fix typing

Embeddings integration

example.py

@@ -1,37 +0,0 @@
-import os
-
-import openai
-
-client = openai.OpenAI(
-    api_key=os.environ["CASHU_TOKEN"],
-    base_url=os.environ.get("ROUTSTR_API_URL", "https://api.routstr.com/v1"),
-    # base_url="http://roustrjfsdgfiueghsklchg.onion/v1",
-    # client=httpx.AsyncClient(
-    #     proxies={"http": "socks5://localhost:9050"},
-    # ),  # to use onion proxy (tor)
-)
-history: list = []
-
-
-def chat() -> None:
-    while True:
-        user_msg = {"role": "user", "content": input("\nYou: ")}
-        history.append(user_msg)
-        ai_msg = {"role": "assistant", "content": ""}
-
-        for chunk in client.chat.completions.create(
-            model=os.environ.get("MODEL", "openai/gpt-4o-mini"),
-            messages=history,
-            stream=True,
-        ):
-            if len(chunk.choices) > 0:
-                content = chunk.choices[0].delta.content
-                if content is not None:
-                    ai_msg["content"] += content
-                    print(content, end="", flush=True)
-        print()
-        history.append(ai_msg)
-
-
-if __name__ == "__main__":
-    chat()

examples/balance/check_balance.py

@@ -0,0 +1,11 @@
+import os
+
+import httpx
+
+# Use your Cashu token or API key as the Bearer token,
+# cashu token is hashed on the server and acts as an Temporary API key
+headers = {"Authorization": f"Bearer {os.environ.get('TOKEN')}"}
+base_url = os.environ.get("API_URL", "https://api.routstr.com/v1")
+
+resp = httpx.get(f"{base_url}/balance/info", headers=headers)
+print(resp.json())

examples/balance/create_balance.py

@@ -0,0 +1,15 @@
+import os
+
+import httpx
+
+# Send a Cashu token to the /create endpoint to get a persistent API key
+token = os.environ.get("TOKEN")
+if not token:
+    print("Please set TOKEN environment variable with a Cashu token")
+    exit(1)
+
+base_url = os.environ.get("API_URL", "https://api.routstr.com/v1")
+
+resp = httpx.get(f"{base_url}/balance/create", params={"initial_balance_token": token})
+
+print(resp.json())

examples/balance/refund_balance.py

@@ -0,0 +1,12 @@
+import os
+
+import httpx
+
+# Use your Cashu token or API key as the Bearer token
+headers = {"Authorization": f"Bearer {os.environ.get('TOKEN')}"}
+base_url = os.environ.get("API_URL", "https://api.routstr.com/v1")
+
+resp = httpx.post(f"{base_url}/balance/refund", headers=headers)
+
+print("Refund successful!")
+print(resp.json())

examples/balance/topup_balance.py

@@ -0,0 +1,16 @@
+import os
+
+import httpx
+
+# Use your Cashu token or API key as the Bearer token
+headers = {"Authorization": f"Bearer {os.environ.get('TOKEN')}"}
+base_url = os.environ.get("API_URL", "https://api.routstr.com/v1")
+
+# The Cashu token to top up with
+cashu_token = input("Enter Cashu token to top up: ")
+
+resp = httpx.post(
+    f"{base_url}/balance/topup", headers=headers, json={"cashu_token": cashu_token}
+)
+
+print(resp.json())

examples/chat_completions.py

@@ -0,0 +1,15 @@
+import os
+
+from openai import OpenAI
+
+client = OpenAI(
+    api_key=os.environ.get("TOKEN"),
+    base_url=os.environ.get("API_URL", "https://api.routstr.com/v1"),
+)
+
+response = client.chat.completions.create(
+    model=os.environ.get("MODEL", "gpt-5-nano"),
+    messages=[{"role": "user", "content": "Hello!"}],
+)
+
+print(response.choices[0].message.content)

examples/list_models.py

@@ -0,0 +1,19 @@
+import os
+
+import httpx
+from openai import OpenAI
+
+client = OpenAI(
+    api_key=os.environ.get("TOKEN", ""),
+    base_url=os.environ.get("API_URL", "https://api.routstr.com/v1"),
+)
+
+for model in client.models.list():
+    print(model.id)
+
+# OR
+
+models = httpx.get(
+    f"{client.base_url}/v1/models",
+    headers={"Authorization": f"Bearer {client.api_key}"},
+).json()

