massive tests refactoring

Author: nikolay-e

failing_tests.yaml

@@ -24,7 +24,7 @@ def _get_encoder(encoding: str) -> Any | None:
         import tiktoken
 
         return tiktoken.get_encoding(encoding)
-    except (ImportError, KeyError, ValueError):
+    except Exception:
         return None
 
 

src/treemapper/tokens.py

@@ -0,0 +1,119 @@
+# Diff Test Cases Enrichment
+
+## Goal
+
+Enrich all diff test cases with:
+
+1. **Realistic code content** - larger, more representative code snippets
+2. **Garbage files** - files with unique markers to test exclusion
+3. **Proper diffctx format** - structured assertions
+
+## File Structure
+
+Each test case consists of 3 files:
+
+- `{name}_before.yaml` - codebase state before the change
+- `{name}_after.yaml` - codebase state after the change
+- `{name}_diffctx.yaml` - assertions for diff context selection
+
+## Garbage Marker Pattern
+
+```
+GARBAGE_{CATEGORY}_{NUM}_{DESCRIPTION}_{LETTER}
+```
+
+Examples:
+
+- `GARBAGE_CICD_003_ROLLBACK_MARKER_A`
+- `GARBAGE_ALGO_005_LOGGING_MARKER_B`
+
+## diffctx Format
+
+```yaml
+must_include_files:
+  - path/to/important/file.py
+must_include_content:
+  - |
+    actual code block that must appear
+must_not_include:
+  - GARBAGE_MARKER_A
+  - GarbageClassName
+  - garbage_file.py
+```
+
+---
+
+## Progress
+
+### ✅ Completed
+
+| Directory | Files | Test Cases | Status |
+|-----------|-------|------------|--------|
+| algorithm | 63 | 21 | ✅ Done |
+| cicd | 45 | 15 | ✅ Done |
+
+### 🔄 Remaining
+
+| Directory | Files | Est. Cases | Priority |
+|-----------|-------|------------|----------|
+| cicd_and_docs | 234 | ~78 | High |
+| comprehensive | 81 | ~27 | High |
+| config | 15 | ~5 | Medium |
+| cpp | 108 | ~36 | Medium |
+| csharp | 78 | ~26 | Medium |
+| dependencies | 123 | ~41 | High |
+| docker | 135 | ~45 | High |
+| fragments | 72 | ~24 | Medium |
+| frontend | 90 | ~30 | High |
+| go | 105 | ~35 | Medium |
+| graph | 51 | ~17 | Medium |
+| helm | 135 | ~45 | High |
+| infrastructure_validation | 15 | ~5 | Low |
+| internals | 195 | ~65 | Medium |
+| java | 78 | ~26 | Medium |
+| javascript | 255 | ~85 | High |
+| javascript_extended | 147 | ~49 | Medium |
+| json | 45 | ~15 | Low |
+| jvm_and_compiled | 216 | ~72 | Medium |
+| kubernetes | 120 | ~40 | High |
+| merging | 36 | ~12 | Low |
+| operations | 18 | ~6 | Low |
+| output | 36 | ~12 | Low |
+| patterns | 135 | ~45 | Medium |
+| php | 60 | ~20 | Low |
+| ppr | 105 | ~35 | Medium |
+| python | 186 | ~62 | High |
+| relations | 21 | ~7 | Low |
+| ruby | 60 | ~20 | Low |
+| rust | 120 | ~40 | Medium |
+| scala | 78 | ~26 | Low |
+| scripting | 90 | ~30 | Medium |
+| selection | 24 | ~8 | Medium |
+| shell | 75 | ~25 | Medium |
+| swift | 60 | ~20 | Low |
+| terraform | 240 | ~80 | High |
+| typescript | 27 | ~9 | Medium |
+| yaml | 60 | ~20 | Medium |
+
+**Total remaining:** ~38 directories, ~1100+ test cases
+
+---
+
+## Process
+
+For each test case:
+
+1. **Read** all 3 files (before, after, diffctx)
+2. **Check** if already enriched (look for garbage markers)
+3. **Enrich** if minimal:
+   - Add realistic code to before/after
+   - Add 2 garbage files with unique markers
+   - Update diffctx to proper format
+4. **Write** updated files
+
+## Notes
+
+- Process one directory at a time
+- Use consistent garbage marker naming per directory
+- Keep code realistic for the test scenario
+- Ensure diffctx assertions match actual diff content

