#274 do not change user for empty response by upstream

routstr/core/settings.py

@@ -6,7 +6,7 @@ import os
 from datetime import datetime, timezone
 from typing import Any
 
-from pydantic.v1 import BaseModel, BaseSettings, Field
+from pydantic.v1 import BaseModel, BaseSettings, Field, validator
 from sqlmodel.ext.asyncio.session import AsyncSession
 
 
@@ -37,6 +37,13 @@ class Settings(BaseSettings):
 
     # Cashu
     cashu_mints: list[str] = Field(default_factory=list, env="CASHU_MINTS")
+
+    @validator("cashu_mints", pre=True, each_item=True)
+    def normalize_mint_url(cls, v: str) -> str:
+        if isinstance(v, str):
+            return v.rstrip("/")
+        return v
+
     receive_ln_address: str = Field(default="", env="RECEIVE_LN_ADDRESS")
     primary_mint: str = Field(default="", env="PRIMARY_MINT_URL")
     primary_mint_unit: str = Field(default="sat", env="PRIMARY_MINT_UNIT")

routstr/payment/cost_calculation.py

@@ -48,13 +48,6 @@ async def calculate_cost(  # todo: can be sync
         },
     )
 
-    cost_data = MaxCostData(
-        base_msats=max_cost,
-        input_msats=0,
-        output_msats=0,
-        total_msats=max_cost,
-    )
-
     if "usage" not in response_data or response_data["usage"] is None:
         logger.warning(
             "No usage data in response, using base cost only",
@@ -63,7 +56,12 @@ async def calculate_cost(  # todo: can be sync
                 "model": response_data.get("model", "unknown"),
             },
         )
-        return cost_data
+        return MaxCostData(
+            base_msats=0,
+            input_msats=0,
+            output_msats=0,
+            total_msats=0,
+        )
 
     usage_data = response_data["usage"]
 
@@ -178,7 +176,12 @@ async def calculate_cost(  # todo: can be sync
                 "model": response_data.get("model", "unknown"),
             },
         )
-        return cost_data
+        return MaxCostData(
+            base_msats=max_cost,
+            input_msats=0,
+            output_msats=0,
+            total_msats=max_cost,
+        )
 
     input_tokens = usage_data.get("prompt_tokens", 0)
     output_tokens = usage_data.get("completion_tokens", 0)

routstr/proxy.py

@@ -16,6 +16,7 @@ from .core.db import (
     create_session,
     get_session,
 )
+from .core.settings import settings
 from .payment.helpers import (
     calculate_discounted_max_cost,
     check_token_balance,
@@ -25,6 +26,7 @@ from .payment.helpers import (
 from .payment.models import Model
 from .upstream import BaseUpstreamProvider
 from .upstream.helpers import init_upstreams
+from .wallet import deserialize_token_from_string
 
 logger = get_logger(__name__)
 proxy_router = APIRouter()
@@ -146,6 +148,24 @@ async def proxy(
     else:
         model_id = request_body_dict.get("model", "unknown")
 
+    if "https://testnut.cashu.space" in settings.cashu_mints:
+        try:
+            token_str = None
+            if x_cashu_header := headers.get("x-cashu"):
+                token_str = x_cashu_header
+            elif auth_header := headers.get("authorization"):
+                parts = auth_header.split(" ")
+                if len(parts) > 1 and not parts[1].startswith("sk-"):
+                    token_str = parts[1]
+
+            if token_str:
+                token_obj = deserialize_token_from_string(token_str)
+                if token_obj.mint == "https://testnut.cashu.space":
+                    model_id = "mock/gpt-420-mock"
+                    request_body_dict["model"] = model_id
+        except Exception:
+            pass
+
     model_obj = get_model_instance(model_id)
     if not model_obj:
         return create_error_response(

