Spaces:
Sleeping
Sleeping
final Changes
Browse files- Dockerfile +6 -0
- README.md +64 -1
- env.py +21 -0
- inference.py +6 -0
- pyproject.toml +37 -0
- requirements-docker.txt +6 -6
- server/__init__.py +1 -0
- server/app.py +41 -0
- test_env.py +1 -1
- uv.lock +0 -0
Dockerfile
CHANGED
|
@@ -13,6 +13,8 @@ COPY env.py .
|
|
| 13 |
COPY inference.py .
|
| 14 |
COPY openenv.yaml .
|
| 15 |
COPY README.md .
|
|
|
|
|
|
|
| 16 |
|
| 17 |
# Non-root user
|
| 18 |
RUN useradd --create-home appuser && chown -R appuser:appuser /app
|
|
@@ -20,4 +22,8 @@ USER appuser
|
|
| 20 |
|
| 21 |
EXPOSE 7860
|
| 22 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
CMD ["uvicorn", "env:app", "--host", "0.0.0.0", "--port", "7860"]
|
|
|
|
| 13 |
COPY inference.py .
|
| 14 |
COPY openenv.yaml .
|
| 15 |
COPY README.md .
|
| 16 |
+
COPY test_env.py .
|
| 17 |
+
COPY audit_reward.py .
|
| 18 |
|
| 19 |
# Non-root user
|
| 20 |
RUN useradd --create-home appuser && chown -R appuser:appuser /app
|
|
|
|
| 22 |
|
| 23 |
EXPOSE 7860
|
| 24 |
|
| 25 |
+
# Health check for HuggingFace Spaces
|
| 26 |
+
HEALTHCHECK --interval=30s --timeout=5s --start-period=10s --retries=3 \
|
| 27 |
+
CMD python -c "import urllib.request; urllib.request.urlopen('http://localhost:7860/')" || exit 1
|
| 28 |
+
|
| 29 |
CMD ["uvicorn", "env:app", "--host", "0.0.0.0", "--port", "7860"]
|
README.md
CHANGED
|
@@ -1,3 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
# ☁️ DevOps/FinOps Cloud Optimizer — OpenEnv Environment
|
| 2 |
|
| 3 |
**An AI agent environment where models learn to optimize cloud infrastructure cost vs. performance under realistic operational constraints.**
|
|
@@ -216,7 +232,7 @@ Generates a unique fleet configuration from the seed. Server count (3–7), work
|
|
| 216 |
|
| 217 |
```bash
|
| 218 |
# 1. Clone and set up
|
| 219 |
-
git clone
|
| 220 |
python -m venv venv && source venv/bin/activate # or .\venv\Scripts\activate on Windows
|
| 221 |
pip install -r requirements.txt
|
| 222 |
|
|
@@ -279,10 +295,57 @@ All scores are deterministic: same actions → same score, always.
|
|
| 279 |
| `/reset` | POST | `{"task_id": "easy\|medium\|hard\|random", "seed": <int>}` | `{session_id, ...Observation}` |
|
| 280 |
| `/step` | POST | `Action JSON` + `?session_id=ID` | `{observation, reward, done, info}` |
|
| 281 |
| `/state` | GET | `?session_id=ID` | Full state with action history |
|
|
|
|
| 282 |
| `/dashboard` | GET | `?session_id=ID` | Live HTML dashboard |
|
| 283 |
|
| 284 |
---
|
| 285 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 286 |
## Technical Stack
|
| 287 |
|
| 288 |
- **Python 3.11** + **FastAPI** + **Pydantic v2** + **uvicorn**
|
|
|
|
| 1 |
+
---
|
| 2 |
+
title: FinOps Cloud Optimizer Environment
|
| 3 |
+
emoji: ☁️
|
| 4 |
+
colorFrom: blue
|
| 5 |
+
colorTo: purple
|
| 6 |
+
sdk: docker
|
| 7 |
+
pinned: false