tests/cases/diff/TODO.md

@@ -0,0 +1,78 @@
+models/user.py: |
+  from dataclasses import dataclass, field
+  from typing import Optional, List
+  from datetime import datetime
+
+  @dataclass
+  class Address:
+      street: str
+      city: str
+      country: str
+      postal_code: str
+
+      def format(self) -> str:
+          return f"{self.street}, {self.city}, {self.postal_code}"
+
+  @dataclass
+  class User:
+      id: int
+      username: str
+      email: str
+      display_name: Optional[str] = None
+      created_at: datetime = field(default_factory=datetime.now)
+      address: Optional[Address] = None
+      roles: List[str] = field(default_factory=list)
+
+      def is_admin(self) -> bool:
+          return "admin" in self.roles
+
+      def has_address(self) -> bool:
+          return self.address is not None
+
+      def get_display_name(self) -> str:
+          return self.display_name or self.username
+
+services/user_service.py: |
+  from models.user import User, Address
+
+  class UserService:
+      def __init__(self, repository):
+          self.repository = repository
+
+      def create_user(self, username: str, email: str) -> User:
+          user = User(
+              id=self.repository.next_id(),
+              username=username,
+              email=email
+          )
+          self.repository.save(user)
+          return user
+
+      def add_role(self, user: User, role: str) -> None:
+          if role not in user.roles:
+              user.roles.append(role)
+              self.repository.save(user)
+
+      def set_address(self, user: User, address: Address) -> None:
+          user.address = address
+          self.repository.save(user)
+
+garbage_analytics.py: |
+  GARBAGE_ALGO_001_ANALYTICS_MARKER_A = "analytics"
+  GARBAGE_ALGO_001_TRACKER_MARKER_B = True
+
+  class AnalyticsTracker:
+      GARBAGE_ALGO_001_CLASS_MARKER_C = "tracker"
+
+      def track(self, event: str) -> None:
+          pass
+
+garbage_notifications.py: |
+  GARBAGE_ALGO_001_NOTIFY_MARKER_D = "notify"
+  GARBAGE_ALGO_001_PUSH_MARKER_E = 42
+
+  class PushService:
+      GARBAGE_ALGO_001_PUSH_CLASS_F = True
+
+      def send(self, token: str, message: str) -> bool:
+          return True

tests/cases/diff/algorithm/algorithm_001_fragment_class_decorator_included_after.yaml

@@ -0,0 +1,74 @@
+models/user.py: |
+  from dataclasses import dataclass, field
+  from typing import Optional, List
+  from datetime import datetime
+
+  @dataclass
+  class Address:
+      street: str
+      city: str
+      country: str
+      postal_code: str
+
+      def format(self) -> str:
+          return f"{self.street}, {self.city}, {self.postal_code}"
+
+  @dataclass
+  class User:
+      id: int
+      username: str
+      email: str
+      created_at: datetime = field(default_factory=datetime.now)
+      address: Optional[Address] = None
+      roles: List[str] = field(default_factory=list)
+
+      def is_admin(self) -> bool:
+          return "admin" in self.roles
+
+      def has_address(self) -> bool:
+          return self.address is not None
+
+services/user_service.py: |
+  from models.user import User, Address
+
+  class UserService:
+      def __init__(self, repository):
+          self.repository = repository
+
+      def create_user(self, username: str, email: str) -> User:
+          user = User(
+              id=self.repository.next_id(),
+              username=username,
+              email=email
+          )
+          self.repository.save(user)
+          return user
+
+      def add_role(self, user: User, role: str) -> None:
+          if role not in user.roles:
+              user.roles.append(role)
+              self.repository.save(user)
+
+      def set_address(self, user: User, address: Address) -> None:
+          user.address = address
+          self.repository.save(user)
+
+garbage_analytics.py: |
+  GARBAGE_ALGO_001_ANALYTICS_MARKER_A = "analytics"
+  GARBAGE_ALGO_001_TRACKER_MARKER_B = True
+
+  class AnalyticsTracker:
+      GARBAGE_ALGO_001_CLASS_MARKER_C = "tracker"
+
+      def track(self, event: str) -> None:
+          pass
+
+garbage_notifications.py: |
+  GARBAGE_ALGO_001_NOTIFY_MARKER_D = "notify"
+  GARBAGE_ALGO_001_PUSH_MARKER_E = 42
+
+  class PushService:
+      GARBAGE_ALGO_001_PUSH_CLASS_F = True
+
+      def send(self, token: str, message: str) -> bool:
+          return True

tests/cases/diff/algorithm/algorithm_001_fragment_class_decorator_included_before.yaml