routstr/upstream/fake.py

@@ -0,0 +1,265 @@
+import asyncio
+import json
+import random
+from typing import AsyncIterator
+
+from fastapi import Request
+from fastapi.responses import Response, StreamingResponse
+
+from ..core.db import ApiKey, AsyncSession
+from ..payment.models import Architecture, Model, Pricing
+from .base import BaseUpstreamProvider
+
+
+class MockUpstreamProvider(BaseUpstreamProvider):
+    """Fack Mock Upstream provider specifically for Testing."""
+
+    provider_type = "mock"
+
+    async def forward_request(
+        self,
+        request: Request,
+        path: str,
+        headers: dict,
+        request_body: bytes | None,
+        key: ApiKey,
+        max_cost_for_model: int,
+        session: AsyncSession,
+        model_obj: Model,
+    ) -> Response | StreamingResponse:
+        if path.endswith("chat/completions"):
+            is_streaming = False
+            if request_body:
+                request_data = json.loads(request_body)
+                is_streaming = request_data.get("stream", False)
+
+            if is_streaming:
+
+                async def fake_streaming_response(
+                    chunk_size: int | None = None,
+                ) -> AsyncIterator[bytes]:
+                    suffix = random.randint(1000, 9999)
+                    req_id = f"gen-mock-stream-{suffix}"
+                    created = 1766138895
+                    model = "mock/gpt-420-mock"
+
+                    def make_chunk(
+                        delta: dict,
+                        finish_reason: str | None = None,
+                        usage: dict | None = None,
+                    ) -> bytes:
+                        chunk = {
+                            "id": req_id,
+                            "provider": "MockProvider",
+                            "model": model,
+                            "object": "chat.completion.chunk",
+                            "created": created,
+                            "choices": [
+                                {
+                                    "index": 0,
+                                    "delta": delta,
+                                    "finish_reason": finish_reason,
+                                    "native_finish_reason": "completed"
+                                    if finish_reason
+                                    else None,
+                                    "logprobs": None,
+                                }
+                            ],
+                        }
+                        if usage:
+                            chunk["usage"] = usage
+                        return f"data: {json.dumps(chunk)}\n\n".encode()
+
+                    # 1. Initial chunk
+                    yield make_chunk({"role": "assistant", "content": ""})
+                    await asyncio.sleep(0.02)
+
+                    # 2. Reasoning chunks
+                    reasoning_tokens = ["Mock", " reason", "ing", "..."]
+                    for token in reasoning_tokens:
+                        delta = {
+                            "role": "assistant",
+                            "content": "",
+                            "reasoning": token,
+                            "reasoning_details": [
+                                {
+                                    "type": "reasoning.summary",
+                                    "summary": token,
+                                    "format": "openai-responses-v1",
+                                    "index": 0,
+                                }
+                            ],
+                        }
+                        yield make_chunk(delta)
+                        await asyncio.sleep(0.03)
+
+                    # 3. Content chunks
+                    content_tokens = ["This", " is", " a", " mock", " stream", "."]
+                    for token in content_tokens:
+                        yield make_chunk({"role": "assistant", "content": token})
+                        await asyncio.sleep(0.03)
+
+                    # 4. Finish chunk
+                    yield make_chunk(
+                        {"role": "assistant", "content": ""}, finish_reason="stop"
+                    )
+
+                    # 5. Usage chunk
+                    usage_data = {
+                        "prompt_tokens": 10,
+                        "completion_tokens": 20,
+                        "total_tokens": 30,
+                        "cost": 0.001,
+                        "is_byok": False,
+                        "prompt_tokens_details": {
+                            "cached_tokens": 0,
+                            "audio_tokens": 0,
+                            "video_tokens": 0,
+                        },
+                        "cost_details": {
+                            "upstream_inference_cost": None,
+                            "upstream_inference_prompt_cost": 0,
+                            "upstream_inference_completions_cost": 0.001,
+                        },
+                        "completion_tokens_details": {
+                            "reasoning_tokens": 10,
+                            "image_tokens": 0,
+                        },
+                    }
+
+                    usage_chunk = {
+                        "id": req_id,
+                        "provider": "MockProvider",
+                        "model": model,
+                        "object": "chat.completion.chunk",
+                        "created": created,
+                        "choices": [
+                            {
+                                "index": 0,
+                                "delta": {"role": "assistant", "content": ""},
+                                "finish_reason": None,
+                                "native_finish_reason": None,
+                                "logprobs": None,
+                            }
+                        ],
+                        "usage": usage_data,
+                    }
