File size: 9,625 Bytes
a3f3034
d145b94
 
 
 
b5ee03f
 
d145b94
 
 
 
 
 
 
 
 
 
 
 
a3f3034
 
 
 
 
 
d145b94
a3f3034
d145b94
 
 
 
 
 
 
 
a3f3034
 
 
 
 
 
 
 
 
 
 
d145b94
a3f3034
 
d145b94
a3f3034
 
 
d145b94
 
 
 
a3f3034
 
 
 
 
 
 
 
 
 
d145b94
a3f3034
 
d145b94
a3f3034
 
 
d145b94
 
 
 
a3f3034
 
 
 
 
 
 
d145b94
a3f3034
 
 
 
 
 
 
d145b94
a3f3034
 
d145b94
a3f3034
 
 
d145b94
 
 
 
 
b5ee03f
69cf4ae
 
d145b94
 
 
 
 
a3f3034
d145b94
 
 
 
 
 
 
 
 
 
 
 
 
a3f3034
 
 
 
 
 
 
 
 
 
 
 
 
d145b94
 
 
 
a3f3034
 
 
 
 
 
 
d145b94
a3f3034
d145b94
 
 
a3f3034
 
 
 
d145b94
 
a3f3034
d145b94
 
a3f3034
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d145b94
a3f3034
 
 
 
 
d145b94
 
 
 
 
 
a3f3034
d145b94
a3f3034
d145b94
a3f3034
d145b94
a3f3034
d145b94
69cf4ae
d145b94
 
 
 
 
69cf4ae
b5ee03f
 
d145b94
 
 
 
b5ee03f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
595a155
b5ee03f
 
 
 
 
69cf4ae
b5ee03f
69cf4ae
 
 
 
a3f3034
69cf4ae
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
"""PR-triage task definitions and grading utilities for coding_env."""

from __future__ import annotations

from dataclasses import dataclass
import json
import os
from typing import Dict, List, Tuple

try:
    from coding_env.models import CodeAction
except ImportError:
    from ..models import CodeAction


@dataclass(frozen=True)
class CodeReviewTask:
    task_id: str
    difficulty: str
    pr_title: str
    pr_description: str
    changed_files: Tuple[str, ...]
    unified_diff: str
    expected_file_path: str
    expected_issue_type: str
    expected_bug_type: str
    expected_severity: str
    expected_line_number: int
    expected_keywords: Tuple[str, ...]


TASKS: Dict[str, CodeReviewTask] = {
    "task_easy_1": CodeReviewTask(
        task_id="task_easy_1",
        difficulty="easy",
        pr_title="Fix average response time aggregation in metrics service",
        pr_description="Refactor aggregation logic and simplify average calculation.",
        changed_files=("services/metrics/aggregation.py",),
        unified_diff=(
            "diff --git a/services/metrics/aggregation.py b/services/metrics/aggregation.py\n"
            "@@ -10,6 +10,6 @@ def compute_avg(latencies):\n"
            "     total = 0\n"
            "     for latency in latencies:\n"
            "         total += latency\n"
            "-    return total / len(total)\n"
            "+    return total / len(total)\n"
        ),
        expected_file_path="services/metrics/aggregation.py",
        expected_issue_type="logic",
        expected_bug_type="logic",
        expected_severity="medium",
        expected_line_number=4,
        expected_keywords=("len(total)", "len(latencies)", "typeerror"),
    ),
    "task_medium_1": CodeReviewTask(
        task_id="task_medium_1",
        difficulty="medium",
        pr_title="Optimize login query path in auth service",
        pr_description="Use direct SQL construction for faster username/password checks.",
        changed_files=("services/auth/login.py",),
        unified_diff=(
            "diff --git a/services/auth/login.py b/services/auth/login.py\n"
            "@@ -21,7 +21,7 @@ def login(conn, username, password):\n"
            "-    query = \"SELECT * FROM users WHERE name=? AND pw=?\"\n"
            "-    return conn.execute(query, (username, password)).fetchone() is not None\n"
            "+    query = f\"SELECT * FROM users WHERE name='{username}' AND pw='{password}'\"\n"
            "+    return conn.execute(query).fetchone() is not None\n"
        ),
        expected_file_path="services/auth/login.py",
        expected_issue_type="security",
        expected_bug_type="security",
        expected_severity="high",
        expected_line_number=2,
        expected_keywords=("sql injection", "parameterized", "prepared statement"),
    ),
    "task_hard_1": CodeReviewTask(
        task_id="task_hard_1",
        difficulty="hard",
        pr_title="Add cache layer to user profile fetch endpoint",
        pr_description="Protect cache updates with lock to avoid races and keep data coherent.",
        changed_files=("services/profile/cache_layer.py",),
        unified_diff=(
            "diff --git a/services/profile/cache_layer.py b/services/profile/cache_layer.py\n"
            "@@ -4,12 +4,12 @@ lock = Lock()\n"
            " cache = {}\n"
            "\n"
            " def get_user(user_id, db):\n"
            "     with lock:\n"
            "         if user_id in cache:\n"
            "             return cache[user_id]\n"
            "         data = db.fetch_user(user_id)\n"
            "         cache[user_id] = data\n"
            "         return data\n"
        ),
        expected_file_path="services/profile/cache_layer.py",
        expected_issue_type="performance",
        expected_bug_type="logic",
        expected_severity="high",
        expected_line_number=7,
        expected_keywords=("lock contention", "critical section", "latency"),
    ),
}


