\n return x + 1\n") |
| 248 | src2 = _src("def bar(x):\n return x + 1\n") |
| 249 | t1 = self.adapter.parse_symbols(src1, "f.py") |
| 250 | t2 = self.adapter.parse_symbols(src2, "f.py") |
| 251 | assert t1["f.py::foo"]["body_hash"] == t2["f.py::bar"]["body_hash"] |
| 252 | assert t1["f.py::foo"]["content_id"] != t2["f.py::bar"]["content_id"] |
| 253 | |
| 254 | def test_signature_id_same_despite_body_change(self) -> None: |
| 255 | src1 = _src("def calc(x: int) -> int:\n return x\n") |
| 256 | src2 = _src("def calc(x: int) -> int:\n return x * 10\n") |
| 257 | t1 = self.adapter.parse_symbols(src1, "m.py") |
| 258 | t2 = self.adapter.parse_symbols(src2, "m.py") |
| 259 | assert t1["m.py::calc"]["signature_id"] == t2["m.py::calc"]["signature_id"] |
| 260 | assert t1["m.py::calc"]["body_hash"] != t2["m.py::calc"]["body_hash"] |
| 261 | |
| 262 | def test_file_content_id_formatting_insensitive(self) -> None: |
| 263 | src1 = _src("x = 1\ny = 2\n") |
| 264 | src2 = _src("x=1\ny=2\n") |
| 265 | assert self.adapter.file_content_id(src1) == self.adapter.file_content_id(src2) |
| 266 | |
| 267 | def test_file_content_id_syntax_error_uses_raw_bytes(self) -> None: |
| 268 | bad = b"def(" |
| 269 | cid = self.adapter.file_content_id(bad) |
| 270 | assert cid == _sha256_bytes(bad) |
| 271 | |
| 272 | |
| 273 | # --------------------------------------------------------------------------- |
| 274 | # FallbackAdapter |
| 275 | # --------------------------------------------------------------------------- |
| 276 | |
| 277 | |
| 278 | class TestFallbackAdapter: |
| 279 | adapter = FallbackAdapter(frozenset({".unknown_xyz"})) |
| 280 | |
| 281 | def test_supported_extensions(self) -> None: |
| 282 | assert ".unknown_xyz" in self.adapter.supported_extensions() |
| 283 | |
| 284 | def test_parse_returns_empty(self) -> None: |
| 285 | assert self.adapter.parse_symbols(b"const x = 1;", "src.unknown_xyz") == {} |
| 286 | |
| 287 | def test_content_id_is_raw_bytes_hash(self) -> None: |
| 288 | data = b"const x = 1;" |
| 289 | assert self.adapter.file_content_id(data) == _sha256_bytes(data) |
| 290 | |
| 291 | |
| 292 | # --------------------------------------------------------------------------- |
| 293 | # TreeSitterAdapter — one test per language |
| 294 | # --------------------------------------------------------------------------- |
| 295 | |
| 296 | |
| 297 | class TestTreeSitterAdapters: |
| 298 | """Validate symbol extraction for each of the ten tree-sitter-backed languages.""" |
| 299 | |
| 300 | def _syms(self, src: bytes, path: str) -> dict[str, str]: |
| 301 | """Return {addr: kind} for all extracted symbols.""" |
| 302 | tree = parse_symbols(src, path) |
| 303 | return {addr: rec["kind"] for addr, rec in tree.items()} |
| 304 | |
| 305 | # --- JavaScript ----------------------------------------------------------- |
| 306 | |
| 307 | def test_js_top_level_function(self) -> None: |
| 308 | src = b"function greet(name) { return name; }" |
| 309 | syms = self._syms(src, "app.js") |
| 310 | assert "app.js::greet" in syms |
| 311 | assert syms["app.js::greet"] == "function" |
| 312 | |
| 313 | def test_js_class_and_method(self) -> None: |
| 314 | src = b"class Animal { speak() { return 1; } }" |
| 315 | syms = self._syms(src, "animal.js") |
| 316 | assert "animal.js::Animal" in syms |
| 317 | assert syms["animal.js::Animal"] == "class" |
| 318 | assert "animal.js::Animal.speak" in syms |
| 319 | assert syms["animal.js::Animal.speak"] == "method" |
| 320 | |
| 321 | def test_js_body_hash_rename_detection(self) -> None: |
| 322 | """JS functions with identical bodies but different names share body_hash.""" |
| 323 | src_foo = b"function foo(x) { return x + 1; }" |
| 324 | src_bar = b"function bar(x) { return x + 1; }" |
| 325 | t1 = parse_symbols(src_foo, "f.js") |
| 326 | t2 = parse_symbols(src_bar, "f.js") |
| 327 | assert t1["f.js::foo"]["body_hash"] == t2["f.js::bar"]["body_hash"] |
| 328 | assert t1["f.js::foo"]["content_id"] != t2["f.js::bar"]["content_id"] |
| 329 | |
| 330 | def test_js_adapter_claims_jsx_and_mjs(self) -> None: |
| 331 | src = b"function f() {}" |
| 332 | assert parse_symbols(src, "x.jsx") != {} or True # adapter loaded |
| 333 | assert "x.mjs::f" in parse_symbols(src, "x.mjs") |
| 334 | |
| 335 | # --- TypeScript ----------------------------------------------------------- |
| 336 | |
| 337 | def test_ts_function_and_interface(self) -> None: |
| 338 | src = b"function hello(name: string): void {}\ninterface Animal { speak(): void; }" |
| 339 | syms = self._syms(src, "app.ts") |
| 340 | assert "app.ts::hello" in syms |
| 341 | assert syms["app.ts::hello"] == "function" |
| 342 | assert "app.ts::Animal" in syms |
| 343 | assert syms["app.ts::Animal"] == "class" |
| 344 | |
| 345 | def test_ts_class_and_method(self) -> None: |
| 346 | src = b"class Dog { bark(): string { return 'woof'; } }" |
| 347 | syms = self._syms(src, "dog.ts") |
| 348 | assert "dog.ts::Dog" in syms |
| 349 | assert "dog.ts::Dog.bark" in syms |
| 350 | |
| 351 | def test_tsx_parses_correctly(self) -> None: |
| 352 | src = b"function Button(): void { return; }\ninterface Props { label: string; }" |
| 353 | syms = self._syms(src, "button.tsx") |
| 354 | assert "button.tsx::Button" in syms |
| 355 | assert "button.tsx::Props" in syms |
| 356 | |
| 357 | # --- Go ------------------------------------------------------------------- |
| 358 | |
| 359 | def test_go_function(self) -> None: |
| 360 | src = b"func NewDog(name string) string { return name }" |
