Files
routstr-core/routstr/models/models.py
T
2025-11-28 13:40:30 +09:00

483 lines
15 KiB
Python

import asyncio
import json
import random
from pathlib import Path
from fastapi import APIRouter, Depends
from pydantic.v1 import BaseModel
from sqlmodel import select
from sqlmodel.ext.asyncio.session import AsyncSession
from ..core.db import ModelRow, create_session, get_session
from ..core.logging import get_logger
from ..core.settings import settings
from ..payment.price import sats_usd_price
from .metadata import async_fetch_openrouter_models
logger = get_logger(__name__)
models_router = APIRouter()
class Architecture(BaseModel):
modality: str
input_modalities: list[str]
output_modalities: list[str]
tokenizer: str
instruct_type: str | None
class Pricing(BaseModel):
prompt: float
completion: float
request: float
image: float
web_search: float
internal_reasoning: float
max_prompt_cost: float = 0.0 # in sats not msats
max_completion_cost: float = 0.0 # in sats not msats
max_cost: float = 0.0 # in sats not msats
class TopProvider(BaseModel):
context_length: int | None = None
max_completion_tokens: int | None = None
is_moderated: bool | None = None
class Model(BaseModel):
id: str
name: str
created: int
description: str
context_length: int
architecture: Architecture
pricing: Pricing
sats_pricing: Pricing | None = None
per_request_limits: dict | None = None
top_provider: TopProvider | None = None
enabled: bool = True
upstream_provider_id: int | None = None
canonical_slug: str | None = None
alias_ids: list[str] | None = None
def __hash__(self) -> int:
return hash(self.id)
def is_openrouter_upstream() -> bool:
try:
base = (settings.upstream_base_url or "").strip().rstrip("/")
except Exception:
return False
return base.lower() == "https://openrouter.ai/api/v1"
def _calculate_usd_max_costs(model: Model) -> tuple[float, float, float]:
"""Calculate max costs in USD based on model context/token limits.
Args:
model: Model object
Returns:
Tuple of (max_prompt_cost, max_completion_cost, max_cost) in USD
"""
min_req_msat = max(1, int(getattr(settings, "min_request_msat", 1)))
min_req_usd = float(min_req_msat) / 1_000_000.0
prompt_price = model.pricing.prompt
completion_price = model.pricing.completion
if model.top_provider and (
model.top_provider.context_length or model.top_provider.max_completion_tokens
):
if (cl := model.top_provider.context_length) and (
mct := model.top_provider.max_completion_tokens
):
if cl <= mct:
return (
cl * prompt_price,
cl * completion_price,
cl * max(completion_price, prompt_price),
)
return (
cl * prompt_price,
mct * completion_price,
(cl - mct) * prompt_price + mct * completion_price,
)
elif cl := model.top_provider.context_length:
return (
cl * prompt_price,
cl * completion_price,
cl * max(completion_price, prompt_price),
)
elif mct := model.top_provider.max_completion_tokens:
return (
mct * prompt_price,
mct * completion_price,
mct * completion_price,
)
elif model.context_length:
return (
model.context_length * prompt_price,
model.context_length * completion_price,
model.context_length * max(completion_price, prompt_price),
)
p = prompt_price * 1_000_000
c = completion_price * 32_000
r = model.pricing.request * 100_000
i = model.pricing.image * 100
w = model.pricing.web_search * 1000
ir = model.pricing.internal_reasoning * 100
return (p, c, max(p + c + r + i + w + ir, min_req_usd))
def _update_model_sats_pricing(model: Model, sats_to_usd: float) -> Model:
"""Update a model's sats_pricing based on USD pricing and exchange rate.
