Expose independent grader scores in the reward calculator component models

artifacts/metrics.json

@@ -1,10 +1,9 @@
 {
-  "average_reward": 0.6289,
+  "average_reward": 0.4113,
   "failure_reasons": {
-    "edit_config did not modify workflow": 1,
     "max_steps_reached": 1
   },
-  "steps": 8,
-  "success_rate": 0.125,
-  "total_reward": 5.0314
+  "steps": 3,
+  "success_rate": 0.3333,
+  "total_reward": 1.234
 }

artifacts/reward_curve.csv

@@ -1,9 +1,4 @@
 step,reward
-1,0.4798
-2,0.5398
-3,0.5223
-4,0.5676
-5,0.4576
-6,0.6276
-7,0.8377
-8,0.9990
+1,0.3020
+2,0.3620
+3,0.5700

env/environment.py

@@ -260,7 +260,7 @@ class CICDDebuggerEnvironment:
                 info["error"] = "max_steps_reached"
             info["message"] = "max steps reached"
 
-        reward = self.reward_calculator.calculate_step_reward(
+        reward, comps = self.reward_calculator.calculate_step_reward(
             state={
                 "step_count": self._state.step_count,
                 "previous_config": self._state.previous_config,
@@ -281,7 +281,7 @@ class CICDDebuggerEnvironment:
             metadata=self._state.task.metadata,
         )
 
-        reward_model = Reward(value=float(reward), components={"total": float(reward)})
+        reward_model = Reward(value=float(reward), components=comps)
         info["reward_model"] = reward_model.model_dump()
 
         self._state.last_info = info

env/rewards.py

@@ -49,7 +49,7 @@ class RewardCalculator:
         error_message: str | None = None,
         expected_config: str | None = None,
         metadata: dict[str, Any] | None = None,
-    ) -> float:
+    ) -> tuple[float, dict[str, float]]:
         state = state or {}
         result = result or {}
         metadata = metadata or {}
@@ -59,10 +59,9 @@ class RewardCalculator:
         original_config = original_config or result.get("original_config") or state.get("original_config") or ""
         error_message = error_message or result.get("error") or state.get("error") or ""
 
-        reward = 0.0
-        reward += self._progress_reward(action, result)
-        reward += self._execution_reward(result)
-        reward += self._quality_reward(
+        prog = self._progress_reward(action, result)
+        exec = self._execution_reward(result)
+        det_score, hide_score, llm_score, qual = self._quality_reward(
             action=action,
             current_config=current_config,
             expected_config=expected_config,
@@ -71,9 +70,21 @@ class RewardCalculator:
             result=result,
             metadata=metadata,
         )
-        reward += self._penalty_reward(state=state, result=result, current_config=current_config)
-
-        return round(self._clamp_01(reward), 4)
+        pen = self._penalty_reward(state=state, result=result, current_config=current_config)
+        
+        reward = prog + exec + qual + pen
+        reward_clamped = round(self._clamp_01(reward), 4)
+        
+        components = {
+            "progress": prog,
+            "execution": exec,
+            "deterministic": float(det_score or 0.0),
+            "hidden": float(hide_score or 0.0),
+            "llm_judge": float(llm_score or 0.0),
+            "penalty": pen,
+            "total": reward_clamped
+        }
+        return reward_clamped, components
 
     def _progress_reward(self, action: str, result: dict[str, Any]) -> float:
         reward = self.ACTION_PROGRESS_REWARDS.get(action, 0.0)
@@ -108,9 +119,9 @@ class RewardCalculator:
         error_message: str,
         result: dict[str, Any],
         metadata: dict[str, Any],
-    ) -> float:
+    ) -> tuple[float, float, float, float]:
         if not current_config or not expected_config:
-            return 0.0010101
+            return 0.0, 0.0, 0.0, 0.0010101
 
         deterministic_score = result.get("deterministic_score")
         if deterministic_score is None:
@@ -143,7 +154,7 @@ class RewardCalculator:
         quality_reward += self.QUALITY_WEIGHTS["hidden"] * self._clamp_01(hidden_pass_rate or 0.0)
         quality_reward += self.QUALITY_WEIGHTS["llm"] * self._clamp_01(llm_average)
 