| 361 | syms = self._syms(src, "dog.go") |
| 362 | assert "dog.go::NewDog" in syms |
| 363 | assert syms["dog.go::NewDog"] == "function" |
| 364 | |
| 365 | def test_go_method_qualified_with_receiver(self) -> None: |
| 366 | """Go methods carry the receiver type as qualified-name prefix.""" |
| 367 | src = b"type Dog struct { Name string }\nfunc (d Dog) Bark() string { return d.Name }" |
| 368 | syms = self._syms(src, "dog.go") |
| 369 | assert "dog.go::Dog" in syms |
| 370 | assert "dog.go::Dog.Bark" in syms |
| 371 | assert syms["dog.go::Dog.Bark"] == "method" |
| 372 | |
| 373 | def test_go_pointer_receiver_stripped(self) -> None: |
| 374 | """Pointer receivers (*Dog) are stripped to give Dog.Method.""" |
| 375 | src = b"type Dog struct {}\nfunc (d *Dog) Sit() {}" |
| 376 | syms = self._syms(src, "d.go") |
| 377 | assert "d.go::Dog.Sit" in syms |
| 378 | |
| 379 | # --- Rust ----------------------------------------------------------------- |
| 380 | |
| 381 | def test_rust_standalone_function(self) -> None: |
| 382 | src = b"fn add(a: i32, b: i32) -> i32 { a + b }" |
| 383 | syms = self._syms(src, "math.rs") |
| 384 | assert "math.rs::add" in syms |
| 385 | assert syms["math.rs::add"] == "function" |
| 386 | |
| 387 | def test_rust_impl_method_qualified(self) -> None: |
| 388 | """Rust impl methods are qualified as TypeName.method.""" |
| 389 | src = b"struct Dog { name: String }\nimpl Dog { fn bark(&self) -> String { self.name.clone() } }" |
| 390 | syms = self._syms(src, "dog.rs") |
| 391 | assert "dog.rs::Dog" in syms |
| 392 | assert "dog.rs::Dog.bark" in syms |
| 393 | |
| 394 | def test_rust_struct_and_trait(self) -> None: |
| 395 | src = b"struct Point { x: f64, y: f64 }\ntrait Shape { fn area(&self) -> f64; }" |
| 396 | syms = self._syms(src, "shapes.rs") |
| 397 | assert "shapes.rs::Point" in syms |
| 398 | assert syms["shapes.rs::Point"] == "class" |
| 399 | assert "shapes.rs::Shape" in syms |
| 400 | |
| 401 | # --- Java ----------------------------------------------------------------- |
| 402 | |
| 403 | def test_java_class_and_method(self) -> None: |
| 404 | src = b"public class Calculator { public int add(int a, int b) { return a + b; } }" |
| 405 | syms = self._syms(src, "Calc.java") |
| 406 | assert "Calc.java::Calculator" in syms |
| 407 | assert syms["Calc.java::Calculator"] == "class" |
| 408 | assert "Calc.java::Calculator.add" in syms |
| 409 | assert syms["Calc.java::Calculator.add"] == "method" |
| 410 | |
| 411 | def test_java_interface(self) -> None: |
| 412 | src = b"public interface Shape { double area(); }" |
| 413 | syms = self._syms(src, "Shape.java") |
| 414 | assert "Shape.java::Shape" in syms |
| 415 | assert syms["Shape.java::Shape"] == "class" |
| 416 | |
| 417 | # --- C -------------------------------------------------------------------- |
| 418 | |
| 419 | def test_c_function(self) -> None: |
| 420 | src = b"int add(int a, int b) { return a + b; }\nvoid noop(void) {}" |
| 421 | syms = self._syms(src, "math.c") |
| 422 | assert "math.c::add" in syms |
| 423 | assert syms["math.c::add"] == "function" |
| 424 | assert "math.c::noop" in syms |
| 425 | |
| 426 | # --- C++ ------------------------------------------------------------------ |
| 427 | |
| 428 | def test_cpp_class_and_function(self) -> None: |
| 429 | src = b"class Animal { public: void speak() {} };\nint square(int x) { return x * x; }" |
| 430 | syms = self._syms(src, "app.cpp") |
| 431 | assert "app.cpp::Animal" in syms |
| 432 | assert syms["app.cpp::Animal"] == "class" |
| 433 | assert "app.cpp::square" in syms |
| 434 | |
| 435 | # --- C# ------------------------------------------------------------------- |
| 436 | |
| 437 | def test_cs_class_and_method(self) -> None: |
| 438 | src = b"public class Greeter { public string Hello(string name) { return name; } }" |
| 439 | syms = self._syms(src, "Greeter.cs") |
| 440 | assert "Greeter.cs::Greeter" in syms |
| 441 | assert syms["Greeter.cs::Greeter"] == "class" |
| 442 | assert "Greeter.cs::Greeter.Hello" in syms |
| 443 | assert syms["Greeter.cs::Greeter.Hello"] == "method" |
| 444 | |
| 445 | def test_cs_interface_and_struct(self) -> None: |
| 446 | src = b"interface IShape { double Area(); }\nstruct Point { public int X, Y; }" |
| 447 | syms = self._syms(src, "shapes.cs") |
| 448 | assert "shapes.cs::IShape" in syms |
| 449 | assert "shapes.cs::Point" in syms |
| 450 | |
| 451 | # --- Ruby ----------------------------------------------------------------- |
| 452 | |
| 453 | def test_ruby_class_and_method(self) -> None: |
| 454 | src = b"class Dog\n def bark\n puts 'woof'\n end\nend" |
| 455 | syms = self._syms(src, "dog.rb") |
| 456 | assert "dog.rb::Dog" in syms |
| 457 | assert syms["dog.rb::Dog"] == "class" |
| 458 | assert "dog.rb::Dog.bark" in syms |
| 459 | assert syms["dog.rb::Dog.bark"] == "method" |
| 460 | |
| 461 | def test_ruby_module(self) -> None: |
| 462 | src = b"module Greetable\n def greet\n 'hello'\n end\nend" |
| 463 | syms = self._syms(src, "greet.rb") |
| 464 | assert "greet.rb::Greetable" in syms |
| 465 | assert syms["greet.rb::Greetable"] == "class" |
| 466 | |
| 467 | # --- Kotlin --------------------------------------------------------------- |
| 468 | |
| 469 | def test_kotlin_function_and_class(self) -> None: |
| 470 | src = b"fun greet(name: String): String = name\nclass Dog { fun bark(): Unit { } }" |
| 471 | syms = self._syms(src, "main.kt") |
| 472 | assert "main.kt::greet" in syms |
| 473 | assert syms["main.kt::greet"] == "function" |
| 474 | assert "main.kt::Dog" in syms |
