File size: 18,688 Bytes
5dd1bb4 9e64e71 5dd1bb4 a001a97 5dd1bb4 9e64e71 5dd1bb4 a001a97 5dd1bb4 9e64e71 a001a97 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 a001a97 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 a001a97 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 a001a97 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 5dd1bb4 9e64e71 a001a97 9e64e71 5dd1bb4 9e64e71 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 | # Architecture
> Last updated: 2026-03-29
System map for SQLEnv, an RL environment where agents learn interactive SQL exploration via the OpenEnv framework.
**Goals:**
- Show how components connect (system map + key flows)
- Make hidden state explicit (what lives where)
- Define shared interfaces (Pydantic models, HTTP API)
- Keep invariants legible (what must stay true)
**Non-goals:**
- Exhaustive API reference
- Training hyperparameter tuning guide
---
## System Map
```text
SQLEnv System
================================================================
RL Training SQLEnv Server (Docker)
βββββββββββββ ββββββββββββββββββββββ
+ββββββββββββββ+ +βββββββββββββββββββββ+
β TRL GRPO β β server/app.py β
β Trainer β HTTP (JSON) β FastAPI + OpenEnv β
β β<========================>β β
β training/ β SQLAction -> server +βββββββββββ¬ββββββββββ+
β trl_adapter β SQLObs <- server β
β .py β v
+ββββββββββββββ+ +βββββββββββββββββββββ+
β β SQLEnvironment β
β OR β (sql_environment.py)β
v β β
+ββββββββββββββ+ β reset() / step() β
β Custom β β action dispatch β
β rollout_func β +βββ¬βββββββ¬βββββββ¬ββββ+
β (rollout.py) β β β β
+ββββββββββββββ+ v v v
+ββββββββββββββββββββββββ+
Evaluation β Action Handlers β
ββββββββββ β DESCRIBE β PRAGMA β
+ββββββββββββββ+ β SAMPLE β SELECT N β
β evaluate() βββ> env.reset/step β QUERY β SQL exec β
β policies β β ANSWER β verifier β
β .py β +βββββββββ¬βββββββββββββββ+
+ββββββββββββββ+ β
β v
+ββββββββββββββ+ +ββββββββββββββββββββββββ+
β Policies β β SQLite (read-only) β
β RandomPolicy β β data/databases/ β
β OraclePolicy β β {db_id}/{db_id}.sqlite β
+ββββββββββββββ+ +ββββββββββββββββββββββββ+
β
βββββββββ΄ββββββββ
v v
+βββββββββββ+ +βββββββββββ+
β reward.py β βverifier.pyβ
β 3-layer β β type-awareβ
β dense β β comparisonβ
+βββββββββββ+ +βββββββββββ+
Data (committed) Synthetic (optional)
ββββββββββββββββ ββββββββββββββββββββ
data/questions/ server/synthetic/
questions_train.json (473 Q) generate.py
questions_eval.json (203 Q) mutations.py
db_list.json (10 databases) validate.py
```
---
## Component Inventory
| Component | Owns | Entrypoint | State / Output |
|-----------|------|------------|----------------|
| **SQLEnvironment** | Episode lifecycle, action dispatch, step budget | `server/sql_environment.py` | `EpisodeContext` (in-memory, per episode) |
| **FastAPI app** | HTTP endpoints, tokenizer factory | `server/app.py` | Stateless (delegates to environment) |
| **SQLEnvClient** | HTTP transport, payload serialization | `client.py` | Stateless (wraps server) |
| **Pydantic models** | Type contracts (action, observation, state) | `models.py` | N/A (data classes) |
| **Reward engine** | 3-layer dense reward computation | `server/reward.py` | Mutates `EpisodeContext` accumulators |
| **Answer verifier** | Type-aware answer comparison | `server/verifier.py` | Stateless (pure function) |
| **GRPO pipeline** | Training orchestration, rollout, reward callables | `training/` (6 modules) | Training artifacts in `outputs/` |
