backend/app/services/submissions.py

"""Pure helpers to validate and build OpenEEGBench eval submissions."""
from __future__ import annotations

import re
from datetime import datetime, timezone
from typing import Any, Dict, List

from app.config.submission_config import (
    FINETUNING_STRATEGIES, PEFT_STRATEGIES, IA3_STRATEGY, HEADS,
    DATASET_KEYS, DEFAULT_STRATEGIES, DEFAULT_HEADS, DEFAULT_N_SEEDS,
    DEFAULT_HEAD_MODULE_NAME, MONTHLY_QUOTA, FRAMEWORK,
)

_DOTTED_RE = re.compile(r"^[A-Za-z_]\w*(\.[A-Za-z_]\w*)+$")
_EMAIL_RE = re.compile(r"^[^@\s]+@[^@\s]+\.[^@\s]+$")


def _as_list(value) -> list:
    if value is None:
        return []
    return value if isinstance(value, list) else [value]


def _str(payload, key) -> str:
    return (payload.get(key) or "").strip()


def validate_submission(payload: Dict[str, Any]) -> List[str]:
    """Return human-readable validation errors (empty list == valid)."""
    errors: List[str] = []

    model_cls = _str(payload, "model_cls")
    if not model_cls:
        errors.append("model_cls is required (dotted path, e.g. braindecode.models.BIOT).")
    elif not _DOTTED_RE.match(model_cls):
        errors.append(f"model_cls '{model_cls}' is not a valid dotted import path.")

    if not _str(payload, "model_name"):
        errors.append("model_name is required.")
    if not _str(payload, "organization"):
        errors.append("organization is required.")

    hub_repo = _str(payload, "hub_repo")
    checkpoint_url = _str(payload, "checkpoint_url")
    if not hub_repo and not checkpoint_url:
        errors.append("Provide a weights source: either hub_repo or checkpoint_url.")
    if hub_repo and checkpoint_url:
        errors.append("Provide only one weights source (hub_repo or checkpoint_url).")

    strategies = _as_list(payload.get("finetuning_strategies")) or DEFAULT_STRATEGIES
    bad = [s for s in strategies if s not in FINETUNING_STRATEGIES]
    if bad:
        errors.append(f"Unknown finetuning strategies: {bad}. Valid: {FINETUNING_STRATEGIES}.")

    heads = _as_list(payload.get("heads")) or DEFAULT_HEADS
    bad = [h for h in heads if h not in HEADS]
    if bad:
        errors.append(f"Unknown heads: {bad}. Valid: {HEADS}.")

    datasets = _as_list(payload.get("datasets")) or DATASET_KEYS
    bad = [d for d in datasets if d not in DATASET_KEYS]
    if bad:
        errors.append(f"Unknown datasets: {bad}. Valid: {DATASET_KEYS}.")

    if any(s in PEFT_STRATEGIES for s in strategies) and not _as_list(payload.get("peft_target_modules")):
        errors.append("peft_target_modules is required when a PEFT strategy (lora/ia3/adalora/dora/oft) is selected.")
    if IA3_STRATEGY in strategies and not _as_list(payload.get("peft_ff_modules")):
        errors.append("peft_ff_modules is required when the ia3 strategy is selected.")

    n_seeds = payload.get("n_seeds", DEFAULT_N_SEEDS)
    if not isinstance(n_seeds, int) or isinstance(n_seeds, bool) or n_seeds < 1:
        errors.append("n_seeds must be an integer >= 1.")

    email = _str(payload, "contact_email")
    if email and not _EMAIL_RE.match(email):
        errors.append("contact_email is not a valid email address.")

    if payload.get("visibility", "Public") not in ("Public", "Private"):
        errors.append("visibility must be 'Public' or 'Private'.")

    return errors


def build_request_entry(payload: Dict[str, Any], submitter: str, now: datetime | None = None) -> Dict[str, Any]:
    now = now or datetime.now(timezone.utc)
    strategies = _as_list(payload.get("finetuning_strategies")) or DEFAULT_STRATEGIES
    heads = _as_list(payload.get("heads")) or DEFAULT_HEADS
    datasets = _as_list(payload.get("datasets")) or DATASET_KEYS
    return {
        "schema_version": 1,
        "framework": FRAMEWORK,
        "model_name": _str(payload, "model_name"),
        "organization": _str(payload, "organization"),
        "submitter": submitter,
        "contact_email": _str(payload, "contact_email"),
        "model_url": _str(payload, "model_url"),
        "paper_url": _str(payload, "paper_url"),
        "additional_info": _str(payload, "additional_info"),
        "visibility": payload.get("visibility", "Public"),
        "benchmark_kwargs": {
            "model_cls": _str(payload, "model_cls"),
            "hub_repo": _str(payload, "hub_repo") or None,
            "checkpoint_url": _str(payload, "checkpoint_url") or None,
            "model_kwargs": payload.get("model_kwargs") or {},
            "head_module_name": payload.get("head_module_name") or DEFAULT_HEAD_MODULE_NAME,
            "peft_target_modules": _as_list(payload.get("peft_target_modules")),
            "peft_ff_modules": _as_list(payload.get("peft_ff_modules")),
            "finetuning_strategies": strategies,
            "heads": heads,
            "datasets": datasets,
            "n_seeds": int(payload.get("n_seeds", DEFAULT_N_SEEDS)),
        },
        "status": "PENDING",
        "submitted_time": now.strftime("%Y-%m-%dT%H:%M:%SZ"),
        "job_id": -1,
    }


def count_user_submissions_this_month(entries, submitter: str, now: datetime | None = None) -> int:
    now = now or datetime.now(timezone.utc)
    month = now.strftime("%Y-%m")
    return sum(
        1 for e in entries
        if e.get("submitter") == submitter and str(e.get("submitted_time", "")).startswith(month)
    )


def monthly_quota_exceeded(entries, submitter: str, now: datetime | None = None) -> bool:
    return count_user_submissions_this_month(entries, submitter, now) >= MONTHLY_QUOTA