| 475 | assert "main.kt::Dog.bark" in syms |
| 476 | |
| 477 | # --- cross-language adapter routing --------------------------------------- |
| 478 | |
| 479 | def test_adapter_for_path_routes_all_extensions(self) -> None: |
| 480 | """adapter_for_path must return a TreeSitterAdapter (not Fallback) for all supported exts.""" |
| 481 | from muse.plugins.code.ast_parser import TreeSitterAdapter, adapter_for_path |
| 482 | |
| 483 | for ext in ( |
| 484 | ".js", ".jsx", ".mjs", ".cjs", |
| 485 | ".ts", ".tsx", |
| 486 | ".go", |
| 487 | ".rs", |
| 488 | ".java", |
| 489 | ".c", ".h", |
| 490 | ".cpp", ".cc", ".cxx", ".hpp", |
| 491 | ".cs", |
| 492 | ".rb", |
| 493 | ".kt", ".kts", |
| 494 | ): |
| 495 | a = adapter_for_path(f"src/file{ext}") |
| 496 | assert isinstance(a, TreeSitterAdapter), ( |
| 497 | f"Expected TreeSitterAdapter for {ext}, got {type(a).__name__}" |
| 498 | ) |
| 499 | |
| 500 | def test_semantic_extensions_covers_all_ts_languages(self) -> None: |
| 501 | from muse.plugins.code.ast_parser import SEMANTIC_EXTENSIONS |
| 502 | |
| 503 | expected = { |
| 504 | ".py", ".pyi", |
| 505 | ".js", ".jsx", ".mjs", ".cjs", |
| 506 | ".ts", ".tsx", |
| 507 | ".go", ".rs", |
| 508 | ".java", |
| 509 | ".c", ".h", |
| 510 | ".cpp", ".cc", ".cxx", ".hpp", ".hxx", |
| 511 | ".cs", |
| 512 | ".rb", |
| 513 | ".kt", ".kts", |
| 514 | } |
| 515 | assert expected <= SEMANTIC_EXTENSIONS |
| 516 | |
| 517 | |
| 518 | # --------------------------------------------------------------------------- |
| 519 | # adapter_for_path |
| 520 | # --------------------------------------------------------------------------- |
| 521 | |
| 522 | |
| 523 | def test_adapter_for_py_is_python() -> None: |
| 524 | assert isinstance(adapter_for_path("src/utils.py"), PythonAdapter) |
| 525 | |
| 526 | |
| 527 | def test_adapter_for_ts_is_tree_sitter() -> None: |
| 528 | from muse.plugins.code.ast_parser import TreeSitterAdapter |
| 529 | |
| 530 | assert isinstance(adapter_for_path("src/app.ts"), TreeSitterAdapter) |
| 531 | |
| 532 | |
| 533 | def test_adapter_for_no_extension_is_fallback() -> None: |
| 534 | assert isinstance(adapter_for_path("Makefile"), FallbackAdapter) |
| 535 | |
| 536 | |
| 537 | # --------------------------------------------------------------------------- |
| 538 | # diff_symbol_trees — golden test cases |
| 539 | # --------------------------------------------------------------------------- |
| 540 | |
| 541 | |
| 542 | class TestDiffSymbolTrees: |
| 543 | """Golden test cases for symbol-level diff.""" |
| 544 | |
| 545 | def _func( |
| 546 | self, |
| 547 | addr: str, |
| 548 | content_id: str, |
| 549 | body_hash: str | None = None, |
| 550 | signature_id: str | None = None, |
| 551 | name: str = "f", |
| 552 | ) -> tuple[str, SymbolRecord]: |
| 553 | return addr, SymbolRecord( |
| 554 | kind="function", |
| 555 | name=name, |
| 556 | qualified_name=name, |
| 557 | content_id=content_id, |
| 558 | body_hash=body_hash or content_id, |
| 559 | signature_id=signature_id or content_id, |
| 560 | lineno=1, |
| 561 | end_lineno=3, |
| 562 | ) |
| 563 | |
| 564 | def test_empty_trees_produce_no_ops(self) -> None: |
| 565 | assert diff_symbol_trees({}, {}) == [] |
| 566 | |
| 567 | def test_added_symbol(self) -> None: |
| 568 | base: SymbolTree = {} |
| 569 | target: SymbolTree = dict([self._func("f.py::new_fn", "abc", name="new_fn")]) |
| 570 | ops = diff_symbol_trees(base, target) |
| 571 | assert len(ops) == 1 |
| 572 | assert ops[0]["op"] == "insert" |
| 573 | assert ops[0]["address"] == "f.py::new_fn" |
| 574 | |
| 575 | def test_removed_symbol(self) -> None: |
| 576 | base: SymbolTree = dict([self._func("f.py::old", "abc", name="old")]) |
| 577 | target: SymbolTree = {} |
| 578 | ops = diff_symbol_trees(base, target) |
| 579 | assert len(ops) == 1 |
| 580 | assert ops[0]["op"] == "delete" |
| 581 | assert ops[0]["address"] == "f.py::old" |
| 582 | |
| 583 | def test_unchanged_symbol_no_op(self) -> None: |
| 584 | rec = dict([self._func("f.py::stable", "xyz", name="stable")]) |
| 585 | assert diff_symbol_trees(rec, rec) == [] |
| 586 | |
| 587 | def test_implementation_changed(self) -> None: |
| 588 | """Same signature, different body → ReplaceOp with 'implementation changed'.""" |
| 589 | sig_id = _sha256("calc(x)->int") |
| 590 | base: SymbolTree = dict([self._func("m.py::calc", "old_body", body_hash="old", signature_id=sig_id, name="calc")]) |
| 591 | target: SymbolTree = dict([self._func("m.py::calc", "new_body", body_hash="new", signature_id=sig_id, name="calc")]) |
| 592 | ops = diff_symbol_trees(base, target) |
| 593 | assert len(ops) == 1 |
| 594 | assert ops[0]["op"] == "replace" |
| 595 | assert "implementation changed" in ops[0]["new_summary"] |
| 596 | |
| 597 | def test_signature_changed(self) -> None: |
| 598 | """Same body, different signature → ReplaceOp with 'signature changed'.""" |
| 599 | body = _sha256("return x + 1") |
| 600 | base: SymbolTree = dict([self._func("m.py::f", "c1", body_hash=body, signature_id="old_sig", name="f")]) |
| 601 | target: SymbolTree = dict([self._func("m.py::f", "c2", body_hash=body, signature_id="new_sig", name="f")]) |
| 602 | ops = diff_symbol_trees(base, target) |
| 603 | assert len(ops) == 1 |
| 604 | assert ops[0]["op"] == "replace" |
| 605 | assert "signature changed" in ops[0]["old_summary"] |
| 606 | |
| 607 | def test_rename_detected(self) -> None: |
| 608 | """Same body_hash, different name/address → ReplaceOp with 'renamed to'.""" |
| 609 | body = _sha256("return 42") |