examples/responses/conversation.py

@@ -0,0 +1,31 @@
+import os
+
+from openai import OpenAI
+
+client = OpenAI(
+    api_key=os.environ.get("TOKEN"),
+    base_url=os.environ.get("API_URL", "https://api.routstr.com/v1"),
+)
+
+conversation = []  # type: ignore
+
+# First turn
+response1 = client.responses.create(  # type: ignore
+    model="o4-mini",
+    input="Hi, my name is Alice.",
+    conversation=conversation,
+)
+print("Response 1:", response1.output)
+
+# Note: The 'conversation' parameter might need to be constructed differently
+# depending on exact SDK/API spec. Typically, you pass back the previous turn's data.
+# Assuming the SDK manages or returns a conversation object/ID:
+# conversation.append(response1)
+
+# Second turn - demonstrating intent, actual implementation depends on strict API spec
+# response2 = client.responses.create(
+#     model="openai/gpt-4o-mini",
+#     input="What is my name?",
+#     conversation=conversation,
+# )
+# print("Response 2:", response2.output)

examples/responses/create.py

@@ -0,0 +1,17 @@
+import os
+
+from openai import OpenAI
+
+# The OpenAI SDK handles the 'responses' endpoint if it's updated to the latest version
+# and the base_url points to a compatible proxy like Routstr.
+client = OpenAI(
+    api_key=os.environ.get("TOKEN"),
+    base_url=os.environ.get("API_URL", "https://api.routstr.com/v1"),
+)
+
+response = client.responses.create(
+    model="gpt-5-mini",
+    input="Tell me a three sentence bedtime story about a unicorn.",
+)
+
+print(response.output)

examples/responses/streaming_response.py

@@ -0,0 +1,20 @@
+import os
+
+from openai import OpenAI
+
+client = OpenAI(
+    api_key=os.environ.get("TOKEN"),
+    base_url=os.environ.get("API_URL", "https://api.routstr.com/v1"),
+)
+
+stream = client.responses.create(
+    model="claude-4.5-sonnet",
+    input="Write a short poem about rust.",
+    stream=True,
+)
+
+for event in stream:
+    # Note: Depending on the SDK version and response structure,
+    # you might access event.output_delta or similar fields
+    print(event, end="", flush=True)
+print()

examples/responses/web_search.py

@@ -0,0 +1,16 @@
+import os
+
+from openai import OpenAI
+
+client = OpenAI(
+    api_key=os.environ.get("TOKEN"),
+    base_url=os.environ.get("API_URL", "https://api.routstr.com/v1"),
+)
+
+response = client.responses.create(
+    model="gpt-5-mini",
+    input="What is the latest news about AI?",
+    tools=[{"type": "web_search"}],  # type: ignore
+)
+
+print(response.output)

examples/streaming.py

@@ -0,0 +1,28 @@
+import os
+
+from openai import OpenAI
+
+client = OpenAI(
+    api_key=os.environ.get("TOKEN"),
+    base_url=os.environ.get("API_URL", "https://api.routstr.com/v1"),
+)
+
+messages = []
+while True:
+    messages.append({"role": "user", "content": input("\nYou: ")})
+
+    stream = client.chat.completions.create(
+        model=os.environ.get("MODEL", "gpt-5.1-mini"),
+        messages=messages,  # type: ignore
+        stream=True,
+    )
+
+    print("AI: ", end="")
+    response_content = ""
+    for chunk in stream:
+        if content := chunk.choices[0].delta.content:  # type: ignore
+            print(content, end="", flush=True)
+            response_content += content
+    print()
+
+    messages.append({"role": "assistant", "content": response_content})