|
| 8 |
+
app_port: 7860
|
| 9 |
+
tags:
|
| 10 |
+
- openenv
|
| 11 |
+
- devops
|
| 12 |
+
- finops
|
| 13 |
+
- reinforcement-learning
|
| 14 |
+
- agents
|
| 15 |
+
---
|
| 16 |
+
|
| 17 |
# ☁️ DevOps/FinOps Cloud Optimizer — OpenEnv Environment
|
| 18 |
|
| 19 |
**An AI agent environment where models learn to optimize cloud infrastructure cost vs. performance under realistic operational constraints.**
|
|
|
|
| 232 |
|
| 233 |
```bash
|
| 234 |
# 1. Clone and set up
|
| 235 |
+
git clone https://github.com/ARYANPATEL-BIT/RLV-2.git && cd RLV-2
|
| 236 |
python -m venv venv && source venv/bin/activate # or .\venv\Scripts\activate on Windows
|
| 237 |
pip install -r requirements.txt
|
| 238 |
|
|
|
|
| 295 |
| `/reset` | POST | `{"task_id": "easy\|medium\|hard\|random", "seed": <int>}` | `{session_id, ...Observation}` |
|
| 296 |
| `/step` | POST | `Action JSON` + `?session_id=ID` | `{observation, reward, done, info}` |
|
| 297 |
| `/state` | GET | `?session_id=ID` | Full state with action history |
|
| 298 |
+
| `/info` | GET | — | Task catalog, instance types, reward formula |
|
| 299 |
| `/dashboard` | GET | `?session_id=ID` | Live HTML dashboard |
|
| 300 |
|
| 301 |
---
|
| 302 |
|
| 303 |
+
## Example Episode (Easy Task)
|
| 304 |
+
|
| 305 |
+
A complete episode walkthrough showing reset, one action, and the final state:
|
| 306 |
+
|
| 307 |
+
```bash
|
| 308 |
+
# 1. Reset to easy task
|
| 309 |
+
curl -X POST http://localhost:7860/reset \
|
| 310 |
+
-H "Content-Type: application/json" \
|
| 311 |
+
-d '{"task_id": "easy"}'
|
| 312 |
+
```
|
| 313 |
+
|
| 314 |
+
**Response:** An xlarge server ($1.60/hr) running two workloads. Budget is $0.25/hr.
|
| 315 |
+
|
| 316 |
+
```json
|
| 317 |
+
{
|
| 318 |
+
"session_id": "a1b2c3d4",
|
| 319 |
+
"servers": [{"server_id": "srv-001", "instance_type": "xlarge", "cpu_cores": 16, "ram_gb": 32, "cost_per_hour": 1.6, "assigned_workloads": ["wl-web", "wl-api"]}],
|
| 320 |
+
"workloads": [
|
| 321 |
+
{"workload_id": "wl-web", "name": "Company Website", "required_cpu": 0.5, "required_ram": 1.0, "current_latency_ms": 40.0, "sla_latency_ms": 200.0, "is_critical": false},
|
| 322 |
+
{"workload_id": "wl-api", "name": "API Gateway", "required_cpu": 1.0, "required_ram": 1.5, "current_latency_ms": 30.0, "sla_latency_ms": 150.0, "is_critical": true}
|
| 323 |
+
],
|
| 324 |
+
"total_cost_per_hour": 1.6, "budget_per_hour": 0.25, "sla_violations": 0,
|
| 325 |
+
"step_number": 0, "max_steps": 5, "done": false
|
| 326 |
+
}
|
| 327 |
+
```
|
| 328 |
+
|
| 329 |
+
```bash
|
| 330 |
+
# 2. Resize the server down to small (2 CPU, 4 GB — fits both workloads)
|
| 331 |
+
curl -X POST "http://localhost:7860/step?session_id=a1b2c3d4" \
|
| 332 |
+
-H "Content-Type: application/json" \
|
| 333 |
+
-d '{"action_type": "resize", "server_id": "srv-001", "instance_type": "small"}'
