ouroboros-ai 0.2.0__py3-none-any.whl → 0.2.2__py3-none-any.whl

ouroboros/bigbang/ambiguity.py

@@ -36,8 +36,8 @@ DEFAULT_MODEL = "openrouter/google/gemini-2.0-flash-001"
 # Temperature for reproducible scoring
 SCORING_TEMPERATURE = 0.1
 
-# Maximum token limit to prevent cost explosion
-MAX_TOKEN_LIMIT = 8192
+# Maximum token limit (None = no limit, rely on model's context window)
+MAX_TOKEN_LIMIT: int | None = None
 
 
 class ComponentScore(BaseModel):
@@ -110,15 +110,15 @@ class AmbiguityScorer:
     from interview conversation, producing reproducible scores.
 
     Uses adaptive token allocation: starts with `initial_max_tokens` and
-    doubles on truncation up to `MAX_TOKEN_LIMIT`. Retries up to `max_retries`
-    times on both provider errors and parse failures.
+    doubles on truncation up to `MAX_TOKEN_LIMIT`. Retries until success
+    by default (unlimited), or up to `max_retries` if specified.
 
     Attributes:
         llm_adapter: The LLM adapter for completions.
         model: Model identifier to use.
         temperature: Temperature for reproducibility (default 0.1).
         initial_max_tokens: Starting token limit (default 2048).
-        max_retries: Maximum retry attempts (default 3).
+        max_retries: Maximum retry attempts, or None for unlimited (default).
 
     Example:
         scorer = AmbiguityScorer(llm_adapter=LiteLLMAdapter())
@@ -138,7 +138,7 @@ class AmbiguityScorer:
     model: str = DEFAULT_MODEL
     temperature: float = SCORING_TEMPERATURE
     initial_max_tokens: int = 2048
-    max_retries: int = 3
+    max_retries: int | None = None  # None = unlimited retries
 
     async def score(
         self, state: InterviewState
@@ -180,8 +180,15 @@ class AmbiguityScorer:
         current_max_tokens = self.initial_max_tokens
         last_error: Exception | ProviderError | None = None
         last_response: str = ""
+        attempt = 0
+
+        while True:
+            # Check retry limit if set
+            if self.max_retries is not None and attempt >= self.max_retries:
+                break
+
+            attempt += 1
 
-        for attempt in range(self.max_retries):
             config = CompletionConfig(
                 model=self.model,
                 temperature=self.temperature,
@@ -190,15 +197,15 @@ class AmbiguityScorer:
 
             result = await self.llm_adapter.complete(messages, config)
 
-            # Fix #3: Retry on provider errors (rate limits, transient failures)
+            # Retry on provider errors (rate limits, transient failures)
             if result.is_err:
                 last_error = result.error
                 log.warning(
                     "ambiguity.scoring.provider_error_retrying",
                     interview_id=state.interview_id,
                     error=str(result.error),
-                    attempt=attempt + 1,
-                    max_retries=self.max_retries,
+                    attempt=attempt,
+                    max_retries=self.max_retries or "unlimited",
                 )
                 continue
 
@@ -221,7 +228,7 @@ class AmbiguityScorer:
                     constraint_clarity=breakdown.constraint_clarity.clarity_score,
                     success_criteria_clarity=breakdown.success_criteria_clarity.clarity_score,
                     tokens_used=current_max_tokens,
-                    attempt=attempt + 1,
+                    attempt=attempt,
                 )
 
                 return Result.ok(ambiguity_score)
@@ -230,17 +237,19 @@ class AmbiguityScorer:
                 last_error = e
                 last_response = result.value.content
 
-                # Fix #2: Only increase tokens if response was truncated
+                # Only increase tokens if response was truncated
                 is_truncated = result.value.finish_reason == "length"
 
                 if is_truncated:
-                    # Fix #1: Cap token growth with MAX_TOKEN_LIMIT
-                    next_tokens = min(current_max_tokens * 2, MAX_TOKEN_LIMIT)
+                    # Double tokens on truncation, capped at MAX_TOKEN_LIMIT if set
+                    next_tokens = current_max_tokens * 2
+                    if MAX_TOKEN_LIMIT is not None:
+                        next_tokens = min(next_tokens, MAX_TOKEN_LIMIT)
                     log.warning(
                         "ambiguity.scoring.truncated_retrying",
                         interview_id=state.interview_id,
                         error=str(e),
-                        attempt=attempt + 1,
+                        attempt=attempt,
                         current_tokens=current_max_tokens,
                         next_tokens=next_tokens,
                     )
@@ -251,11 +260,11 @@ class AmbiguityScorer:
                         "ambiguity.scoring.format_error_retrying",
                         interview_id=state.interview_id,
                         error=str(e),
-                        attempt=attempt + 1,
+                        attempt=attempt,
                         finish_reason=result.value.finish_reason,
                     )
 