examples/tor.py

@@ -0,0 +1,20 @@
+import os
+
+import httpx
+from openai import OpenAI
+
+# Requires `pip install "httpx[socks]"` and a running Tor proxy on port 9050
+client = OpenAI(
+    api_key=os.environ.get("TOKEN"),
+    base_url=os.environ.get("ONION_URL", "http://roustrjfsdgfiueghsklchg.onion/v1"),
+    http_client=httpx.Client(proxies="socks5://localhost:9050"),
+)
+
+print(
+    client.chat.completions.create(
+        model="openai/gpt-4o-mini",
+        messages=[{"role": "user", "content": "Hello from Tor!"}],
+    )
+    .choices[0]
+    .message.content
+)

pyproject.toml

@@ -73,6 +73,7 @@ packages = ["routstr"]
 [tool.ruff.lint]
 select = ["E", "F", "I"]
 ignore = ["E501"]
+exclude = ["examples"]
 
 [tool.mypy]
 python_version = "3.11"

routstr/payment/cost_calculation.py

@@ -5,6 +5,7 @@ from pydantic.v1 import BaseModel
 from ..core import get_logger
 from ..core.db import AsyncSession
 from ..core.settings import settings
+from .price import sats_usd_price
 
 logger = get_logger(__name__)
 
@@ -64,6 +65,56 @@ async def calculate_cost(  # todo: can be sync
         )
         return cost_data
 
+    usage_data = response_data["usage"]
+
+    usd_cost = 0.0
+
+    # Prioritize cost_details.upstream_inference_cost
+    if "cost_details" in usage_data:
+        usd_cost = float(
+            usage_data["cost_details"].get("upstream_inference_cost", 0) or 0
+        )
+
+    # Fallback to cost field if upstream_inference_cost is 0
+    if usd_cost == 0 and "cost" in usage_data:
+        try:
+            usd_cost = float(usage_data.get("cost", 0) or 0)
+        except Exception:
+            pass
+
+    if usd_cost > 0:
+        try:
+            sats_per_usd = 1.0 / sats_usd_price()
+            cost_in_sats = usd_cost * sats_per_usd
+            cost_in_msats = math.ceil(cost_in_sats * 1000)
+
+            logger.info(
+                "Using cost from usage data/details",
+                extra={
+                    "usd_cost": usd_cost,
+                    "cost_in_sats": cost_in_sats,
+                    "cost_in_msats": cost_in_msats,
+                    "model": response_data.get("model", "unknown"),
+                },
+            )
+
+            return CostData(
+                base_msats=-1,
+                input_msats=-1,  # Cost field doesn't break down by token type
+                output_msats=-1,
+                total_msats=cost_in_msats,
+            )
+        except Exception as e:
+            logger.warning(
+                "Error calculating cost from usage data",
+                extra={
+                    "error": str(e),
+                    "usd_cost": usd_cost,
+                    "model": response_data.get("model", "unknown"),
+                },
+            )
+            # Fall through to token-based calculation
+
     MSATS_PER_1K_INPUT_TOKENS: float = (
         float(settings.fixed_per_1k_input_tokens) * 1000.0
     )
@@ -129,10 +180,19 @@ async def calculate_cost(  # todo: can be sync
         )
         return cost_data
 
-    input_tokens = response_data.get("usage", {}).get("prompt_tokens", 0)
-    output_tokens = response_data.get("usage", {}).get("completion_tokens", 0)
+    input_tokens = usage_data.get("prompt_tokens", 0)
+    output_tokens = usage_data.get("completion_tokens", 0)
+
+    # added for response api
+    input_tokens = (
+        input_tokens if input_tokens != 0 else usage_data.get("input_tokens", 0)
+    )
+    output_tokens = (
+        output_tokens if output_tokens != 0 else usage_data.get("output_tokens", 0)
+    )
 
     input_msats = round(input_tokens / 1000 * MSATS_PER_1K_INPUT_TOKENS, 3)
+
     output_msats = round(output_tokens / 1000 * MSATS_PER_1K_OUTPUT_TOKENS, 3)
     token_based_cost = math.ceil(input_msats + output_msats)
 

tests/integration/test_wallet_authentication.py

@@ -3,7 +3,6 @@ Integration tests for wallet authentication system including API key generation
 Tests POST /v1/wallet/topup endpoint and authorization header validation.
 """
 