|
| 334 |
+
```
|
| 335 |
+
|
| 336 |
+
**Response:** Cost drops from $1.60 to $0.20. SLAs still met. Score 0.718.
|
| 337 |
+
|
| 338 |
+
```json
|
| 339 |
+
{
|
| 340 |
+
"observation": {"total_cost_per_hour": 0.2, "sla_violations": 0, "step_number": 1},
|
| 341 |
+
"reward": {"score": 0.718, "cost_efficiency": 0.92, "performance_score": 1.0, "penalty": 0.0,
|
| 342 |
+
"breakdown": "cost_eff=0.920×0.4 + perf=1.000×0.35 - penalty=0.000×0.25 = 0.7180"},
|
| 343 |
+
"done": false
|
| 344 |
+
}
|
| 345 |
+
```
|
| 346 |
+
|
| 347 |
+
---
|
| 348 |
+
|
| 349 |
## Technical Stack
|
| 350 |
|
| 351 |
- **Python 3.11** + **FastAPI** + **Pydantic v2** + **uvicorn**
|
env.py
CHANGED
|
@@ -23,6 +23,7 @@ from typing import Any, Dict, List, Literal, Optional, TypedDict
|
|
| 23 |
import uvicorn
|
| 24 |
from fastapi import FastAPI, HTTPException, Query
|
| 25 |
from fastapi.responses import HTMLResponse
|
|
|
|
| 26 |
from pydantic import BaseModel, Field
|
| 27 |
|
| 28 |
|
|
@@ -899,6 +900,15 @@ app = FastAPI(
|
|
| 899 |
description="AI agent environment for cloud infrastructure cost vs. performance optimization.",
|
| 900 |
version="2.0.0",
|
| 901 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 902 |
sessions = SessionManager()
|
| 903 |
|
| 904 |
|
|
@@ -906,6 +916,17 @@ sessions = SessionManager()
|
|
| 906 |
def health_check():
|
| 907 |
return {"status": "ok", "environment": "devops-finops-cloud-optimizer", "version": "2.0.0"}
|
| 908 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 909 |
|
| 910 |
@app.post("/reset")
|
| 911 |
def reset_endpoint(body: ResetRequest = ResetRequest()):
|
|
|
|
| 23 |
import uvicorn
|
| 24 |
from fastapi import FastAPI, HTTPException, Query
|
| 25 |
from fastapi.responses import HTMLResponse
|
| 26 |
+
from fastapi.middleware.cors import CORSMiddleware
|
| 27 |
from pydantic import BaseModel, Field
|
| 28 |
|
| 29 |
|
|
|
|
| 900 |
description="AI agent environment for cloud infrastructure cost vs. performance optimization.",
|
| 901 |
version="2.0.0",
|
| 902 |
)
|
| 903 |
+
|
| 904 |
+
app.add_middleware(
|
| 905 |
+
CORSMiddleware,
|
| 906 |
+
allow_origins=["*"],
|
| 907 |
+
allow_credentials=True,
|
| 908 |
+
allow_methods=["*"],
|
| 909 |
+
allow_headers=["*"],
|
| 910 |
+
)
|
| 911 |
+
|
| 912 |
sessions = SessionManager()
|
| 913 |
|
| 914 |
|
|
|
|
| 916 |
def health_check():
|
| 917 |
return {"status": "ok", "environment": "devops-finops-cloud-optimizer", "version": "2.0.0"}
|
| 918 |
|
| 919 |
+
@app.get("/info")
|
| 920 |
+
def info_endpoint():
|
| 921 |
+
return {
|
| 922 |
+
"environment": "devops-finops-cloud-optimizer",
|
| 923 |
+
"version": "2.0.0",
|
| 924 |
+
"tasks": TASK_CONFIGS,
|
| 925 |
+
"instance_catalog": INSTANCE_CATALOG,
|
| 926 |
+
"reward_formula": "score = (cost_efficiency × 0.40) + (performance_score × 0.35) - (penalty × 0.25)",
|
| 927 |
+
"endpoints": ["/reset", "/step", "/state", "/info", "/dashboard"]
|
| 928 |
+
}
|
| 929 |
+
|
| 930 |
|
| 931 |
@app.post("/reset")
|
| 932 |
def reset_endpoint(body: ResetRequest = ResetRequest()):
|
inference.py
CHANGED
|
@@ -211,6 +211,7 @@ def run_task(client: OpenAI, task_id: str) -> float:
|
|
| 211 |
max_steps = obs.get("max_steps", 5)
