fix: reuse token truncation helper

Maintainer edit: make token truncation include the suffix within the budget and route the consolidator through the shared helper so recent-history and archive truncation keep the same semantics.

nanobot/agent/memory.py

@@ -13,7 +13,6 @@ from datetime import datetime
 from pathlib import Path
 from typing import TYPE_CHECKING, Any, Callable, Iterator
 
-import tiktoken
 from loguru import logger
 
 from nanobot.session.manager import Session
@@ -25,6 +24,7 @@ from nanobot.utils.helpers import (
     find_legal_message_start,
     strip_think,
     truncate_text,
+    truncate_text_to_tokens,
 )
 from nanobot.utils.prompt_templates import render_template
 
@@ -806,14 +806,7 @@ class Consolidator:
         budget = self._input_token_budget
         if budget <= 0:
             return truncate_text(text, _RAW_ARCHIVE_MAX_CHARS)
-        try:
-            enc = tiktoken.get_encoding("cl100k_base")
-            tokens = enc.encode(text)
-            if len(tokens) <= budget:
-                return text
-            return enc.decode(tokens[:budget]) + "\n... (truncated)"
-        except Exception:
-            return truncate_text(text, budget * 4)
+        return truncate_text_to_tokens(text, budget)
 
     async def archive(
         self,

nanobot/utils/helpers.py

@@ -218,6 +218,7 @@ _TOOL_RESULT_PREVIEW_CHARS = 1200
 _TOOL_RESULTS_DIR = ".nanobot/tool-results"
 _TOOL_RESULT_RETENTION_SECS = 7 * 24 * 60 * 60
 _TOOL_RESULT_MAX_BUCKETS = 32
+_TRUNCATED_SUFFIX = "\n... (truncated)"
 
 
 def safe_filename(name: str) -> str:
@@ -234,7 +235,7 @@ def truncate_text(text: str, max_chars: int) -> str:
     """Truncate text with a stable suffix."""
     if max_chars <= 0 or len(text) <= max_chars:
         return text
-    return text[:max_chars] + "\n... (truncated)"
+    return text[:max_chars] + _TRUNCATED_SUFFIX
 
 
 def truncate_text_to_tokens(text: str, max_tokens: int) -> str:
@@ -252,9 +253,21 @@ def truncate_text_to_tokens(text: str, max_tokens: int) -> str:
         tokens = enc.encode(text)
         if len(tokens) <= max_tokens:
             return text
-        return enc.decode(tokens[:max_tokens]) + "\n... (truncated)"
+        suffix_tokens = enc.encode(_TRUNCATED_SUFFIX)
+        body_budget = max_tokens - len(suffix_tokens)
+        if body_budget <= 0:
+            return enc.decode(tokens[:max_tokens])
+        result = enc.decode(tokens[:body_budget]) + _TRUNCATED_SUFFIX
+        while len(enc.encode(result)) > max_tokens and body_budget > 0:
+            body_budget -= 1
+            result = enc.decode(tokens[:body_budget]) + _TRUNCATED_SUFFIX
+        return result
     except Exception:
-        return truncate_text(text, max_tokens * 4)
+        max_chars = max_tokens * 4
+        suffix_chars = len(_TRUNCATED_SUFFIX)
+        if max_chars <= suffix_chars:
+            return text[:max_chars]
+        return truncate_text(text, max_chars - suffix_chars)
 
 
 def find_legal_message_start(messages: list[dict[str, Any]]) -> int:

tests/agent/test_consolidator.py

@@ -825,4 +825,4 @@ class TestArchiveTruncation:
         enc = tiktoken.get_encoding("cl100k_base")
         sent_content = mock_provider.chat_with_retry.call_args.kwargs["messages"][1]["content"]
         token_count = len(enc.encode(sent_content))
-        assert token_count <= 9_900 + 10  # small margin for truncation suffix
+        assert token_count <= 9_900

tests/agent/test_context_prompt_cache.py

@@ -237,8 +237,7 @@ def test_recent_history_truncated_at_max_tokens(tmp_path) -> None:
     assert len(history_section) == 2
 
     enc = tiktoken.get_encoding("cl100k_base")
-    # Small margin for the truncation suffix appended after the token slice.
-    assert len(enc.encode(history_section[1])) <= builder._MAX_HISTORY_TOKENS + 50
+    assert len(enc.encode(history_section[1])) <= builder._MAX_HISTORY_TOKENS
 
 
 def test_no_recent_history_when_dream_has_processed_all(tmp_path) -> None:

tests/utils/test_helpers.py

@@ -24,8 +24,7 @@ def test_truncate_text_to_tokens_truncates_over_budget():
     result = truncate_text_to_tokens(text, 50)
 
     assert result.endswith("\n... (truncated)")
-    body = result[: -len("\n... (truncated)")]
-    assert len(enc.encode(body)) <= 50
+    assert len(enc.encode(result)) <= 50
 
 
 def test_truncate_text_to_tokens_non_positive_budget_returns_text():