| **TRL adapter** | `environment_factory` for TRL GRPOTrainer | `training/trl_adapter.py` | Per-session environment instances |
| **Evaluation** | Policy protocol, evaluate() runner | `evaluation/policies.py` | `EvaluationResult` metrics |
| **Oracle policy** | Deterministic upper-bound baseline | `evaluation/oracle_policy.py` | Stateless per-step |
| **Synthetic DB gen** | Metamorphic testing via data mutations | `server/synthetic/` | Variant SQLite files |
| **Question dataset** | 676 curated Spider questions across 10 DBs | `data/questions/` | JSON files |
### External Dependencies
| Dependency | Purpose | Required |
|------------|---------|----------|
| SQLite (stdlib) | Database execution | Yes |
| OpenEnv (`openenv-core`) | Environment protocol, `create_app` | Yes |
| TRL (`trl`) | GRPO training | Only for training |
| HuggingFace Transformers | Tokenizer loading | Only for production server |
---
## Key Flows
### Flow: Episode (Reset + Multi-Turn Steps)
```text
Client / Policy SQLEnvironment
β β
βββ reset(seed=42) ββββββββββββββββ> β
β βββ pick question (random or seeded)
β βββ open read-only SQLite connection
β βββ execute gold_sql β store gold_rows
β βββ init EpisodeContext (budget=15)
β <ββ SQLObservation βββββββββββββββββ
β .question="How many students?" β
β .schema_info="Tables: student" β (column details hidden)
β .budget_remaining=15 β
β β
βββ step(DESCRIBE student) ββββββββ> β
β βββ PRAGMA table_info(student)
β βββ add to described_tables
β βββ compute_step_reward()
β <ββ SQLObservation βββββββββββββββββ
β .schema_info="student: id INT" β (columns now revealed)
β .result="5 columns, 20 rows" β
β .reward=0.02 β
β .budget_remaining=14 β
β β
βββ step(QUERY "SELECT COUNT(*)...") β
β βββ validate (SELECT-only, single stmt)
β βββ execute with 5s timeout
β βββ compute_step_reward() (L1 + L2)
β <ββ SQLObservation βββββββββββββββββ
β .result="| COUNT(*) |\n| 20 |" β
β .reward=0.035 β
β β
βββ step(ANSWER "20") βββββββββββββ> β
β βββ verify_answer("20", gold, type)
β βββ terminal reward: +1.0 or 0.0
β <ββ SQLObservation βββββββββββββββββ
β .done=true β
β .reward=1.0 β
```
### Flow: 3-Layer Reward Computation
```text
step() called with action
β
v
Layer 1: Operational Shaping (every action)
βββ exec_ok? β +0.02
βββ new SQL hash? β +0.01 (per unique query, no cumulative cap)
βββ repeated SQL? β -0.01
βββ step cost β -0.005
β
v (only if action_type == QUERY and no error)
Layer 2: Progress Shaping (delta-from-previous, PBRS)
βββ cardinality score (25%) β |pred_rows - gold_rows| / max
βββ value overlap (50%) β Jaccard of cell values
βββ numeric range (25%) β log-distance proximity
β
v
bin to {0.0, 0.25, 0.5, 0.75, 1.0}
delta = binned - previous_progress β delta * 0.15
(positive = improvement, negative = regression)
β
v
Clip per step to [-0.05, +0.15]
No cumulative tracking
β
v (on ANSWER action)
Layer 3: Terminal Correctness
βββ verify_answer() β +1.0 (correct) or 0.0 (wrong)
```
### Flow: TRL Training Integration
```text
GRPOTrainer
β
βββ discovers tool methods via docstrings
β (describe, sample, query, answer)
β
βββ per rollout:
β SQLEnvTRL() β SQLEnvironment (internal)
β .reset() β observation string
β .describe(table) β schema string
β .query(sql) β result string
β .answer(value) β final string
β
βββ reward:
β sql_env_reward_func() β accumulated .reward
β
v
Training loop (GRPO: generate N completions, rank by reward)
```
---
## Shared Data Models
Defined in `models.py`. These cross the HTTP boundary between client and server.