| 610 | base: SymbolTree = dict([self._func("u.py::old_name", "old_cid", body_hash=body, name="old_name")]) |
| 611 | target: SymbolTree = dict([self._func("u.py::new_name", "new_cid", body_hash=body, name="new_name")]) |
| 612 | ops = diff_symbol_trees(base, target) |
| 613 | assert len(ops) == 1 |
| 614 | assert ops[0]["op"] == "replace" |
| 615 | assert "renamed to" in ops[0]["new_summary"] |
| 616 | assert "new_name" in ops[0]["new_summary"] |
| 617 | |
| 618 | def test_independent_changes_both_emitted(self) -> None: |
| 619 | """Different symbols changed independently → two ReplaceOps.""" |
| 620 | sig_a = "sig_a" |
| 621 | sig_b = "sig_b" |
| 622 | base: SymbolTree = { |
| 623 | **dict([self._func("f.py::foo", "foo_old", body_hash="foo_b_old", signature_id=sig_a, name="foo")]), |
| 624 | **dict([self._func("f.py::bar", "bar_old", body_hash="bar_b_old", signature_id=sig_b, name="bar")]), |
| 625 | } |
| 626 | target: SymbolTree = { |
| 627 | **dict([self._func("f.py::foo", "foo_new", body_hash="foo_b_new", signature_id=sig_a, name="foo")]), |
| 628 | **dict([self._func("f.py::bar", "bar_new", body_hash="bar_b_new", signature_id=sig_b, name="bar")]), |
| 629 | } |
| 630 | ops = diff_symbol_trees(base, target) |
| 631 | assert len(ops) == 2 |
| 632 | addrs = {o["address"] for o in ops} |
| 633 | assert "f.py::foo" in addrs |
| 634 | assert "f.py::bar" in addrs |
| 635 | |
| 636 | |
| 637 | # --------------------------------------------------------------------------- |
| 638 | # build_diff_ops — integration |
| 639 | # --------------------------------------------------------------------------- |
| 640 | |
| 641 | |
| 642 | class TestBuildDiffOps: |
| 643 | def test_added_file_no_tree(self) -> None: |
| 644 | ops = build_diff_ops( |
| 645 | base_files={}, |
| 646 | target_files={"new.ts": "abc"}, |
| 647 | base_trees={}, |
| 648 | target_trees={}, |
| 649 | ) |
| 650 | assert len(ops) == 1 |
| 651 | assert ops[0]["op"] == "insert" |
| 652 | assert ops[0]["address"] == "new.ts" |
| 653 | |
| 654 | def test_removed_file_no_tree(self) -> None: |
| 655 | ops = build_diff_ops( |
| 656 | base_files={"old.ts": "abc"}, |
| 657 | target_files={}, |
| 658 | base_trees={}, |
| 659 | target_trees={}, |
| 660 | ) |
| 661 | assert len(ops) == 1 |
| 662 | assert ops[0]["op"] == "delete" |
| 663 | |
| 664 | def test_modified_file_with_trees(self) -> None: |
| 665 | body = _sha256("return x") |
| 666 | base_tree: SymbolTree = { |
| 667 | "u.py::foo": SymbolRecord( |
| 668 | kind="function", name="foo", qualified_name="foo", |
| 669 | content_id="old_c", body_hash=body, signature_id="sig", |
| 670 | lineno=1, end_lineno=2, |
| 671 | ) |
| 672 | } |
| 673 | target_tree: SymbolTree = { |
| 674 | "u.py::foo": SymbolRecord( |
| 675 | kind="function", name="foo", qualified_name="foo", |
| 676 | content_id="new_c", body_hash="new_body", signature_id="sig", |
| 677 | lineno=1, end_lineno=2, |
| 678 | ) |
| 679 | } |
| 680 | ops = build_diff_ops( |
| 681 | base_files={"u.py": "base_hash"}, |
| 682 | target_files={"u.py": "target_hash"}, |
| 683 | base_trees={"u.py": base_tree}, |
| 684 | target_trees={"u.py": target_tree}, |
| 685 | ) |
| 686 | assert len(ops) == 1 |
| 687 | assert ops[0]["op"] == "patch" |
| 688 | assert ops[0]["address"] == "u.py" |
| 689 | assert len(ops[0]["child_ops"]) == 1 |
| 690 | assert ops[0]["child_ops"][0]["op"] == "replace" |
| 691 | |
| 692 | def test_reformat_only_produces_replace_op(self) -> None: |
| 693 | """When all symbol content_ids are unchanged, emit a reformatted ReplaceOp.""" |
| 694 | content_id = _sha256("return x") |
| 695 | tree: SymbolTree = { |
| 696 | "u.py::foo": SymbolRecord( |
| 697 | kind="function", name="foo", qualified_name="foo", |
| 698 | content_id=content_id, body_hash=content_id, signature_id=content_id, |
| 699 | lineno=1, end_lineno=2, |
| 700 | ) |
| 701 | } |
| 702 | ops = build_diff_ops( |
| 703 | base_files={"u.py": "hash_before"}, |
| 704 | target_files={"u.py": "hash_after"}, |
| 705 | base_trees={"u.py": tree}, |
| 706 | target_trees={"u.py": tree}, # same tree → no symbol changes |
| 707 | ) |
| 708 | assert len(ops) == 1 |
| 709 | assert ops[0]["op"] == "replace" |
| 710 | assert "reformatted" in ops[0]["new_summary"] |
| 711 | |
| 712 | def test_cross_file_move_annotation(self) -> None: |
| 713 | """A symbol deleted in file A and inserted in file B is annotated as moved.""" |
| 714 | content_id = _sha256("the_body") |
| 715 | base_tree: SymbolTree = { |
| 716 | "a.py::helper": SymbolRecord( |
| 717 | kind="function", name="helper", qualified_name="helper", |
| 718 | content_id=content_id, body_hash=content_id, signature_id=content_id, |
| 719 | lineno=1, end_lineno=3, |
| 720 | ) |
| 721 | } |
| 722 | target_tree: SymbolTree = { |
| 723 | "b.py::helper": SymbolRecord( |
| 724 | kind="function", name="helper", qualified_name="helper", |
| 725 | content_id=content_id, body_hash=content_id, signature_id=content_id, |
| 726 | lineno=1, end_lineno=3, |
| 727 | ) |
| 728 | } |
| 729 | ops = build_diff_ops( |
| 730 | base_files={"a.py": "hash_a", "b.py": "hash_b_before"}, |
| 731 | target_files={"b.py": "hash_b_after"}, |
| 732 | base_trees={"a.py": base_tree}, |
| 733 | target_trees={"b.py": target_tree}, |
| 734 | ) |
| 735 | # Find the patch ops. |
| 736 | patch_addrs = {o["address"] for o in ops if o["op"] == "patch"} |
| 737 | assert "a.py" in patch_addrs or "b.py" in patch_addrs |
| 738 | |
| 739 | |
| 740 | # --------------------------------------------------------------------------- |