-import hashlib
 from datetime import datetime, timedelta
 from typing import Any
 
@@ -15,7 +14,6 @@ from routstr.core.db import ApiKey
 
 from .utils import (
     CashuTokenGenerator,
-    ConcurrencyTester,
     ResponseValidator,
 )
 
@@ -389,69 +387,6 @@ async def test_api_key_with_expiry_time(
     # The expiry time and refund address functionality is tested elsewhere
 
 
-@pytest.mark.integration
-@pytest.mark.asyncio
-async def test_concurrent_token_submissions(
-    integration_client: AsyncClient, testmint_wallet: Any, integration_session: Any
-) -> None:
-    """Test concurrent submissions of different tokens"""
-
-    # Generate multiple unique tokens with known amounts
-    num_tokens = 10
-    tokens = []
-    expected_balances = {}
-
-    for i in range(num_tokens):
-        amount = 100 + i * 10
-        token = await testmint_wallet.mint_tokens(amount)
-        tokens.append(token)
-        # Store expected balance by token hash
-        hashed_key = hashlib.sha256(token.encode()).hexdigest()
-        expected_balances[hashed_key] = amount * 1000  # msats
-
-    # Create concurrent requests
-    requests = [
-        {
-            "method": "GET",
-            "url": "/v1/wallet/info",
-            "headers": {"Authorization": f"Bearer {token}"},
-        }
-        for token in tokens
-    ]
-
-    # Execute concurrently
-    tester = ConcurrencyTester()
-    responses = await tester.run_concurrent_requests(
-        integration_client, requests, max_concurrent=5
-    )
-
-    # All should succeed
-    assert len(responses) == num_tokens
-    api_keys = set()
-
-    for response in responses:
-        assert response.status_code == 200
-        data = response.json()
-        api_key = data["api_key"]
-        api_keys.add(api_key)
-
-        # Verify balance matches the expected amount
-        hashed_key = api_key[3:]  # Remove "sk-" prefix
-        assert data["balance"] == expected_balances[hashed_key]
-
-    # Should have created unique API keys
-    assert len(api_keys) == num_tokens
-
-    # Verify all keys exist in database
-    for api_key in api_keys:
-        hashed_key = api_key[3:]  # Remove "sk-" prefix
-        result = await integration_session.execute(
-            select(ApiKey).where(ApiKey.hashed_key == hashed_key)  # type: ignore[arg-type]
-        )
-        db_key = result.scalar_one()
-        assert db_key.balance == expected_balances[hashed_key]
-
-
 @pytest.mark.integration
 @pytest.mark.asyncio
 async def test_authorization_with_cashu_token_directly(
@@ -504,48 +439,6 @@ async def test_x_cashu_header_support(
     assert response.status_code == 200
 
 
-@pytest.mark.integration
-@pytest.mark.asyncio
-@pytest.mark.slow
-async def test_api_key_consistency_under_load(
-    integration_client: AsyncClient, testmint_wallet: Any, integration_session: Any
-) -> None:
-    """Test API key generation consistency under concurrent load"""
-
-    # Generate a single token
-    token = await testmint_wallet.mint_tokens(1000)
-
-    # First request to create the API key
-    integration_client.headers["Authorization"] = f"Bearer {token}"
-    initial_response = await integration_client.get("/v1/wallet/info")
-    assert initial_response.status_code == 200
-    expected_api_key = initial_response.json()["api_key"]
-    expected_balance = initial_response.json()["balance"]
-
-    # Try to use the same token concurrently multiple times
-    # All should return the same API key since it's already created
-    requests = [
-        {
-            "method": "GET",
-            "url": "/v1/wallet/info",
-            "headers": {"Authorization": f"Bearer {token}"},
-        }
-        for _ in range(20)  # 20 concurrent attempts
-    ]
-
-    tester = ConcurrencyTester()
-    responses = await tester.run_concurrent_requests(
-        integration_client, requests, max_concurrent=10
-    )
-
-    # All should succeed and return the same API key
-    for response in responses:
-        assert response.status_code == 200
-        data = response.json()
-        assert data["api_key"] == expected_api_key
-        assert data["balance"] == expected_balance
-
-
 @pytest.mark.integration
 @pytest.mark.asyncio
 async def test_database_timestamp_accuracy(

tests/integration/test_wallet_information.py

@@ -13,7 +13,7 @@ from sqlmodel import select, update
 
 from routstr.core.db import ApiKey
 
-from .utils import ConcurrencyTester, ResponseValidator
+from .utils import ResponseValidator
 