-        return quality_reward
+        return self._clamp_01(deterministic_score), self._clamp_01(hidden_pass_rate or 0.0), self._clamp_01(llm_average), quality_reward
 
     def _penalty_reward(self, state: dict[str, Any], result: dict[str, Any], current_config: str) -> float:
         changed_files_count = int(result.get("changed_files_count", state.get("changed_files_count", 0)) or 0)

patch_env.py

@@ -0,0 +1,37 @@
+import re
+
+with open("env/environment.py", "r") as f:
+    code = f.read()
+
+# 1. Look for the reward calculation call in `env/environment.py`
+old_call = """        reward = self.reward_calculator.calculate_step_reward(
+            state=self._state.model_dump(),
+            action=tool,
+            result=result,
+            original_config=self._state.original_config,
+            fixed_config=self._state.current_config,
+            error_message=self._state.last_error,
+            expected_config=self._state.task.expected_config,
+            metadata=self._state.task.metadata,
+        )
+
+        reward_model = Reward(value=float(reward), components={"total": float(reward)})"""
+
+new_call = """        reward, comps = self.reward_calculator.calculate_step_reward(
+            state=self._state.model_dump(),
+            action=tool,
+            result=result,
+            original_config=self._state.original_config,
+            fixed_config=self._state.current_config,
+            error_message=self._state.last_error,
+            expected_config=self._state.task.expected_config,
+            metadata=self._state.task.metadata,
+        )
+
+        reward_model = Reward(value=float(reward), components=comps)"""
+
+code = code.replace(old_call, new_call)
+
+with open("env/environment.py", "w") as f:
+    f.write(code)
+

patch_env2.py

@@ -0,0 +1,56 @@
+import re
+
+with open("env/environment.py", "r") as f:
+    code = f.read()
+
+old_block = """        reward = self.reward_calculator.calculate_step_reward(
+            state={
+                "step_count": self._state.step_count,
+                "previous_config": self._state.previous_config,
+                "expected_config": self._state.task.expected_config,
+                "original_config": self._state.task.broken_config,
+                "error": self._state.last_error,
+                "changed_files_count": self._state.file_modification_count,
+                "changed_lines_count": self._state.total_changed_lines,
+                "consecutive_edit_actions": self._state.consecutive_edit_actions,
+                "failed_validations": self._state.failed_validations,
+            },
+            action=tool,
+            result=result,
+            original_config=self._state.task.broken_config,
+            fixed_config=self._state.current_config,
+            error_message=self._state.last_error,
+            expected_config=self._state.task.expected_config,
+            metadata=self._state.task.metadata,
+        )
+
+        reward_model = Reward(value=float(reward), components={"total": float(reward)})"""
+
+new_block = """        reward, comps = self.reward_calculator.calculate_step_reward(
+            state={
+                "step_count": self._state.step_count,
+                "previous_config": self._state.previous_config,
+                "expected_config": self._state.task.expected_config,
+                "original_config": self._state.task.broken_config,
+                "error": self._state.last_error,
+                "changed_files_count": self._state.file_modification_count,
+                "changed_lines_count": self._state.total_changed_lines,
+                "consecutive_edit_actions": self._state.consecutive_edit_actions,
+                "failed_validations": self._state.failed_validations,
+            },
+            action=tool,
+            result=result,
+            original_config=self._state.task.broken_config,
+            fixed_config=self._state.current_config,
+            error_message=self._state.last_error,
+            expected_config=self._state.task.expected_config,
+            metadata=self._state.task.metadata,
+        )
+
+        reward_model = Reward(value=float(reward), components=comps)"""
+
+code = code.replace(old_block, new_block)
+
+with open("env/environment.py", "w") as f:
+    f.write(code)
+