+                    yield f"data: {json.dumps(usage_chunk)}\n\n".encode()
+
+                    # 6. DONE
+                    yield b"data: [DONE]\n\n"
+
+                    # 7. Cost
+                    cost_chunk = {
+                        "cost": {
+                            "base_msats": 0,
+                            "input_msats": 2,
+                            "output_msats": 10,
+                            "total_msats": 12,
+                        }
+                    }
+                    yield f"data: {json.dumps(cost_chunk)}\n\n".encode()
+
+                return StreamingResponse(
+                    fake_streaming_response(),
+                    200,
+                )
+
+            else:
+                suffix = random.randint(1000, 9999)
+                content_dict = {
+                    "id": f"gen-mock-{suffix}",
+                    "provider": "MockProvider",
+                    "model": "mock/gpt-5-mini",
+                    "object": "chat.completion",
+                    "created": 1766138655,
+                    "choices": [
+                        {
+                            "logprobs": None,
+                            "finish_reason": "length",
+                            "native_finish_reason": "max_output_tokens",
+                            "index": 0,
+                            "message": {
+                                "role": "assistant",
+                                "content": f"Mock Content {suffix}",
+                                "refusal": None,
+                                "reasoning": f"Mock Reasoning {suffix}",
+                                "reasoning_details": [
+                                    {
+                                        "format": "openai-responses-v1",
+                                        "index": 0,
+                                        "type": "reasoning.summary",
+                                        "summary": f"Mock Summary {suffix}",
+                                    },
+                                    {
+                                        "id": f"rs_mock_{suffix}",
+                                        "format": "openai-responses-v1",
+                                        "index": 0,
+                                        "type": "reasoning.encrypted",
+                                        "data": "mock_encrypted_data",
+                                    },
+                                ],
+                            },
+                        }
+                    ],
+                    "usage": {
+                        "prompt_tokens": 10,
+                        "completion_tokens": 10,
+                        "total_tokens": 20,
+                        "cost": 0,
+                        "is_byok": False,
+                        "prompt_tokens_details": {
+                            "cached_tokens": 0,
+                            "audio_tokens": 0,
+                            "video_tokens": 0,
+                        },
+                        "cost_details": {
+                            "upstream_inference_cost": None,
+                            "upstream_inference_prompt_cost": 0,
+                            "upstream_inference_completions_cost": 0,
+                        },
+                        "completion_tokens_details": {
+                            "reasoning_tokens": 5,
+                            "image_tokens": 0,
+                        },
+                    },
+                    "cost": {
+                        "base_msats": 0,
+                        "input_msats": 0,
+                        "output_msats": 0,
+                        "total_msats": 0,
+                    },
+                }
+                return Response(json.dumps(content_dict).encode(), 200)
+
+        elif path.endswith("embeddings"):
+            raise NotImplementedError
+        elif path.endswith("responses"):
+            raise NotImplementedError
+        else:
+            raise NotImplementedError
+
+    async def fetch_models(self) -> list[Model]:
+        return [
+            Model(
+                id="mock/gpt-420-mock",
+                name="mock/gpt-420-mock",
+                created=0,
+                description="mock model for testing",
+                context_length=8192,
+                architecture=Architecture(
+                    modality="text",
+                    input_modalities=["text"],
+                    output_modalities=["text"],
+                    tokenizer="",
+                    instruct_type=None,
+                ),
+                pricing=Pricing(prompt=0.01, completion=0.01),
+            ),
+        ]
+
+    def transform_model_name(self, model_id: str) -> str:
+        return "fake-model"
+
+    async def get_balance(self) -> float | None:
+        return 420.69

routstr/upstream/helpers.py

@@ -218,6 +218,14 @@ async def init_upstreams() -> list[BaseUpstreamProvider]:
         results = await asyncio.gather(*tasks)
         upstreams = [p for p in results if p is not None]
 
+        if "https://testnut.cashu.space" in settings.cashu_mints:
+            from .fake import MockUpstreamProvider
+
+            mock_provider = MockUpstreamProvider("mock", "mock")
+            await mock_provider.refresh_models_cache()
+            upstreams.append(mock_provider)
+            logger.info("Initialized MockUpstreamProvider for testnut mint")
+
         return upstreams
 
 

routstr/wallet.py

@@ -313,6 +313,8 @@ async def periodic_payout() -> None:
         try:
             async with db.create_session() as session:
                 for mint_url in settings.cashu_mints:
+                    if mint_url == "https://testnut.cashu.space":
+                        continue
                     for unit in ["sat", "msat"]:
                         wallet = await get_wallet(mint_url, unit)
                         proofs = get_proofs_per_mint_and_unit(