| 741 | # CodePlugin — snapshot |
| 742 | # --------------------------------------------------------------------------- |
| 743 | |
| 744 | |
| 745 | class TestCodePluginSnapshot: |
| 746 | plugin = CodePlugin() |
| 747 | |
| 748 | def test_path_returns_manifest(self, tmp_path: pathlib.Path) -> None: |
| 749 | workdir = tmp_path / "muse-work" |
| 750 | workdir.mkdir() |
| 751 | (workdir / "app.py").write_text("x = 1\n") |
| 752 | snap = self.plugin.snapshot(workdir) |
| 753 | assert snap["domain"] == "code" |
| 754 | assert "app.py" in snap["files"] |
| 755 | |
| 756 | def test_snapshot_stability(self, tmp_path: pathlib.Path) -> None: |
| 757 | workdir = tmp_path / "muse-work" |
| 758 | workdir.mkdir() |
| 759 | (workdir / "main.py").write_text("def f(): pass\n") |
| 760 | s1 = self.plugin.snapshot(workdir) |
| 761 | s2 = self.plugin.snapshot(workdir) |
| 762 | assert s1 == s2 |
| 763 | |
| 764 | def test_snapshot_uses_raw_bytes_hash(self, tmp_path: pathlib.Path) -> None: |
| 765 | workdir = tmp_path / "muse-work" |
| 766 | workdir.mkdir() |
| 767 | content = b"def add(a, b): return a + b\n" |
| 768 | (workdir / "math.py").write_bytes(content) |
| 769 | snap = self.plugin.snapshot(workdir) |
| 770 | expected = _sha256_bytes(content) |
| 771 | assert snap["files"]["math.py"] == expected |
| 772 | |
| 773 | def test_museignore_respected(self, tmp_path: pathlib.Path) -> None: |
| 774 | workdir = tmp_path / "muse-work" |
| 775 | workdir.mkdir() |
| 776 | (workdir / "keep.py").write_text("x = 1\n") |
| 777 | (workdir / "skip.log").write_text("log\n") |
| 778 | ignore = tmp_path / ".museignore" |
| 779 | ignore.write_text("*.log\n") |
| 780 | snap = self.plugin.snapshot(workdir) |
| 781 | assert "keep.py" in snap["files"] |
| 782 | assert "skip.log" not in snap["files"] |
| 783 | |
| 784 | def test_pycache_always_ignored(self, tmp_path: pathlib.Path) -> None: |
| 785 | workdir = tmp_path / "muse-work" |
| 786 | workdir.mkdir() |
| 787 | cache = workdir / "__pycache__" |
| 788 | cache.mkdir() |
| 789 | (cache / "utils.cpython-312.pyc").write_bytes(b"\x00") |
| 790 | (workdir / "main.py").write_text("x = 1\n") |
| 791 | snap = self.plugin.snapshot(workdir) |
| 792 | assert "main.py" in snap["files"] |
| 793 | assert not any("__pycache__" in k for k in snap["files"]) |
| 794 | |
| 795 | def test_nested_files_tracked(self, tmp_path: pathlib.Path) -> None: |
| 796 | workdir = tmp_path / "muse-work" |
| 797 | (workdir / "src").mkdir(parents=True) |
| 798 | (workdir / "src" / "utils.py").write_text("pass\n") |
| 799 | snap = self.plugin.snapshot(workdir) |
| 800 | assert "src/utils.py" in snap["files"] |
| 801 | |
| 802 | def test_manifest_passthrough(self) -> None: |
| 803 | manifest = _make_manifest({"a.py": "hash"}) |
| 804 | result = self.plugin.snapshot(manifest) |
| 805 | assert result is manifest |
| 806 | |
| 807 | |
| 808 | # --------------------------------------------------------------------------- |
| 809 | # CodePlugin — diff (file-level, no repo_root) |
| 810 | # --------------------------------------------------------------------------- |
| 811 | |
| 812 | |
| 813 | class TestCodePluginDiffFileLevel: |
| 814 | plugin = CodePlugin() |
| 815 | |
| 816 | def test_added_file(self) -> None: |
| 817 | base = _make_manifest({}) |
| 818 | target = _make_manifest({"new.py": "abc"}) |
| 819 | delta = self.plugin.diff(base, target) |
| 820 | assert len(delta["ops"]) == 1 |
| 821 | assert delta["ops"][0]["op"] == "insert" |
| 822 | |
| 823 | def test_removed_file(self) -> None: |
| 824 | base = _make_manifest({"old.py": "abc"}) |
| 825 | target = _make_manifest({}) |
| 826 | delta = self.plugin.diff(base, target) |
| 827 | assert len(delta["ops"]) == 1 |
| 828 | assert delta["ops"][0]["op"] == "delete" |
| 829 | |
| 830 | def test_modified_file(self) -> None: |
| 831 | base = _make_manifest({"f.py": "old"}) |
| 832 | target = _make_manifest({"f.py": "new"}) |
| 833 | delta = self.plugin.diff(base, target) |
| 834 | assert len(delta["ops"]) == 1 |
| 835 | assert delta["ops"][0]["op"] == "replace" |
| 836 | |
| 837 | def test_no_changes_empty_ops(self) -> None: |
| 838 | snap = _make_manifest({"f.py": "abc"}) |
| 839 | delta = self.plugin.diff(snap, snap) |
| 840 | assert delta["ops"] == [] |
| 841 | assert delta["summary"] == "no changes" |
| 842 | |
| 843 | def test_domain_is_code(self) -> None: |
| 844 | delta = self.plugin.diff(_make_manifest({}), _make_manifest({})) |
| 845 | assert delta["domain"] == "code" |
| 846 | |
| 847 | |
| 848 | # --------------------------------------------------------------------------- |
| 849 | # CodePlugin — diff (semantic, with repo_root) |
| 850 | # --------------------------------------------------------------------------- |
| 851 | |
| 852 | |
| 853 | class TestCodePluginDiffSemantic: |
| 854 | plugin = CodePlugin() |
| 855 | |
| 856 | def _setup_repo( |
| 857 | self, tmp_path: pathlib.Path |
| 858 | ) -> tuple[pathlib.Path, pathlib.Path]: |
| 859 | repo_root = tmp_path / "repo" |
| 860 | repo_root.mkdir() |
| 861 | workdir = repo_root / "muse-work" |
| 862 | workdir.mkdir() |
| 863 | return repo_root, workdir |
| 864 | |
| 865 | def test_add_function_produces_patch_op(self, tmp_path: pathlib.Path) -> None: |
| 866 | repo_root, _ = self._setup_repo(tmp_path) |
| 867 | base_src = _src("x = 1\n") |
| 868 | target_src = _src("x = 1\n\ndef greet(name: str) -> str:\n return f'Hello {name}'\n") |
| 869 | |
| 870 | base_oid = _store_blob(repo_root, base_src) |
| 871 | target_oid = _store_blob(repo_root, target_src) |