 
 @pytest.mark.integration
@@ -204,45 +204,6 @@ async def test_expired_api_key_behavior(
     assert db_key.refund_address == "test@lightning.address"
 
 
-@pytest.mark.integration
-@pytest.mark.asyncio
-async def test_concurrent_access_same_api_key(
-    integration_client: AsyncClient, authenticated_client: AsyncClient
-) -> None:
-    """Test concurrent access with the same API key"""
-
-    # Get the API key from authenticated client
-    response = await authenticated_client.get("/v1/wallet/")
-    api_key = response.json()["api_key"]
-    initial_balance = response.json()["balance"]
-
-    # Create multiple concurrent requests
-    requests = []
-    for i in range(20):
-        # Alternate between both endpoints
-        endpoint = "/v1/wallet/" if i % 2 == 0 else "/v1/wallet/info"
-        requests.append(
-            {
-                "method": "GET",
-                "url": endpoint,
-                "headers": {"Authorization": f"Bearer {api_key}"},
-            }
-        )
-
-    # Execute concurrently
-    tester = ConcurrencyTester()
-    responses = await tester.run_concurrent_requests(
-        integration_client, requests, max_concurrent=10
-    )
-
-    # All should succeed with consistent data
-    for response in responses:
-        assert response.status_code == 200
-        data = response.json()
-        assert data["api_key"] == api_key
-        assert data["balance"] == initial_balance
-
-
 @pytest.mark.integration
 @pytest.mark.asyncio
 async def test_wallet_info_data_consistency(

tests/integration/test_wallet_topup.py

@@ -15,7 +15,6 @@ from routstr.core.db import ApiKey
 
 from .utils import (
     CashuTokenGenerator,
-    ConcurrencyTester,
     ResponseValidator,
 )
 
@@ -284,60 +283,6 @@ async def test_transaction_history_tracking(  # type: ignore[no-untyped-def]
     assert response.status_code == 400
 
 
-@pytest.mark.integration
-@pytest.mark.asyncio
-async def test_concurrent_topups_same_api_key(  # type: ignore[no-untyped-def]
-    integration_client: AsyncClient,
-    authenticated_client: AsyncClient,
-    testmint_wallet: Any,
-) -> None:
-    """Test concurrent top-ups to the same API key"""
-
-    # Get API key
-    response = await authenticated_client.get("/v1/wallet/")
-    api_key = response.json()["api_key"]
-    initial_balance = response.json()["balance"]
-
-    # Generate multiple unique tokens
-    num_tokens = 10
-    tokens = []
-    total_amount = 0
-
-    for i in range(num_tokens):
-        amount = 100 + i * 10  # Different amounts
-        token = await testmint_wallet.mint_tokens(amount)
-        tokens.append(token)
-        total_amount += amount
-
-    # Create concurrent top-up requests
-    requests = [
-        {
-            "method": "POST",
-            "url": "/v1/wallet/topup",
-            "params": {"cashu_token": token},
-            "headers": {"Authorization": f"Bearer {api_key}"},
-        }
-        for token in tokens
-    ]
-
-    # Execute concurrently
-    tester = ConcurrencyTester()
-    responses = await tester.run_concurrent_requests(
-        integration_client, requests, max_concurrent=5
-    )
-
-    # All should succeed
-    for response in responses:
-        assert response.status_code == 200
-        assert "msats" in response.json()
-
-    # Verify final balance is correct
-    final_response = await authenticated_client.get("/v1/wallet/")
-    final_balance = final_response.json()["balance"]
-    expected_balance = initial_balance + (total_amount * 1000)
-    assert final_balance == expected_balance
-
-
 @pytest.mark.integration
 @pytest.mark.asyncio
 async def test_topup_during_active_proxy_request(  # type: ignore[no-untyped-def]