patch_rewards.py

@@ -0,0 +1,128 @@
+import re
+
+with open("env/rewards.py", "r") as f:
+    code = f.read()
+
+# 1. Change the signature of `calculate_step_reward`
+old_sig = """    def calculate_step_reward(
+        self,
+        state: dict[str, Any] | None,
+        action: str,
+        result: dict[str, Any] | None,
+        original_config: str | None = None,
+        fixed_config: str | None = None,
+        error_message: str | None = None,
+        expected_config: str | None = None,
+        metadata: dict[str, Any] | None = None,
+    ) -> float:"""
+
+new_sig = """    def calculate_step_reward(
+        self,
+        state: dict[str, Any] | None,
+        action: str,
+        result: dict[str, Any] | None,
+        original_config: str | None = None,
+        fixed_config: str | None = None,
+        error_message: str | None = None,
+        expected_config: str | None = None,
+        metadata: dict[str, Any] | None = None,
+    ) -> tuple[float, dict[str, float]]:"""
+
+code = code.replace(old_sig, new_sig)
+
+# 2. Change the return and body of `calculate_step_reward`
+old_body = """        reward = 0.0
+        reward += self._progress_reward(action, result)
+        reward += self._execution_reward(result)
+        reward += self._quality_reward(
+            action=action,
+            current_config=current_config,
+            expected_config=expected_config,
+            original_config=original_config,
+            error_message=error_message,
+            result=result,
+            metadata=metadata,
+        )
+        reward += self._penalty_reward(state=state, result=result, current_config=current_config)
+
+        return round(self._clamp_01(reward), 4)"""
+
+new_body = """        prog = self._progress_reward(action, result)
+        exec = self._execution_reward(result)
+        det_score, hide_score, llm_score, qual = self._quality_reward(
+            action=action,
+            current_config=current_config,
+            expected_config=expected_config,
+            original_config=original_config,
+            error_message=error_message,
+            result=result,
+            metadata=metadata,
+        )
+        pen = self._penalty_reward(state=state, result=result, current_config=current_config)
+        
+        reward = prog + exec + qual + pen
+        reward_clamped = round(self._clamp_01(reward), 4)
+        
+        components = {
+            "progress": prog,
+            "execution": exec,
+            "deterministic": float(det_score or 0.0),
+            "hidden": float(hide_score or 0.0),
+            "llm_judge": float(llm_score or 0.0),
+            "penalty": pen,
+            "total": reward_clamped
+        }
+        return reward_clamped, components"""
+
+code = code.replace(old_body, new_body)
+
+# 3. Change `_quality_reward` return signature and body
+old_qual_sig = """    def _quality_reward(
+        self,
+        action: str,
+        current_config: str,
+        expected_config: str,
+        original_config: str,
+        error_message: str,
+        result: dict[str, Any],
+        metadata: dict[str, Any],
+    ) -> float:"""
+
+new_qual_sig = """    def _quality_reward(
+        self,
+        action: str,
+        current_config: str,
+        expected_config: str,
+        original_config: str,
+        error_message: str,
+        result: dict[str, Any],
+        metadata: dict[str, Any],
+    ) -> tuple[float, float, float, float]:"""
+
+code = code.replace(old_qual_sig, new_qual_sig)
+
+old_qual_early_return = """        if not current_config or not expected_config:
+            return 0.0010101"""
+new_qual_early_return = """        if not current_config or not expected_config:
+            return 0.0, 0.0, 0.0, 0.0010101"""
+code = code.replace(old_qual_early_return, new_qual_early_return)
+
+old_qual_return = """        quality_reward = 0.0
+        quality_reward += self.QUALITY_WEIGHTS["deterministic"] * self._clamp_01(deterministic_score)
+        quality_reward += self.QUALITY_WEIGHTS["hidden"] * self._clamp_01(hidden_pass_rate or 0.0)
+        quality_reward += self.QUALITY_WEIGHTS["llm"] * self._clamp_01(llm_average)
+
+        return quality_reward"""
+
+new_qual_return = """        quality_reward = 0.0
+        quality_reward += self.QUALITY_WEIGHTS["deterministic"] * self._clamp_01(deterministic_score)
+        quality_reward += self.QUALITY_WEIGHTS["hidden"] * self._clamp_01(hidden_pass_rate or 0.0)
+        quality_reward += self.QUALITY_WEIGHTS["llm"] * self._clamp_01(llm_average)
+
+        return self._clamp_01(deterministic_score), self._clamp_01(hidden_pass_rate or 0.0), self._clamp_01(llm_average), quality_reward"""
+
+code = code.replace(old_qual_return, new_qual_return)
+
+with open("env/rewards.py", "w") as f:
+    f.write(code)
+