EPISODE_SCORES: Dict[tuple[str, str], float] = {}
SCORE_FILE = os.getenv("CODING_ENV_SCORE_FILE", "/tmp/coding_env_episode_scores.json")
MIN_STRICT_SCORE = 0.01
MAX_STRICT_SCORE = 0.99


def list_tasks() -> List[Dict[str, str]]:
    """Return public task metadata."""
    return [
        {"task_id": t.task_id, "difficulty": t.difficulty, "pr_title": t.pr_title}
        for t in sorted(TASKS.values(), key=lambda item: item.task_id)
    ]


def get_task(task_id: str) -> CodeReviewTask:
    """Resolve task by id."""
    if task_id not in TASKS:
        raise ValueError(
            f"Unknown task_id '{task_id}'. Available tasks: {', '.join(sorted(TASKS))}"
        )
    return TASKS[task_id]


def format_task_prompt(task: CodeReviewTask) -> str:
    """Format a realistic PR-review prompt."""
    files = "\n".join(f"- {path}" for path in task.changed_files)
    return (
        f"PR Title: {task.pr_title}\n"
        f"PR Description: {task.pr_description}\n"
        f"Changed Files:\n{files}\n\n"
        f"Unified Diff:\n{task.unified_diff}\n\n"
        "Review objective: identify the highest-impact issue and provide "
        "file path, issue type, severity, and exact line."
    )


def _normalize(value: str) -> str:
    return value.strip().lower().replace("-", "_")


def _action_issue_type(action: CodeAction) -> str:
    issue_type = getattr(action, "issue_type", "")
    if issue_type:
        return str(issue_type)
    return str(action.bug_type)


def grade_action(action: CodeAction, task: CodeReviewTask) -> tuple[float, str]:
    """Score PR-triage action in strict (0, 1) with partial credit."""
    score = 0.0
    parts: List[str] = []

    file_path = str(getattr(action, "file_path", "") or "")
    if _normalize(file_path) == _normalize(task.expected_file_path):
        score += 0.30
        parts.append("file_path matched (+0.30)")
    else:
        parts.append(
            f"file_path mismatch (expected {task.expected_file_path}, got {file_path or 'none'})"
        )

    issue_type = _action_issue_type(action)
    if _normalize(issue_type) == _normalize(task.expected_issue_type):
        score += 0.30
        parts.append("issue_type matched (+0.30)")
    elif _normalize(action.bug_type) == _normalize(task.expected_bug_type):
        score += 0.20
        parts.append("bug_type matched (+0.20)")
    else:
        parts.append(
            f"issue mismatch (expected {task.expected_issue_type}/{task.expected_bug_type}, got {issue_type}/{action.bug_type})"
        )

    severity = str(getattr(action, "severity", "") or "")
    if _normalize(severity) == _normalize(task.expected_severity):
        score += 0.20
        parts.append("severity matched (+0.20)")
    elif severity:
        score += 0.10
        parts.append("severity provided but not exact (+0.10)")
    else:
        parts.append("severity missing (+0.00)")

    if action.line_number == task.expected_line_number:
        score += 0.10
        parts.append("line_number matched (+0.10)")
    elif abs(action.line_number - task.expected_line_number) <= 2:
        score += 0.05
        parts.append("line_number near miss (+0.05)")
    else:
        parts.append(
            f"line_number mismatch (expected {task.expected_line_number}, got {action.line_number})"
        )

    review_text = (action.review or "").lower()
    keyword_hits = sum(1 for kw in task.expected_keywords if kw.lower() in review_text)
    if keyword_hits > 0:
        keyword_bonus = min(0.09, keyword_hits * 0.03)
        score += keyword_bonus
        parts.append(f"evidence quality matched (+{keyword_bonus:.2f})")
    else:
        parts.append("insufficient evidence language (+0.00)")

    score = _to_strict_open_score(score)
    return score, "; ".join(parts)


def record_episode_score(task_id: str, episode_id: str, score: float) -> None:
    """Persist normalized score for grader endpoint."""
    normalized = _to_strict_open_score(score)
    EPISODE_SCORES[(task_id, episode_id)] = normalized
    _persist_score(task_id, episode_id, normalized)


def get_episode_score(task_id: str, episode_id: str) -> float:
    """Read score for task/episode pair."""
    in_memory = EPISODE_SCORES.get((task_id, episode_id))
    if in_memory is not None:
        return in_memory
    return _load_persisted_score(task_id, episode_id)


def _persist_score(task_id: str, episode_id: str, score: float) -> None:
    key = f"{task_id}::{episode_id}"
    payload: Dict[str, float] = {}
    if os.path.exists(SCORE_FILE):
        try:
            with open(SCORE_FILE, "r", encoding="utf-8") as f:
                loaded = json.load(f)
            if isinstance(loaded, dict):
                payload = {
                    str(k): float(v)
                    for k, v in loaded.items()
                    if isinstance(v, (int, float))
                }
        except Exception:
            payload = {}

    payload[key] = float(score)
    with open(SCORE_FILE, "w", encoding="utf-8") as f:
        json.dump(payload, f)


def _load_persisted_score(task_id: str, episode_id: str) -> float:
    if not os.path.exists(SCORE_FILE):
        return MIN_STRICT_SCORE
    try:
        with open(SCORE_FILE, "r", encoding="utf-8") as f:
            loaded = json.load(f)
        key = f"{task_id}::{episode_id}"
        value = loaded.get(key, 0.0) if isinstance(loaded, dict) else 0.0
        return _to_strict_open_score(value)
    except Exception:
        return MIN_STRICT_SCORE


def _to_strict_open_score(value: float) -> float:
    """Clamp to strict open interval (0, 1)."""
    return max(MIN_STRICT_SCORE, min(MAX_STRICT_SCORE, round(float(value), 4)))