### SQLAction (agent -> server)
```python
class SQLAction(Action):
action_type: str # DESCRIBE | SAMPLE | QUERY | ANSWER
argument: str # table name, SQL string, or answer value
```
### SQLObservation (server -> agent)
```python
class SQLObservation(Observation):
question: str # NL question to answer
schema_info: str # known schema (incrementally revealed)
result: str # last action result (truncated)
error: str # error message if action failed
step_count: int # current step number
budget_remaining: int # steps left
action_history: list[str] # summary of prior actions
# Inherited: done (bool), reward (float | None)
```
### SQLState (metadata endpoint)
```python
class SQLState(State):
history_messages: list[Message]
current_action_type: str
```
### Server-Only Types (never sent to agent)
```python
@dataclass
class QuestionRecord:
question_id: str
question_text: str
database_name: str
gold_sql: str
gold_answer: str
answer_type: str # integer | float | string | list
difficulty: str # easy | medium | hard
tables_involved: list[str]
@dataclass
class EpisodeContext:
episode_id: str
db_connection: sqlite3.Connection
question_record: QuestionRecord
step_count: int = 0
budget: int = 15
described_tables: set[str]
action_log: list[str]
done: bool = False
gold_answer: str | None
gold_rows: list[tuple]
# Reward accumulators
query_hashes: set[str]
best_progress: float = 0.0
cumulative_step_reward: float = 0.0
cumulative_new_info_reward: float = 0.0
```
**POMDP design:** The agent sees `SQLObservation`. The server holds `EpisodeContext`. The agent never sees gold answers, progress scores, or the full database. This separation forces exploration.
---
## API Contracts
### HTTP (OpenEnv Protocol)
The server exposes HTTP endpoints via `openenv.core.env_server.create_app()`.
| Operation | Method | Payload | Response |
|-----------|--------|---------|----------|
| Reset | `POST /reset` | `{seed: int}` (optional) | `SQLObservation` (JSON) |
| Step | `POST /step` | `{action_type, argument, metadata}` | `{observation, reward, done, info}` |
| State | `GET /state` | β | `SQLState` (JSON) |
### Evaluation API
```python
# Policy protocol
class Policy(Protocol):
def select_action(self, observation: SQLObservation) -> SQLAction: ...
# Built-in policies
RandomPolicy() # random baseline
OraclePolicy(questions) # gold-answer upper bound
# Runner
evaluate(env, policy, n_episodes, seed) -> EvaluationResult
# .success_rate, .avg_reward, .avg_steps, .episodes[]
```
### TRL Adapter API
```python
SQLEnvTRL.configure(questions_path, db_dir, step_budget) # class method
# Tool methods (auto-discovered by TRL):
SQLEnvTRL.describe(table_name: str) -> str
SQLEnvTRL.sample(table_name: str) -> str
SQLEnvTRL.query(sql: str) -> str
SQLEnvTRL.answer(value: str) -> str
```
---
## Cross-Cutting Concerns
### SQL Safety
All database access enforces:
- **Read-only** SQLite connections (`file:...?mode=ro`)
- **SELECT-only** β rejects INSERT, UPDATE, DELETE, ALTER, DROP
- **Single statement** β rejects `; ...` (no stacked queries)
- **5-second timeout** via SQLite progress handler
- **20-row truncation** on all result sets
### POMDP Structure
The partial observability is deliberate and load-bearing:
- Agent sees table names at reset but **not column details** (must DESCRIBE)
- Query results are **truncated** (at most 20 rows)
- Agent never sees `gold_answer`, `best_progress`, or `gold_rows`
- Step budget (default 15) forces strategic allocation of exploration
### Import Compatibility
Dual import paths throughout for local vs Docker execution:
```python
try:
from sql_env.models import SQLAction # local / pip install
except ImportError:
from models import SQLAction # Docker (PYTHONPATH=/app/env)
```
### Configuration
| Variable | Required | Default | Purpose |
|----------|----------|---------|---------|
| `QUESTIONS_PATH` | No | `data/questions/student_assessment.json` | Questions JSON |
| `DB_DIR` | No | `data/databases/` | SQLite database directory |
| `TOKENIZER_NAME` | No | `mistralai/Mistral-7B-Instruct-v0.1` | HuggingFace tokenizer |
| `PORT` | No | `8000` | Server port |
---
## Data, State, and Storage
### Committed Data
| Path | Contents |
|------|----------|
| `data/questions/questions_train.json` | 473 training questions across 10 DBs |
| `data/questions/questions_eval.json` | 203 evaluation questions across 10 DBs |
| `data/questions/db_list.json` | 10 Spider database IDs |
| `data/databases/models.py` | Legacy SQLAlchemy ORM models |
### Downloaded Data (gitignored)
Spider SQLite databases in `data/databases/{db_id}/{db_id}.sqlite`. Downloaded via `scripts/download_spider_databases.py`. The 10 databases: student_assessment, concert_singer, world_1, car_1, employee_hire_evaluation, pets_1, cre_Doc_Template_Mgt, dog_kennels, flight_2, poker_player.