| 872 | |
| 873 | base = _make_manifest({"hello.py": base_oid}) |
| 874 | target = _make_manifest({"hello.py": target_oid}) |
| 875 | delta = self.plugin.diff(base, target, repo_root=repo_root) |
| 876 | |
| 877 | patch_ops = [o for o in delta["ops"] if o["op"] == "patch"] |
| 878 | assert len(patch_ops) == 1 |
| 879 | assert patch_ops[0]["address"] == "hello.py" |
| 880 | child_ops = patch_ops[0]["child_ops"] |
| 881 | assert any(c["op"] == "insert" and "greet" in c.get("content_summary", "") for c in child_ops) |
| 882 | |
| 883 | def test_remove_function_produces_patch_op(self, tmp_path: pathlib.Path) -> None: |
| 884 | repo_root, _ = self._setup_repo(tmp_path) |
| 885 | base_src = _src("def old_fn() -> None:\n pass\n") |
| 886 | target_src = _src("# removed\n") |
| 887 | |
| 888 | base_oid = _store_blob(repo_root, base_src) |
| 889 | target_oid = _store_blob(repo_root, target_src) |
| 890 | |
| 891 | base = _make_manifest({"mod.py": base_oid}) |
| 892 | target = _make_manifest({"mod.py": target_oid}) |
| 893 | delta = self.plugin.diff(base, target, repo_root=repo_root) |
| 894 | |
| 895 | patch_ops = [o for o in delta["ops"] if o["op"] == "patch"] |
| 896 | assert len(patch_ops) == 1 |
| 897 | child_ops = patch_ops[0]["child_ops"] |
| 898 | assert any(c["op"] == "delete" and "old_fn" in c.get("content_summary", "") for c in child_ops) |
| 899 | |
| 900 | def test_rename_function_detected(self, tmp_path: pathlib.Path) -> None: |
| 901 | repo_root, _ = self._setup_repo(tmp_path) |
| 902 | base_src = _src("def compute(x: int) -> int:\n return x * 2\n") |
| 903 | target_src = _src("def calculate(x: int) -> int:\n return x * 2\n") |
| 904 | |
| 905 | base_oid = _store_blob(repo_root, base_src) |
| 906 | target_oid = _store_blob(repo_root, target_src) |
| 907 | |
| 908 | base = _make_manifest({"ops.py": base_oid}) |
| 909 | target = _make_manifest({"ops.py": target_oid}) |
| 910 | delta = self.plugin.diff(base, target, repo_root=repo_root) |
| 911 | |
| 912 | patch_ops = [o for o in delta["ops"] if o["op"] == "patch"] |
| 913 | assert len(patch_ops) == 1 |
| 914 | child_ops = patch_ops[0]["child_ops"] |
| 915 | rename_ops = [ |
| 916 | c for c in child_ops |
| 917 | if c["op"] == "replace" and "renamed to" in c.get("new_summary", "") |
| 918 | ] |
| 919 | assert len(rename_ops) == 1 |
| 920 | assert "calculate" in rename_ops[0]["new_summary"] |
| 921 | |
| 922 | def test_implementation_change_detected(self, tmp_path: pathlib.Path) -> None: |
| 923 | repo_root, _ = self._setup_repo(tmp_path) |
| 924 | base_src = _src("def double(x: int) -> int:\n return x * 2\n") |
| 925 | target_src = _src("def double(x: int) -> int:\n return x + x\n") |
| 926 | |
| 927 | base_oid = _store_blob(repo_root, base_src) |
| 928 | target_oid = _store_blob(repo_root, target_src) |
| 929 | |
| 930 | base = _make_manifest({"math.py": base_oid}) |
| 931 | target = _make_manifest({"math.py": target_oid}) |
| 932 | delta = self.plugin.diff(base, target, repo_root=repo_root) |
| 933 | |
| 934 | patch_ops = [o for o in delta["ops"] if o["op"] == "patch"] |
| 935 | child_ops = patch_ops[0]["child_ops"] |
| 936 | impl_ops = [c for c in child_ops if "implementation changed" in c.get("new_summary", "")] |
| 937 | assert len(impl_ops) == 1 |
| 938 | |
| 939 | def test_reformat_only_produces_replace_with_reformatted( |
| 940 | self, tmp_path: pathlib.Path |
| 941 | ) -> None: |
| 942 | repo_root, _ = self._setup_repo(tmp_path) |
| 943 | base_src = _src("def add(a,b):\n return a+b\n") |
| 944 | # Same semantics, different formatting — ast.unparse normalizes both. |
| 945 | target_src = _src("def add(a, b):\n return a + b\n") |
| 946 | |
| 947 | base_oid = _store_blob(repo_root, base_src) |
| 948 | target_oid = _store_blob(repo_root, target_src) |
| 949 | |
| 950 | base = _make_manifest({"f.py": base_oid}) |
| 951 | target = _make_manifest({"f.py": target_oid}) |
| 952 | delta = self.plugin.diff(base, target, repo_root=repo_root) |
| 953 | |
| 954 | # The diff should produce a reformatted ReplaceOp rather than a PatchOp. |
| 955 | replace_ops = [o for o in delta["ops"] if o["op"] == "replace"] |
| 956 | patch_ops = [o for o in delta["ops"] if o["op"] == "patch"] |
| 957 | # Reformatting: either zero ops (if raw hashes are identical) or a |
| 958 | # reformatted replace (if raw hashes differ but symbols unchanged). |
| 959 | if delta["ops"]: |
| 960 | assert replace_ops or patch_ops # something was emitted |
| 961 | if replace_ops: |
| 962 | assert any("reformatted" in o.get("new_summary", "") for o in replace_ops) |
| 963 | |
| 964 | def test_missing_object_falls_back_to_file_level( |
| 965 | self, tmp_path: pathlib.Path |
| 966 | ) -> None: |
| 967 | repo_root, _ = self._setup_repo(tmp_path) |
| 968 | # Objects NOT written to store — should fall back gracefully. |
| 969 | base = _make_manifest({"f.py": "deadbeef" * 8}) |
| 970 | target = _make_manifest({"f.py": "cafebabe" * 8}) |
| 971 | delta = self.plugin.diff(base, target, repo_root=repo_root) |
| 972 | assert len(delta["ops"]) == 1 |
| 973 | assert delta["ops"][0]["op"] == "replace" |
| 974 | |
| 975 | |
| 976 | # --------------------------------------------------------------------------- |
| 977 | # CodePlugin — merge |
| 978 | # --------------------------------------------------------------------------- |
| 979 | |
| 980 | |
| 981 | class TestCodePluginMerge: |
| 982 | plugin = CodePlugin() |
| 983 | |
| 984 | def test_only_one_side_changed(self) -> None: |
| 985 | base = _make_manifest({"f.py": "v1"}) |