|
| 212 |
final_score = 0.0
|
| 213 |
reward = {}
|
|
|
|
| 214 |
|
| 215 |
print(f" Session: {session_id}")
|
| 216 |
print(f" Budget: ${obs.get('budget_per_hour', 0):.2f}/hr | "
|
|
@@ -224,6 +225,10 @@ def run_task(client: OpenAI, task_id: str) -> float:
|
|
| 224 |
break
|
| 225 |
|
| 226 |
user_prompt = build_user_prompt(obs)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 227 |
messages.append({"role": "user", "content": user_prompt})
|
| 228 |
|
| 229 |
# Keep conversation manageable (last 6 exchanges + system)
|
|
@@ -249,6 +254,7 @@ def run_task(client: OpenAI, task_id: str) -> float:
|
|
| 249 |
done = step_result.get("done", False)
|
| 250 |
final_score = reward.get("score", 0.0)
|
| 251 |
info_msg = step_result.get("info", {}).get("action_result", "")
|
|
|
|
| 252 |
|
| 253 |
print(f" -> score={final_score:.4f} | {info_msg[:70]}")
|
| 254 |
|
|
|
|
| 211 |
max_steps = obs.get("max_steps", 5)
|
| 212 |
final_score = 0.0
|
| 213 |
reward = {}
|
| 214 |
+
action_log = []
|
| 215 |
|
| 216 |
print(f" Session: {session_id}")
|
| 217 |
print(f" Budget: ${obs.get('budget_per_hour', 0):.2f}/hr | "
|
|
|
|
| 225 |
break
|
| 226 |
|
| 227 |
user_prompt = build_user_prompt(obs)
|
| 228 |
+
if action_log:
|
| 229 |
+
user_prompt += f"\n\n## Action History\nYou have taken {len(action_log)} actions so far:\n"
|
| 230 |
+
for past_step, (past_a, past_m) in enumerate(action_log):
|
| 231 |
+
user_prompt += f" Step {past_step+1}: {past_a.get('action_type', '?')} -> {past_m[:70]}\n"
|
| 232 |
messages.append({"role": "user", "content": user_prompt})
|
| 233 |
|
| 234 |
# Keep conversation manageable (last 6 exchanges + system)
|
|
|
|
| 254 |
done = step_result.get("done", False)
|
| 255 |
final_score = reward.get("score", 0.0)
|
| 256 |
info_msg = step_result.get("info", {}).get("action_result", "")
|
| 257 |
+
action_log.append((action, info_msg))
|
| 258 |
|
| 259 |
print(f" -> score={final_score:.4f} | {info_msg[:70]}")
|
| 260 |
|
pyproject.toml
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2026. All rights reserved.
|
| 2 |
+
|
| 3 |
+
[build-system]
|
| 4 |
+
requires = ["setuptools>=45", "wheel"]
|
| 5 |
+
build-backend = "setuptools.build_meta"
|
| 6 |
+
|
| 7 |
+
[project]
|
| 8 |
+
name = "openenv-devops-finops"
|
| 9 |
+
version = "2.0.0"
|
| 10 |
+
description = "AI agent environment for cloud infrastructure cost vs performance optimization."
|
| 11 |
+
authors = [
|
| 12 |
+
{ name = "aryanpatel" }
|
| 13 |
+
]
|
| 14 |
+
readme = "README.md"
|
| 15 |
+
requires-python = ">=3.10"
|
| 16 |
+
dependencies = [
|
| 17 |
+
"fastapi>=0.104.0",
|
| 18 |
+
"uvicorn>=0.24.0",
|
| 19 |
+
"pydantic>=2.0.0",
|
| 20 |
+
"openai>=1.0.0",
|
| 21 |
+
"requests>=2.31.0",
|
| 22 |
+
"openenv-core[core]>=0.2.0",
|
| 23 |
+
]
|
| 24 |
+
|
| 25 |
+
[project.optional-dependencies]
|
| 26 |
+
dev = [
|
| 27 |
+
"pytest>=8.0.0",
|
| 28 |
+
]
|
| 29 |
+
|
| 30 |
+
[project.scripts]
|
| 31 |
+
# Server entry point - enables running via: uv run --project . server
|
| 32 |
+
server = "devops_finops.server.app:main"
|
| 33 |
+
|
| 34 |
+
[tool.setuptools]
|
| 35 |
+
include-package-data = true
|
| 36 |
+
packages = ["devops_finops", "devops_finops.server"]