-        # All retries exhausted
+        # All retries exhausted (only reached if max_retries is set)
         log.warning(
             "ambiguity.scoring.failed",
             interview_id=state.interview_id,

{ouroboros_ai-0.2.0.dist-info → ouroboros_ai-0.2.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ouroboros-ai
-Version: 0.2.0
+Version: 0.2.2
 Summary: Self-Improving AI Workflow System
 Author-email: Q00 <jqyu.lee@gmail.com>
 License-File: LICENSE

{ouroboros_ai-0.2.0.dist-info → ouroboros_ai-0.2.2.dist-info}/RECORD

@@ -2,7 +2,7 @@ ouroboros/__init__.py,sha256=lmQgHmNOWxGlmwayNvp1ckCuJycL8WzX5Y-7IzrFaVM,701
 ouroboros/__main__.py,sha256=f_qnL0zPJwh9kfQqynX5adpqzj8ilj94zW5Q2loqGxE,168
 ouroboros/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ouroboros/bigbang/__init__.py,sha256=9xGqOYwMKBifb7QVwonc_wndNLMZb7ZH7xgMHaz_70A,951
-ouroboros/bigbang/ambiguity.py,sha256=hm-6LeuD_j14uzgZ2wnbBYq4Q24J7kEk4ag0DO0JtAU,18516
+ouroboros/bigbang/ambiguity.py,sha256=pmvahQP41OVL8o6N7DkNHM8zqsLYYhem_CfM2X5xuGE,18911
 ouroboros/bigbang/interview.py,sha256=zm1VrDNqE8ouGG62h8qnNkIpnUf3HHv4NjzMKDIaWcY,17147
 ouroboros/bigbang/seed_generator.py,sha256=7MY9a7Eua_zVGDWIVDlzOZJjeAwz0DRatXJg0PvMgiY,20082
 ouroboros/cli/__init__.py,sha256=CRpxsqJadZL7bCS-yrULWC51tqPKfPsxQLgt0JiwP4g,225
@@ -75,8 +75,8 @@ ouroboros/routing/tiers.py,sha256=QhBQUOo2-h5Z3dEtC0lcOzkRnqTi2W7Jl46750AVNig,73
 ouroboros/secondary/__init__.py,sha256=kYQ7C4bnBzwDlPrU8qZrOPr2ZuTBaftGktOXl5WZl5Q,1123
 ouroboros/secondary/scheduler.py,sha256=sPVVWJ1q0yewRAM-Rm1j_HMerSe4cavIvP9z4xlUuL4,13737
 ouroboros/secondary/todo_registry.py,sha256=4W3C9Uro29VrVLCPKUlpH_BYpzQSbRNW1oMnDYyEhEw,13880
-ouroboros_ai-0.2.0.dist-info/METADATA,sha256=-znRAEKqEghugiU67FXrH52Hyt4kBtigrQvwXW-3J_E,19661
-ouroboros_ai-0.2.0.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
-ouroboros_ai-0.2.0.dist-info/entry_points.txt,sha256=MoETHup6rVkR6AsyjoRzAgIuvVtYYm3Jw40itV3_VyI,53
-ouroboros_ai-0.2.0.dist-info/licenses/LICENSE,sha256=n2X-q26TqpXnoBo0t_WouhFxWw663_q5FmbYDZayoHo,1060
-ouroboros_ai-0.2.0.dist-info/RECORD,,
+ouroboros_ai-0.2.2.dist-info/METADATA,sha256=v_WwLBVh1qLB3vaFS2PwNCiODepMmq-zitj82ty4jPs,19661
+ouroboros_ai-0.2.2.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+ouroboros_ai-0.2.2.dist-info/entry_points.txt,sha256=MoETHup6rVkR6AsyjoRzAgIuvVtYYm3Jw40itV3_VyI,53
+ouroboros_ai-0.2.2.dist-info/licenses/LICENSE,sha256=n2X-q26TqpXnoBo0t_WouhFxWw663_q5FmbYDZayoHo,1060
+ouroboros_ai-0.2.2.dist-info/RECORD,,