| 986 | left = _make_manifest({"f.py": "v1"}) |
| 987 | right = _make_manifest({"f.py": "v2"}) |
| 988 | result = self.plugin.merge(base, left, right) |
| 989 | assert result.is_clean |
| 990 | assert result.merged["files"]["f.py"] == "v2" |
| 991 | |
| 992 | def test_both_sides_same_change(self) -> None: |
| 993 | base = _make_manifest({"f.py": "v1"}) |
| 994 | left = _make_manifest({"f.py": "v2"}) |
| 995 | right = _make_manifest({"f.py": "v2"}) |
| 996 | result = self.plugin.merge(base, left, right) |
| 997 | assert result.is_clean |
| 998 | assert result.merged["files"]["f.py"] == "v2" |
| 999 | |
| 1000 | def test_conflict_when_both_sides_differ(self) -> None: |
| 1001 | base = _make_manifest({"f.py": "v1"}) |
| 1002 | left = _make_manifest({"f.py": "v2"}) |
| 1003 | right = _make_manifest({"f.py": "v3"}) |
| 1004 | result = self.plugin.merge(base, left, right) |
| 1005 | assert not result.is_clean |
| 1006 | assert "f.py" in result.conflicts |
| 1007 | |
| 1008 | def test_disjoint_additions_auto_merge(self) -> None: |
| 1009 | base = _make_manifest({}) |
| 1010 | left = _make_manifest({"a.py": "hash_a"}) |
| 1011 | right = _make_manifest({"b.py": "hash_b"}) |
| 1012 | result = self.plugin.merge(base, left, right) |
| 1013 | assert result.is_clean |
| 1014 | assert "a.py" in result.merged["files"] |
| 1015 | assert "b.py" in result.merged["files"] |
| 1016 | |
| 1017 | def test_deletion_on_one_side(self) -> None: |
| 1018 | base = _make_manifest({"f.py": "v1"}) |
| 1019 | left = _make_manifest({}) |
| 1020 | right = _make_manifest({"f.py": "v1"}) |
| 1021 | result = self.plugin.merge(base, left, right) |
| 1022 | assert result.is_clean |
| 1023 | assert "f.py" not in result.merged["files"] |
| 1024 | |
| 1025 | |
| 1026 | # --------------------------------------------------------------------------- |
| 1027 | # CodePlugin — merge_ops (symbol-level OT) |
| 1028 | # --------------------------------------------------------------------------- |
| 1029 | |
| 1030 | |
| 1031 | class TestCodePluginMergeOps: |
| 1032 | plugin = CodePlugin() |
| 1033 | |
| 1034 | def _py_snap(self, file_path: str, src: bytes, repo_root: pathlib.Path) -> SnapshotManifest: |
| 1035 | oid = _store_blob(repo_root, src) |
| 1036 | return _make_manifest({file_path: oid}) |
| 1037 | |
| 1038 | def test_different_symbols_auto_merge(self, tmp_path: pathlib.Path) -> None: |
| 1039 | """Two agents modify different functions → no conflict.""" |
| 1040 | repo_root = tmp_path / "repo" |
| 1041 | repo_root.mkdir() |
| 1042 | |
| 1043 | base_src = _src("""\ |
| 1044 | def foo(x: int) -> int: |
| 1045 | return x |
| 1046 | |
| 1047 | def bar(y: int) -> int: |
| 1048 | return y |
| 1049 | """) |
| 1050 | # Ours: modify foo. |
| 1051 | ours_src = _src("""\ |
| 1052 | def foo(x: int) -> int: |
| 1053 | return x * 2 |
| 1054 | |
| 1055 | def bar(y: int) -> int: |
| 1056 | return y |
| 1057 | """) |
| 1058 | # Theirs: modify bar. |
| 1059 | theirs_src = _src("""\ |
| 1060 | def foo(x: int) -> int: |
| 1061 | return x |
| 1062 | |
| 1063 | def bar(y: int) -> int: |
| 1064 | return y + 1 |
| 1065 | """) |
| 1066 | |
| 1067 | base_snap = self._py_snap("m.py", base_src, repo_root) |
| 1068 | ours_snap = self._py_snap("m.py", ours_src, repo_root) |
| 1069 | theirs_snap = self._py_snap("m.py", theirs_src, repo_root) |
| 1070 | |
| 1071 | ours_delta = self.plugin.diff(base_snap, ours_snap, repo_root=repo_root) |
| 1072 | theirs_delta = self.plugin.diff(base_snap, theirs_snap, repo_root=repo_root) |
| 1073 | |
| 1074 | result = self.plugin.merge_ops( |
| 1075 | base_snap, |
| 1076 | ours_snap, |
| 1077 | theirs_snap, |
| 1078 | ours_delta["ops"], |
| 1079 | theirs_delta["ops"], |
| 1080 | repo_root=repo_root, |
| 1081 | ) |
| 1082 | # Different symbol addresses → ops commute → no conflict. |
| 1083 | assert result.is_clean, f"Expected no conflicts, got: {result.conflicts}" |
| 1084 | |
| 1085 | def test_same_symbol_conflict(self, tmp_path: pathlib.Path) -> None: |
| 1086 | """Both agents modify the same function → conflict at symbol address.""" |
| 1087 | repo_root = tmp_path / "repo" |
| 1088 | repo_root.mkdir() |
| 1089 | |
| 1090 | base_src = _src("def calc(x: int) -> int:\n return x\n") |
| 1091 | ours_src = _src("def calc(x: int) -> int:\n return x * 2\n") |
| 1092 | theirs_src = _src("def calc(x: int) -> int:\n return x + 100\n") |
| 1093 | |
| 1094 | base_snap = self._py_snap("calc.py", base_src, repo_root) |
| 1095 | ours_snap = self._py_snap("calc.py", ours_src, repo_root) |
| 1096 | theirs_snap = self._py_snap("calc.py", theirs_src, repo_root) |
| 1097 | |
| 1098 | ours_delta = self.plugin.diff(base_snap, ours_snap, repo_root=repo_root) |
| 1099 | theirs_delta = self.plugin.diff(base_snap, theirs_snap, repo_root=repo_root) |
| 1100 | |
| 1101 | result = self.plugin.merge_ops( |
| 1102 | base_snap, |
| 1103 | ours_snap, |
| 1104 | theirs_snap, |
| 1105 | ours_delta["ops"], |
| 1106 | theirs_delta["ops"], |
| 1107 | repo_root=repo_root, |
| 1108 | ) |
| 1109 | assert not result.is_clean |
| 1110 | # Conflict should be at file or symbol level. |
| 1111 | assert len(result.conflicts) > 0 |
| 1112 | |
| 1113 | def test_disjoint_files_auto_merge(self, tmp_path: pathlib.Path) -> None: |
| 1114 | """Agents modify completely different files → auto-merge.""" |
| 1115 | repo_root = tmp_path / "repo" |
| 1116 | repo_root.mkdir() |
| 1117 | |
| 1118 | base = _make_manifest({"a.py": "v1", "b.py": "v1"}) |
| 1119 | ours = _make_manifest({"a.py": "v2", "b.py": "v1"}) |