@@ -0,0 +1,33 @@
+must_include_files:
+  - models/user.py
+must_include_content:
+  - |
+    @dataclass
+    class User:
+        id: int
+        username: str
+        email: str
+        display_name: Optional[str] = None
+        created_at: datetime = field(default_factory=datetime.now)
+        address: Optional[Address] = None
+        roles: List[str] = field(default_factory=list)
+
+        def is_admin(self) -> bool:
+            return "admin" in self.roles
+
+        def has_address(self) -> bool:
+            return self.address is not None
+
+        def get_display_name(self) -> str:
+            return self.display_name or self.username
+must_not_include:
+  - GARBAGE_ALGO_001_ANALYTICS_MARKER_A
+  - GARBAGE_ALGO_001_TRACKER_MARKER_B
+  - GARBAGE_ALGO_001_CLASS_MARKER_C
+  - GARBAGE_ALGO_001_NOTIFY_MARKER_D
+  - GARBAGE_ALGO_001_PUSH_MARKER_E
+  - GARBAGE_ALGO_001_PUSH_CLASS_F
+  - AnalyticsTracker
+  - PushService
+  - garbage_analytics.py
+  - garbage_notifications.py

tests/cases/diff/algorithm/algorithm_001_fragment_class_decorator_included_diffctx.yaml

@@ -0,0 +1,72 @@
+decorators.py: |
+  from functools import wraps
+  from typing import Callable, Any
+
+  def log_calls(func: Callable) -> Callable:
+      @wraps(func)
+      def wrapper(*args: Any, **kwargs: Any) -> Any:
+          print(f"Calling {func.__name__}")
+          result = func(*args, **kwargs)
+          print(f"Finished {func.__name__}")
+          return result
+      return wrapper
+
+  def retry(max_attempts: int = 3) -> Callable:
+      def decorator(func: Callable) -> Callable:
+          @wraps(func)
+          def wrapper(*args: Any, **kwargs: Any) -> Any:
+              for attempt in range(max_attempts):
+                  try:
+                      return func(*args, **kwargs)
+                  except Exception as e:
+                      if attempt == max_attempts - 1:
+                          raise
+              return None
+          return wrapper
+      return decorator
+
+  def validate_args(func: Callable) -> Callable:
+      @wraps(func)
+      def wrapper(*args: Any, **kwargs: Any) -> Any:
+          if not args and not kwargs:
+              raise ValueError("No arguments provided")
+          return func(*args, **kwargs)
+      return wrapper
+
+services/api_client.py: |
+  from decorators import log_calls, retry, validate_args
+
+  class ApiClient:
+      def __init__(self, base_url: str):
+          self.base_url = base_url
+          self.timeout = 30
+
+      @log_calls
+      @retry(max_attempts=5)
+      def fetch_data(self, endpoint: str) -> dict:
+          url = f"{self.base_url}/{endpoint}"
+          return {"url": url, "status": "ok", "timeout": self.timeout}
+
+      @validate_args
+      def post_data(self, endpoint: str, data: dict) -> dict:
+          return {"endpoint": endpoint, "data": data}
+
+garbage_cache.py: |
+  GARBAGE_ALGO_002_CACHE_MARKER_A = "cache"
+  GARBAGE_ALGO_002_REDIS_MARKER_B = True
+
+  class CacheManager:
+      GARBAGE_ALGO_002_CLASS_MARKER_C = "manager"
+
+      def get(self, key: str) -> None:
+          pass
+
+garbage_metrics.py: |
+  GARBAGE_ALGO_002_METRICS_MARKER_D = "metrics"
+  GARBAGE_ALGO_002_COUNTER_MARKER_E = 0
+
+  class MetricsCollector:
+      GARBAGE_ALGO_002_COLLECTOR_F = True
+
+      def record(self, name: str, value: float) -> None:
+          pass