Args:
model: Model object to update
sats_to_usd: Current sats to USD exchange rate
Returns:
Updated Model object with new sats_pricing
"""
try:
min_req_msat = max(1, int(getattr(settings, "min_request_msat", 1)))
min_req_sats = float(min_req_msat) / 1000.0
sats = Pricing.parse_obj(
{k: v / sats_to_usd for k, v in model.pricing.dict().items()}
)
if sats.request <= 0.0:
sats.request = min_req_sats
if (sats.max_cost or 0.0) < min_req_sats:
sats.max_cost = min_req_sats
return Model(
id=model.id,
name=model.name,
created=model.created,
description=model.description,
context_length=model.context_length,
architecture=model.architecture,
pricing=model.pricing,
sats_pricing=sats,
per_request_limits=model.per_request_limits,
top_provider=model.top_provider,
enabled=model.enabled,
upstream_provider_id=model.upstream_provider_id,
canonical_slug=model.canonical_slug,
alias_ids=model.alias_ids,
)
except Exception as e:
logger.error(
"Failed to update sats pricing for model",
extra={
"model_id": model.id,
"error": str(e),
"error_type": type(e).__name__,
},
)
return model
async def ensure_models_bootstrapped() -> None:
from .crud import _model_to_row_payload
async with create_session() as s:
existing = (await s.exec(select(ModelRow.id).limit(1))).all() # type: ignore
if existing:
return
try:
models_path = Path(settings.models_path)
except Exception:
models_path = Path("models.json")
models_to_insert: list[dict] = []
if models_path.exists():
try:
with models_path.open("r") as f:
data = json.load(f)
models_to_insert = data.get("models", [])
logger.info(
f"Bootstrapping {len(models_to_insert)} models from {models_path}"
)
except Exception as e:
logger.error(f"Error loading models from {models_path}: {e}")
if not models_to_insert and is_openrouter_upstream():
logger.info("Bootstrapping models from OpenRouter API")
source_filter = None
try:
src = settings.source or None
source_filter = src if src and src.strip() else None
except Exception:
pass
models_to_insert = await async_fetch_openrouter_models(
source_filter=source_filter
)
elif not models_to_insert:
logger.info(
"No models.json found and upstream is not OpenRouter; skipping bootstrap"
)
for m in models_to_insert:
try:
model = Model(**m) # type: ignore
except Exception:
# Some OpenRouter models include extra fields; only map required ones
continue
exists = await s.get(ModelRow, model.id)
if exists:
continue
payload = _model_to_row_payload(model)
s.add(ModelRow(**payload)) # type: ignore
await s.commit()
async def _update_sats_pricing_once() -> None:
"""Update sats pricing once for all provider models (in-memory only)."""
from ..proxy import get_upstreams
upstreams = get_upstreams()
sats_to_usd = sats_usd_price()
updated_count = 0
for upstream in upstreams:
updated_models = [
_update_model_sats_pricing(m, sats_to_usd)
for m in upstream.get_cached_models()
]
upstream._models_cache = updated_models
upstream._models_by_id = {m.id: m for m in updated_models}
updated_count += len(updated_models)
if updated_count > 0:
logger.info("Updated sats pricing", extra={"models_updated": updated_count})
async def update_sats_pricing() -> None:
"""Periodically update sats pricing for all provider models and database overrides."""
try:
if not settings.enable_pricing_refresh:
return
except Exception:
pass
await _update_sats_pricing_once()
while True:
try:
interval = getattr(settings, "pricing_refresh_interval_seconds", 120)
jitter = max(0.0, float(interval) * 0.1)
await asyncio.sleep(interval + random.uniform(0, jitter))
except asyncio.CancelledError:
break
try:
try:
if not settings.enable_pricing_refresh:
return
except Exception:
pass
await _update_sats_pricing_once()
except asyncio.CancelledError:
break
except Exception as e:
logger.error(f"Error updating sats pricing: {e}")
async def cleanup_enabled_models_periodically() -> None:
"""Background task to clean up enabled models that match upstream pricing.
When model is enabled (enabled=True), remove it from DB if it matches upstream pricing.
Keep it in DB only if pricing differs from upstream or if it's disabled.