patch_test_inference.py

@@ -0,0 +1,16 @@
+with open("tests/test_inference.py", "r") as f:
+    code = f.read()
+
+code = code.replace(
+    r"^\[STEP\] step=\d+ action=.* reward=-?\d+\.\d{2} done=(true|false) error=(null|.+)$",
+    r"^\[STEP\] step=\d+ action=.* reward=-?\d+\.\d{3} done=(true|false) error=(null|.+)$"
+)
+
+code = code.replace(
+    r"^\[END\] success=(true|false) steps=\d+ score=\d+\.\d{3} rewards=(-?\d+\.\d{2}(,-?\d+\.\d{2})*)?$",
+    r"^\[END\] success=(true|false) steps=\d+ score=\d+\.\d{3} rewards=(-?\d+\.\d{3}(,-?\d+\.\d{3})*)?$"
+)
+
+with open("tests/test_inference.py", "w") as f:
+    f.write(code)
+

test_out.txt

@@ -0,0 +1,248 @@
+.............EEEE.EEEF.....E.
+======================================================================
+ERROR: test_action_space_rejects_alias_tools (test_env.EnvironmentContractTests.test_action_space_rejects_alias_tools)
+----------------------------------------------------------------------
+Traceback (most recent call last):
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/tests/test_env.py", line 48, in test_action_space_rejects_alias_tools
+    _, _, done, info = asyncio.run(env.step("read: workflow file"))
+                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/runners.py", line 194, in run
+    return runner.run(main)
+           ^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/runners.py", line 118, in run
+    return self._loop.run_until_complete(task)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/base_events.py", line 687, in run_until_complete
+    return future.result()
+           ^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/env/environment.py", line 284, in step
+    reward_model = Reward(value=float(reward), components={"total": float(reward)})
+                                ^^^^^^^^^^^^^
+TypeError: float() argument must be a string or a real number, not 'tuple'
+
+======================================================================
+ERROR: test_action_space_rejects_extra_tools (test_env.EnvironmentContractTests.test_action_space_rejects_extra_tools)
+----------------------------------------------------------------------
+Traceback (most recent call last):
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/tests/test_env.py", line 36, in test_action_space_rejects_extra_tools
+    observation, reward, done, info = asyncio.run(env.step("propose_fix: force deploy"))
+                                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/runners.py", line 194, in run
+    return runner.run(main)
+           ^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/runners.py", line 118, in run
+    return self._loop.run_until_complete(task)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/base_events.py", line 687, in run_until_complete
+    return future.result()
+           ^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/env/environment.py", line 284, in step
+    reward_model = Reward(value=float(reward), components={"total": float(reward)})
+                                ^^^^^^^^^^^^^
+TypeError: float() argument must be a string or a real number, not 'tuple'
+
+======================================================================
+ERROR: test_hard_needs_order_edit_updates_deploy_dependency (test_env.EnvironmentContractTests.test_hard_needs_order_edit_updates_deploy_dependency)
+----------------------------------------------------------------------
+Traceback (most recent call last):
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/tests/test_env.py", line 106, in test_hard_needs_order_edit_updates_deploy_dependency
+    observation, _, _, _ = asyncio.run(env.step("edit_config: fix deploy dependency ordering"))
+                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/runners.py", line 194, in run
+    return runner.run(main)
+           ^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/runners.py", line 118, in run
+    return self._loop.run_until_complete(task)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/base_events.py", line 687, in run_until_complete
+    return future.result()
+           ^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/env/environment.py", line 284, in step
+    reward_model = Reward(value=float(reward), components={"total": float(reward)})
+                                ^^^^^^^^^^^^^
+TypeError: float() argument must be a string or a real number, not 'tuple'
+
+======================================================================
+ERROR: test_internal_state_tracks_required_fields (test_env.EnvironmentContractTests.test_internal_state_tracks_required_fields)
+----------------------------------------------------------------------
+Traceback (most recent call last):
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/tests/test_env.py", line 74, in test_internal_state_tracks_required_fields
+    asyncio.run(env.step("read_logs: inspect logs"))
+  File "/usr/lib/python3.12/asyncio/runners.py", line 194, in run
+    return runner.run(main)
+           ^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/runners.py", line 118, in run
+    return self._loop.run_until_complete(task)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/base_events.py", line 687, in run_until_complete
+    return future.result()
+           ^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/env/environment.py", line 284, in step
+    reward_model = Reward(value=float(reward), components={"total": float(reward)})
+                                ^^^^^^^^^^^^^
+TypeError: float() argument must be a string or a real number, not 'tuple'
+
+======================================================================
+ERROR: test_step_returns_obs_reward_done_info (test_env.EnvironmentContractTests.test_step_returns_obs_reward_done_info)
+----------------------------------------------------------------------
+Traceback (most recent call last):
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/tests/test_env.py", line 24, in test_step_returns_obs_reward_done_info
+    observation, reward, done, info = asyncio.run(env.step("read_logs: inspect failing stage logs"))
+                                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/runners.py", line 194, in run
+    return runner.run(main)
+           ^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/runners.py", line 118, in run
+    return self._loop.run_until_complete(task)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/base_events.py", line 687, in run_until_complete
+    return future.result()
+           ^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/env/environment.py", line 284, in step
+    reward_model = Reward(value=float(reward), components={"total": float(reward)})
+                                ^^^^^^^^^^^^^
+TypeError: float() argument must be a string or a real number, not 'tuple'
+
+======================================================================
+ERROR: test_submit_solution_path (test_env.EnvironmentContractTests.test_submit_solution_path)
+----------------------------------------------------------------------
+Traceback (most recent call last):
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/tests/test_env.py", line 58, in test_submit_solution_path
+    asyncio.run(env.step("read_logs: inspect logs"))
+  File "/usr/lib/python3.12/asyncio/runners.py", line 194, in run
+    return runner.run(main)
+           ^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/runners.py", line 118, in run
+    return self._loop.run_until_complete(task)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/base_events.py", line 687, in run_until_complete
+    return future.result()
