LCORE-1569,LCORE-1570: appease pylint, pyright, and ruff

Fix-ups discovered by `uv run make verify`. No behavior change.

* utils/compaction: hoist `from utils.token_estimator import ...`
  to the module top. The original local imports were a paranoid
  hedge against a hypothetical circular import that does not
  actually exist (token_estimator does not import compaction).
  Top-level imports also let pylint's import-outside-toplevel rule
  pass cleanly.

* utils/compaction: switch `timezone.utc` to the Python 3.13
  `datetime.UTC` alias per ruff UP017.

* tests: `import tiktoken` at module top instead of inside a test
  method (test_token_estimator.py).

* tests: `# pylint: disable=too-few-public-methods` at module top
  for the test files that use one-attribute stand-in classes for
  message and content-part fixtures. These are intentional
  duck-type stubs, not real abstractions, and adding methods just
  to satisfy a count would obscure the intent.

* tests: `Configuration.model_fields.get(...)` instead of
  `"compaction" in Configuration.model_fields`. The dict accessor
  pylint cannot infer; the .get() form sidesteps the
  unsupported-membership-test / unsubscriptable-object false
  positives without weakening the assertion.

* models/config: `compaction` Field's `default_factory` now
  constructs CompactionConfiguration with every field passed
  explicitly. Matches the established pattern used by
  `conversation_cache` and `quota_handlers` on the same Configuration
  class, and resolves pyright reportCallIssue: the bare class form
  trips pyright's "constructor requires positional args" check
  because pyright treats the kwargs-with-defaults init as
  zero-positional but more-than-zero-required-keyword.

Author: max-svistunov

src/models/config.py

@@ -2007,7 +2007,13 @@ class Configuration(ConfigurationBase):
     )
 
     compaction: CompactionConfiguration = Field(
-        default_factory=CompactionConfiguration,
+        default_factory=lambda: CompactionConfiguration(
+            enabled=False,
+            threshold_ratio=0.7,
+            token_floor=4096,
+            buffer_turns=4,
+            buffer_max_ratio=0.3,
+        ),
         title="Conversation compaction configuration",
         description="Controls when conversation history is summarized "
         "to keep the model's input below the context window limit. "

src/utils/compaction.py

@@ -27,13 +27,14 @@
 to disentangle a tangle of side effects.
 """
 
-from datetime import datetime, timezone
+from datetime import UTC, datetime
 from typing import Any
 
 from llama_stack_client import AsyncLlamaStackClient
 
 from log import get_logger
 from models.compaction import ConversationSummary
+from utils.token_estimator import estimate_conversation_tokens, estimate_tokens
 
 logger = get_logger(__name__)
 
@@ -190,10 +191,6 @@ def partition_conversation(
         compaction needed); ``recent_items`` may be empty (everything
         had to be summarized).
     """
-    # Import locally to avoid a top-level circular import once the
-    # request-flow integration in LCORE-1572 imports both modules.
-    from utils.token_estimator import estimate_conversation_tokens
-
     msg_indices = _message_indices(items)
     if not msg_indices:
         return [], items
@@ -325,8 +322,6 @@ async def summarize_chunk(
             cannot be persisted (PositiveInt) and is also useless as
             context, so propagating an error is the honest behavior.
     """
-    from utils.token_estimator import estimate_tokens
-
     prompt = _build_summarization_prompt_body(old_items)
     logger.info(
         "Summarizing %d conversation items (%d messages) for model %s.",
@@ -351,7 +346,7 @@ async def summarize_chunk(
         summary_text=summary_text,
         summarized_through_turn=summarized_through_turn,
         token_count=token_count,
-        created_at=datetime.now(timezone.utc).isoformat(),
+        created_at=datetime.now(UTC).isoformat(),
         model_used=model,
     )
 
@@ -423,8 +418,6 @@ async def recursively_resummarize(
             fold is needed) or when the LLM call yields no
             extractable text.
     """
-    from utils.token_estimator import estimate_tokens
-
     if len(summaries) < 2:
         raise ValueError(
             "recursively_resummarize requires at least 2 summary chunks "
@@ -459,6 +452,6 @@ async def recursively_resummarize(
         summary_text=folded_text,
         summarized_through_turn=summaries[-1].summarized_through_turn,
         token_count=estimate_tokens(folded_text, encoding_name=encoding_name),
-        created_at=datetime.now(timezone.utc).isoformat(),
+        created_at=datetime.now(UTC).isoformat(),
         model_used=model,
     )

tests/unit/models/config/test_compaction_configuration.py

@@ -97,9 +97,8 @@ def test_root_configuration_has_compaction_field() -> None:
     """The root Configuration declares a `compaction` field typed as
     CompactionConfiguration with a default-factory that produces a
     fresh CompactionConfiguration with the spec defaults."""
-    fields = Configuration.model_fields
-    assert "compaction" in fields, "Configuration must declare a compaction field"
-    field_info = fields["compaction"]
+    field_info = Configuration.model_fields.get("compaction")
+    assert field_info is not None, "Configuration must declare a compaction field"
     assert field_info.annotation is CompactionConfiguration
 
     # default_factory must produce a CompactionConfiguration with disabled

tests/unit/utils/test_compaction.py

@@ -1,5 +1,7 @@
 """Unit tests for utils/compaction — partitioning, prompt, summarization."""
 
+# pylint: disable=too-few-public-methods
+
 from typing import Any
 
 import pytest

tests/unit/utils/test_token_estimator.py

@@ -1,8 +1,11 @@
 """Unit tests for utils/token_estimator."""
 
+# pylint: disable=too-few-public-methods
+
 from typing import Any
 
 import pytest
+import tiktoken
 
 from models.config import InferenceConfiguration
 from utils.token_estimator import (
@@ -93,8 +96,6 @@ def test_within_5pct_of_explicit_tiktoken_call(self) -> None:
         is at most 5% — establishing that the estimator does not
         introduce its own off-by-some error.
         """
-        import tiktoken
-
         encoding = tiktoken.get_encoding(DEFAULT_ENCODING_NAME)
         text = (
             "The quick brown fox jumps over the lazy dog. "