Spaces:
Sleeping
Sleeping
Deploy: openenv.yaml
Browse files- openenv.yaml +47 -0
openenv.yaml
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
name: codearena-rl-benchmark
|
| 2 |
+
description: "RL Benchmark for Autonomous Code Repair — iterative debugging with execution feedback"
|
| 3 |
+
version: "1.0.0"
|
| 4 |
+
entrypoint: server.app:CodeArenaEnv
|
| 5 |
+
|
| 6 |
+
runtime:
|
| 7 |
+
language: python
|
| 8 |
+
python_version: "3.11"
|
| 9 |
+
|
| 10 |
+
api:
|
| 11 |
+
reset: /reset
|
| 12 |
+
step: /step
|
| 13 |
+
state: /state
|
| 14 |
+
|
| 15 |
+
observation_space:
|
| 16 |
+
type: json
|
| 17 |
+
schema:
|
| 18 |
+
buggy_code: string
|
| 19 |
+
error_log: string
|
| 20 |
+
test_results: string
|
| 21 |
+
previous_attempts: list[string]
|
| 22 |
+
|
| 23 |
+
action_space:
|
| 24 |
+
type: json
|
| 25 |
+
schema:
|
| 26 |
+
proposed_fix: string
|
| 27 |
+
|
| 28 |
+
tasks:
|
| 29 |
+
- id: easy
|
| 30 |
+
path: tasks/easy.json
|
| 31 |
+
grader: server.grader:grade
|
| 32 |
+
- id: medium
|
| 33 |
+
path: tasks/medium.json
|
| 34 |
+
grader: server.grader:grade
|
| 35 |
+
- id: hard
|
| 36 |
+
path: tasks/hard.json
|
| 37 |
+
grader: server.grader:grade
|
| 38 |
+
- id: type_errors
|
| 39 |
+
path: tasks/type_errors/type_error_1.json
|
| 40 |
+
grader: server.grader:grade
|
| 41 |
+
- id: security_bugs
|
| 42 |
+
path: tasks/security_bugs/security_bug_1.json
|
| 43 |
+
grader: server.grader:grade
|
| 44 |
+
|
| 45 |
+
limits:
|
| 46 |
+
step_timeout_seconds: 2
|
| 47 |
+
max_runtime_minutes: 20
|