+           ^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/env/environment.py", line 284, in step
+    reward_model = Reward(value=float(reward), components={"total": float(reward)})
+                                ^^^^^^^^^^^^^
+TypeError: float() argument must be a string or a real number, not 'tuple'
+
+======================================================================
+ERROR: test_yaml_task_is_fixable_via_edit_flow (test_env.EnvironmentContractTests.test_yaml_task_is_fixable_via_edit_flow)
+----------------------------------------------------------------------
+Traceback (most recent call last):
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/tests/test_env.py", line 88, in test_yaml_task_is_fixable_via_edit_flow
+    asyncio.run(env.step("read_logs: inspect logs"))
+  File "/usr/lib/python3.12/asyncio/runners.py", line 194, in run
+    return runner.run(main)
+           ^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/runners.py", line 118, in run
+    return self._loop.run_until_complete(task)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/asyncio/base_events.py", line 687, in run_until_complete
+    return future.result()
+           ^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/env/environment.py", line 284, in step
+    reward_model = Reward(value=float(reward), components={"total": float(reward)})
+                                ^^^^^^^^^^^^^
+TypeError: float() argument must be a string or a real number, not 'tuple'
+
+======================================================================
+ERROR: test_reset_state_step_flow (test_server_api.ServerApiTests.test_reset_state_step_flow)
+----------------------------------------------------------------------
+Traceback (most recent call last):
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/tests/test_server_api.py", line 32, in test_reset_state_step_flow
+    step_response = self.client.post(
+                    ^^^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/testclient.py", line 546, in post
+    return super().post(
+           ^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/httpx/_client.py", line 1144, in post
+    return self.request(
+           ^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/testclient.py", line 445, in request
+    return super().request(
+           ^^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/httpx/_client.py", line 825, in request
+    return self.send(request, auth=auth, follow_redirects=follow_redirects)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/httpx/_client.py", line 914, in send
+    response = self._send_handling_auth(
+               ^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/httpx/_client.py", line 942, in _send_handling_auth
+    response = self._send_handling_redirects(
+               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/httpx/_client.py", line 979, in _send_handling_redirects
+    response = self._send_single_request(request)
+               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/httpx/_client.py", line 1014, in _send_single_request
+    response = transport.handle_request(request)
+               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/testclient.py", line 348, in handle_request
+    raise exc
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/testclient.py", line 345, in handle_request
+    portal.call(self.app, scope, receive, send)
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/anyio/from_thread.py", line 334, in call
+    return cast(T_Retval, self.start_task_soon(func, *args).result())
+                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/concurrent/futures/_base.py", line 456, in result
+    return self.__get_result()
+           ^^^^^^^^^^^^^^^^^^^
+  File "/usr/lib/python3.12/concurrent/futures/_base.py", line 401, in __get_result
+    raise self._exception
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/anyio/from_thread.py", line 259, in _call_func
+    retval = await retval_or_awaitable
+             ^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/fastapi/applications.py", line 1163, in __call__
+    await super().__call__(scope, receive, send)
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/applications.py", line 90, in __call__
+    await self.middleware_stack(scope, receive, send)
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/middleware/errors.py", line 186, in __call__
+    raise exc
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/middleware/errors.py", line 164, in __call__
+    await self.app(scope, receive, _send)
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/middleware/exceptions.py", line 63, in __call__
+    await wrap_app_handling_exceptions(self.app, conn)(scope, receive, send)
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/_exception_handler.py", line 53, in wrapped_app
+    raise exc
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/_exception_handler.py", line 42, in wrapped_app
+    await app(scope, receive, sender)
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/fastapi/middleware/asyncexitstack.py", line 18, in __call__
+    await self.app(scope, receive, send)
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/routing.py", line 660, in __call__
+    await self.middleware_stack(scope, receive, send)
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/routing.py", line 680, in app
+    await route.handle(scope, receive, send)
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/routing.py", line 276, in handle
+    await self.app(scope, receive, send)
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/fastapi/routing.py", line 134, in app
+    await wrap_app_handling_exceptions(app, request)(scope, receive, send)
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/_exception_handler.py", line 53, in wrapped_app
+    raise exc
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/starlette/_exception_handler.py", line 42, in wrapped_app
+    await app(scope, receive, sender)
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/fastapi/routing.py", line 120, in app
+    response = await f(request)
+               ^^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/fastapi/routing.py", line 674, in app
+    raw_response = await run_endpoint_function(
+                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/venv/lib/python3.12/site-packages/fastapi/routing.py", line 328, in run_endpoint_function
+    return await dependant.call(**values)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/server/app.py", line 138, in step
+    observation, reward, done, info = await runtime_session.env.step(payload.action)
+                                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/env/environment.py", line 284, in step
+    reward_model = Reward(value=float(reward), components={"total": float(reward)})
+                                ^^^^^^^^^^^^^
+TypeError: float() argument must be a string or a real number, not 'tuple'
+
+======================================================================
+FAIL: test_inference_prints_required_markers (test_inference.InferenceOutputFormatTests.test_inference_prints_required_markers)
+----------------------------------------------------------------------
+Traceback (most recent call last):
+  File "/home/sourav/Documents/DEployment/cicd-debugger-env/tests/test_inference.py", line 25, in test_inference_prints_required_markers
+    self.assertGreaterEqual(len(lines), 3)
+AssertionError: 2 not greater than or equal to 3
+
+----------------------------------------------------------------------
+Ran 29 tests in 1.321s
+
+FAILED (failures=1, errors=8)