### Runtime State (in-memory, per episode)
`EpisodeContext` holds all episode state: DB connection, gold data, reward accumulators, action history. Created on `reset()`, discarded when episode ends. Nothing persists between episodes.
---
<!-- ARCHITECTURE-SNAPSHOT-BEGIN -->
Snapshot (auto-managed)
- Repo signals: Python (pyproject.toml)
- Roots: tests/
- Entrypoint candidates: (none detected)
```text
tests/
e2e/
test_training_e2e.py
integration/
test_training_pipeline.py
unit/
test_error_handling.py
test_grpo_config.py
test_oracle_policy.py
test_prompts.py
test_reward.py
test_rewards.py
test_rollout.py
test_sft_terminal_message.py
test_trl_adapter.py
test_evaluation.py
test_smoke.py
test_synthetic.py
test_trl_adapter.py
test_verifier.py
test_verifier_integration.py
```
<!-- ARCHITECTURE-SNAPSHOT-END -->
---
## Infrastructure
### Development
**Prerequisites:** Python 3.11-3.12, `uv`, Docker (for deployment)
```bash
git clone <repo-url> && cd sql-env
uv sync
uv run python scripts/download_spider_databases.py
uv run pytest tests/ -v
```
### Deployment
**Target:** HuggingFace Spaces (Docker, free tier)
```bash
uv run openenv build # build Docker image
uv run openenv push # push to HF Spaces
```
The Dockerfile uses multi-stage build with `openenv-base`, runs as non-root `appuser`, bundles Spider databases, and exposes port 8000.
---
## Invariants
- Token tensors in `SQLState` grow monotonically across turns (never shrink mid-episode)
- `EpisodeContext` is server-only β leaking gold data to the agent breaks the POMDP
- Per-step rewards clipped to `[-0.05, 0.15]` β terminal reward (+1.0) always dominates exploration (~0.3 max)
- `tests/` must pass without GPU, without network, without downloaded databases (mocks/fixtures)
- SQL execution never mutates the database (read-only mode enforced at connection level)
---
## Glossary
| Term | Definition |
|------|------------|
| Episode | One question-answering session: reset -> N steps -> terminal |
| Action type | One of: DESCRIBE, SAMPLE, QUERY, ANSWER |
| POMDP | Partially observable MDP. Agent acts under uncertainty |
| Spider | Academic text-to-SQL benchmark dataset (10 DBs used) |
| OpenEnv | Meta's RL environment framework (Environment, EnvClient) |
| Green Agent | OpenEnv's evaluation wrapper pattern |
| Oracle policy | Baseline that uses gold SQL/answer for reward ceiling validation |
| TRL | HuggingFace Transformer Reinforcement Learning library |
| GRPO | Group Relative Policy Optimization (RL algorithm used for training) |
| Dense reward | Per-step reward signal (vs sparse terminal-only reward) |
---
## References
- OpenEnv framework: https://github.com/meta-pytorch/OpenEnv
- Spider dataset: https://huggingface.co/datasets/xlangai/spider
- TRL OpenEnv docs: https://huggingface.co/docs/trl/openenv
|