| 1120 | theirs = _make_manifest({"a.py": "v1", "b.py": "v2"}) |
| 1121 | |
| 1122 | ours_delta = self.plugin.diff(base, ours) |
| 1123 | theirs_delta = self.plugin.diff(base, theirs) |
| 1124 | |
| 1125 | result = self.plugin.merge_ops( |
| 1126 | base, ours, theirs, |
| 1127 | ours_delta["ops"], |
| 1128 | theirs_delta["ops"], |
| 1129 | ) |
| 1130 | assert result.is_clean |
| 1131 | |
| 1132 | |
| 1133 | # --------------------------------------------------------------------------- |
| 1134 | # CodePlugin — drift |
| 1135 | # --------------------------------------------------------------------------- |
| 1136 | |
| 1137 | |
| 1138 | class TestCodePluginDrift: |
| 1139 | plugin = CodePlugin() |
| 1140 | |
| 1141 | def test_no_drift(self, tmp_path: pathlib.Path) -> None: |
| 1142 | workdir = tmp_path / "muse-work" |
| 1143 | workdir.mkdir() |
| 1144 | (workdir / "app.py").write_text("x = 1\n") |
| 1145 | snap = self.plugin.snapshot(workdir) |
| 1146 | report = self.plugin.drift(snap, workdir) |
| 1147 | assert not report.has_drift |
| 1148 | |
| 1149 | def test_has_drift_after_edit(self, tmp_path: pathlib.Path) -> None: |
| 1150 | workdir = tmp_path / "muse-work" |
| 1151 | workdir.mkdir() |
| 1152 | f = workdir / "app.py" |
| 1153 | f.write_text("x = 1\n") |
| 1154 | snap = self.plugin.snapshot(workdir) |
| 1155 | f.write_text("x = 2\n") |
| 1156 | report = self.plugin.drift(snap, workdir) |
| 1157 | assert report.has_drift |
| 1158 | |
| 1159 | def test_has_drift_after_add(self, tmp_path: pathlib.Path) -> None: |
| 1160 | workdir = tmp_path / "muse-work" |
| 1161 | workdir.mkdir() |
| 1162 | (workdir / "a.py").write_text("a = 1\n") |
| 1163 | snap = self.plugin.snapshot(workdir) |
| 1164 | (workdir / "b.py").write_text("b = 2\n") |
| 1165 | report = self.plugin.drift(snap, workdir) |
| 1166 | assert report.has_drift |
| 1167 | |
| 1168 | def test_has_drift_after_delete(self, tmp_path: pathlib.Path) -> None: |
| 1169 | workdir = tmp_path / "muse-work" |
| 1170 | workdir.mkdir() |
| 1171 | f = workdir / "gone.py" |
| 1172 | f.write_text("x = 1\n") |
| 1173 | snap = self.plugin.snapshot(workdir) |
| 1174 | f.unlink() |
| 1175 | report = self.plugin.drift(snap, workdir) |
| 1176 | assert report.has_drift |
| 1177 | |
| 1178 | |
| 1179 | # --------------------------------------------------------------------------- |
| 1180 | # CodePlugin — apply (passthrough) |
| 1181 | # --------------------------------------------------------------------------- |
| 1182 | |
| 1183 | |
| 1184 | def test_apply_returns_live_state_unchanged(tmp_path: pathlib.Path) -> None: |
| 1185 | plugin = CodePlugin() |
| 1186 | workdir = tmp_path / "muse-work" |
| 1187 | workdir.mkdir() |
| 1188 | delta = plugin.diff(_make_manifest({}), _make_manifest({})) |
| 1189 | result = plugin.apply(delta, workdir) |
| 1190 | assert result is workdir |
| 1191 | |
| 1192 | |
| 1193 | # --------------------------------------------------------------------------- |
| 1194 | # CodePlugin — schema |
| 1195 | # --------------------------------------------------------------------------- |
| 1196 | |
| 1197 | |
| 1198 | class TestCodePluginSchema: |
| 1199 | plugin = CodePlugin() |
| 1200 | |
| 1201 | def test_schema_domain(self) -> None: |
| 1202 | assert self.plugin.schema()["domain"] == "code" |
| 1203 | |
| 1204 | def test_schema_merge_mode(self) -> None: |
| 1205 | assert self.plugin.schema()["merge_mode"] == "three_way" |
| 1206 | |
| 1207 | def test_schema_version(self) -> None: |
| 1208 | assert self.plugin.schema()["schema_version"] == 1 |
| 1209 | |
| 1210 | def test_schema_dimensions(self) -> None: |
| 1211 | dims = self.plugin.schema()["dimensions"] |
| 1212 | names = {d["name"] for d in dims} |
| 1213 | assert "structure" in names |
| 1214 | assert "symbols" in names |
| 1215 | assert "imports" in names |
| 1216 | |
| 1217 | def test_schema_top_level_is_tree(self) -> None: |
| 1218 | top = self.plugin.schema()["top_level"] |
| 1219 | assert top["kind"] == "tree" |
| 1220 | |
| 1221 | def test_schema_description_non_empty(self) -> None: |
| 1222 | assert len(self.plugin.schema()["description"]) > 0 |
| 1223 | |
| 1224 | |
| 1225 | # --------------------------------------------------------------------------- |
| 1226 | # delta_summary |
| 1227 | # --------------------------------------------------------------------------- |
| 1228 | |
| 1229 | |
| 1230 | class TestDeltaSummary: |
| 1231 | def test_empty_ops(self) -> None: |
| 1232 | assert delta_summary([]) == "no changes" |
| 1233 | |
| 1234 | def test_file_added(self) -> None: |
| 1235 | from muse.domain import DomainOp |
| 1236 | ops: list[DomainOp] = [InsertOp( |
| 1237 | op="insert", address="f.py", position=None, |
| 1238 | content_id="abc", content_summary="added f.py", |
| 1239 | )] |
| 1240 | summary = delta_summary(ops) |
| 1241 | assert "added" in summary |
| 1242 | assert "file" in summary |
| 1243 | |
| 1244 | def test_symbols_counted_from_patch(self) -> None: |
| 1245 | from muse.domain import DomainOp, PatchOp |
| 1246 | child: list[DomainOp] = [ |
| 1247 | InsertOp(op="insert", address="f.py::foo", position=None, content_id="a", content_summary="added function foo"), |
| 1248 | InsertOp(op="insert", address="f.py::bar", position=None, content_id="b", content_summary="added function bar"), |
| 1249 | ] |
| 1250 | ops: list[DomainOp] = [PatchOp(op="patch", address="f.py", child_ops=child, child_domain="code_symbols", child_summary="2 added")] |
| 1251 | summary = delta_summary(ops) |
| 1252 | assert "symbol" in summary |