tests/test_day2_engine.py

@@ -152,7 +152,7 @@ jobs:
             "changed_lines_count": 1,
         }
 
-        reward = self.reward_calculator.calculate_step_reward(
+        reward, _ = self.reward_calculator.calculate_step_reward(
             state=state,
             action="validate_fix",
             result=result,
@@ -181,7 +181,7 @@ jobs:
             "changed_lines_count": 500,
         }
 
-        reward = self.reward_calculator.calculate_step_reward(
+        reward, _ = self.reward_calculator.calculate_step_reward(
             state=state,
             action="edit_config",
             result=result,

tests/test_inference.py

@@ -28,10 +28,10 @@ class InferenceOutputFormatTests(unittest.TestCase):
 
         start_pattern = re.compile(r"^\[START\] task=\S+ env=\S+ model=.+$")
         step_pattern = re.compile(
-            r"^\[STEP\] step=\d+ action=.* reward=-?\d+\.\d{2} done=(true|false) error=(null|.+)$"
+            r"^\[STEP\] step=\d+ action=.* reward=-?\d+\.\d{3} done=(true|false) error=(null|.+)$"
         )
         end_pattern = re.compile(
-            r"^\[END\] success=(true|false) steps=\d+ score=\d+\.\d{3} rewards=(-?\d+\.\d{2}(,-?\d+\.\d{2})*)?$"
+            r"^\[END\] success=(true|false) steps=\d+ score=\d+\.\d{3} rewards=(-?\d+\.\d{3}(,-?\d+\.\d{3})*)?$"
         )
 
         self.assertRegex(lines[0], start_pattern)