|
| 37 |
+
package-dir = { "devops_finops" = ".", "devops_finops.server" = "server" }
|
requirements-docker.txt
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
# requirements-docker.txt — Minimal deps for the OpenEnv environment
|
| 2 |
-
# Used by Dockerfile only
|
| 3 |
|
| 4 |
-
fastapi
|
| 5 |
-
uvicorn
|
| 6 |
-
pydantic
|
| 7 |
-
openai
|
| 8 |
-
requests
|
|
|
|
| 1 |
# requirements-docker.txt — Minimal deps for the OpenEnv environment
|
| 2 |
+
# Used by Dockerfile only. Versions pinned for reproducible builds.
|
| 3 |
|
| 4 |
+
fastapi==0.115.6
|
| 5 |
+
uvicorn==0.34.0
|
| 6 |
+
pydantic==2.10.4
|
| 7 |
+
openai==1.58.1
|
| 8 |
+
requests==2.32.3
|
server/__init__.py
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
# Server package for DevOps/FinOps OpenEnv
|
server/app.py
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""
|
| 2 |
+
FastAPI application entry point for the DevOps/FinOps OpenEnv Environment.
|
| 3 |
+
|
| 4 |
+
This module creates an HTTP server exposing the cloud optimization
|
| 5 |
+
environment over HTTP endpoints.
|
| 6 |
+
|
| 7 |
+
Usage:
|
| 8 |
+
# Development:
|
| 9 |
+
uvicorn server.app:app --reload --host 0.0.0.0 --port 7860
|
| 10 |
+
|
| 11 |
+
# Production:
|
| 12 |
+
uvicorn server.app:app --host 0.0.0.0 --port 7860
|
| 13 |
+
|
| 14 |
+
# Or run directly:
|
| 15 |
+
uv run --project . server
|
| 16 |
+
"""
|
| 17 |
+
|
| 18 |
+
import sys
|
| 19 |
+
import os
|
| 20 |
+
|
| 21 |
+
# Ensure the root directory is in the path so we can import env.py
|
| 22 |
+
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
|
| 23 |
+
|
| 24 |
+
from env import app # noqa: E402 — re-export the FastAPI app
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
def main():
|
| 28 |
+
"""
|
| 29 |
+
Entry point for direct execution via uv run or python -m.
|
| 30 |
+
|
| 31 |
+
This function enables running the server without Docker:
|
| 32 |
+
uv run --project . server
|
| 33 |
+
python -m devops_finops.server.app
|
| 34 |
+
"""
|
| 35 |
+
import uvicorn
|
| 36 |
+
|
| 37 |
+
uvicorn.run(app, host="0.0.0.0", port=7860)
|
| 38 |
+
|
| 39 |
+
|
| 40 |
+
if __name__ == "__main__":
|
| 41 |
+
main()
|
test_env.py
CHANGED
|
@@ -62,7 +62,7 @@ def test_cost_efficiency():
|
|
| 62 |
_compute_cost_efficiency(0.25, budget) > _compute_cost_efficiency(budget, budget))
|
| 63 |
check("Over budget < at budget",
|
| 64 |
_compute_cost_efficiency(0.80, budget) < _compute_cost_efficiency(budget, budget))
|
| 65 |
-
check("At
|
| 66 |
check("Monotonically decreasing",
|
| 67 |
_compute_cost_efficiency(0.10, budget) >
|
| 68 |
_compute_cost_efficiency(0.30, budget) >
|
|
|
|
| 62 |
_compute_cost_efficiency(0.25, budget) > _compute_cost_efficiency(budget, budget))
|
| 63 |
check("Over budget < at budget",
|
| 64 |
_compute_cost_efficiency(0.80, budget) < _compute_cost_efficiency(budget, budget))
|
| 65 |
+
check("At 3x budget → 0.0", _compute_cost_efficiency(1.50, budget) == 0.0)
|
| 66 |
check("Monotonically decreasing",
|
| 67 |
_compute_cost_efficiency(0.10, budget) >
|
| 68 |
_compute_cost_efficiency(0.30, budget) >
|
uv.lock
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|