"""
interval = getattr(
settings, "models_cleanup_interval_seconds", 300
) # 5 minutes default
if not interval or interval <= 0:
return
while True:
try:
await _cleanup_enabled_models_once()
except asyncio.CancelledError:
break
except Exception as e:
logger.error(
"Error during enabled models cleanup",
extra={"error": str(e), "error_type": type(e).__name__},
)
try:
jitter = max(0.0, float(interval) * 0.1)
await asyncio.sleep(interval + random.uniform(0, jitter))
except asyncio.CancelledError:
break
async def _cleanup_enabled_models_once() -> None:
"""Clean up enabled models that match upstream pricing."""
from ..proxy import get_upstreams
async with create_session() as session:
# Get all enabled models from DB
result = await session.exec(
select(ModelRow).where(
ModelRow.enabled, # Only enabled models
)
)
db_models = result.all()
if not db_models:
return
upstreams = get_upstreams()
models_to_remove = []
for db_model in db_models:
# Find corresponding upstream model
upstream_model = None
for upstream in upstreams:
upstream_model = upstream.get_cached_model_by_id(db_model.id)
if upstream_model:
break
if not upstream_model:
continue
# Compare pricing to see if they match
db_pricing = json.loads(db_model.pricing)
upstream_pricing = upstream_model.pricing.dict()
# Check if pricing matches (with small tolerance for float comparison)
pricing_matches = _pricing_matches(db_pricing, upstream_pricing)
if pricing_matches:
models_to_remove.append(db_model)
logger.info(
f"Removing enabled model {db_model.id} - matches upstream pricing",
extra={"model_id": db_model.id},
)
# Remove models that match upstream pricing
for model in models_to_remove:
await session.delete(model)
if models_to_remove:
await session.commit()
logger.info(
f"Cleaned up {len(models_to_remove)} enabled models that match upstream pricing"
)
def _pricing_matches(
db_pricing: dict, upstream_pricing: dict, tolerance: float = 0.0
) -> bool:
"""Check if pricing dictionaries match within tolerance."""
keys_to_compare = [
"prompt",
"completion",
"request",
"image",
"web_search",
"internal_reasoning",
]
for key in keys_to_compare:
db_val = int(float(db_pricing.get(key, 0.0)) * 1000000)
upstream_val = int(float(upstream_pricing.get(key, 0.0)) * 1000000)
if abs(db_val - upstream_val) > tolerance:
return False
return True
async def refresh_models_periodically() -> None:
"""Background task: periodically fetch OpenRouter models and insert new ones.
- Respects optional SOURCE filter from settings
- Does not overwrite existing rows
- Sleeps according to settings.models_refresh_interval_seconds; disabled when 0
"""
from .crud import _model_to_row_payload
interval = getattr(settings, "models_refresh_interval_seconds", 0)
if not interval or interval <= 0:
return
# Only refresh from OpenRouter when upstream is OpenRouter
if not is_openrouter_upstream():
logger.info("Skipping models refresh: upstream_base_url is not OpenRouter")
return
while True:
try:
try:
if not settings.enable_models_refresh:
return
except Exception:
pass
try:
src = settings.source or None
source_filter = src if src and src.strip() else None
except Exception:
source_filter = None
models = await async_fetch_openrouter_models(source_filter=source_filter)
if not models:
await asyncio.sleep(interval)
continue
async with create_session() as s:
result = await s.exec(select(ModelRow.id)) # type: ignore
existing_ids = {
row[0] if isinstance(row, tuple) else row for row in result.all()
}
inserted = 0
for m in models:
try:
model = Model(**m) # type: ignore
except Exception:
continue
if model.id in existing_ids:
continue
payload = _model_to_row_payload(model)
try:
s.add(ModelRow(**payload)) # type: ignore
except Exception:
pass
inserted += 1
if inserted:
await s.commit()
logger.info(f"Inserted {inserted} new models from OpenRouter")
except asyncio.CancelledError:
break
except Exception as e:
logger.error(
"Error during models refresh",
extra={"error": str(e), "error_type": type(e).__name__},
)
try:
jitter = max(0.0, float(interval) * 0.1)
await asyncio.sleep(interval + random.uniform(0, jitter))
except asyncio.CancelledError:
break
@models_router.get("/v1/models")
@models_router.get("/models", include_in_schema=False)
async def models(session: AsyncSession = Depends(get_session)) -> dict:
"""Get all available models from all providers with database overrides applied."""
from ..proxy import get_unique_models
items = get_unique_models()
return {"data": items}