prompture 0.0.40.dev1__py3-none-any.whl → 0.0.42__py3-none-any.whl

prompture/drivers/async_moonshot_driver.py

@@ -0,0 +1,312 @@
+"""Async Moonshot AI (Kimi) driver using httpx.
+
+All pricing comes from models.dev (provider: "moonshotai") — no hardcoded pricing.
+
+Moonshot-specific constraints:
+- Temperature clamped to [0, 1] (OpenAI allows [0, 2])
+- tool_choice: "required" not supported — only "auto" or "none"
+"""
+
+from __future__ import annotations
+
+import json
+import os
+from collections.abc import AsyncIterator
+from typing import Any
+
+import httpx
+
+from ..async_driver import AsyncDriver
+from ..cost_mixin import CostMixin, prepare_strict_schema
+from .moonshot_driver import MoonshotDriver
+
+
+class AsyncMoonshotDriver(CostMixin, AsyncDriver):
+    supports_json_mode = True
+    supports_json_schema = True
+    supports_tool_use = True
+    supports_streaming = True
+    supports_vision = True
+
+    MODEL_PRICING = MoonshotDriver.MODEL_PRICING
+
+    def __init__(
+        self,
+        api_key: str | None = None,
+        model: str = "kimi-k2-0905-preview",
+        endpoint: str = "https://api.moonshot.ai/v1",
+    ):
+        self.api_key = api_key or os.getenv("MOONSHOT_API_KEY")
+        if not self.api_key:
+            raise ValueError("Moonshot API key not found. Set MOONSHOT_API_KEY env var.")
+        self.model = model
+        self.base_url = endpoint.rstrip("/")
+        self.headers = {
+            "Authorization": f"Bearer {self.api_key}",
+            "Content-Type": "application/json",
+        }
+
+    supports_messages = True
+
+    def _prepare_messages(self, messages: list[dict[str, Any]]) -> list[dict[str, Any]]:
+        from .vision_helpers import _prepare_openai_vision_messages
+
+        return _prepare_openai_vision_messages(messages)
+
+    async def generate(self, prompt: str, options: dict[str, Any]) -> dict[str, Any]:
+        messages = [{"role": "user", "content": prompt}]
+        return await self._do_generate(messages, options)
+
+    async def generate_messages(self, messages: list[dict[str, str]], options: dict[str, Any]) -> dict[str, Any]:
+        return await self._do_generate(self._prepare_messages(messages), options)
+
+    async def _do_generate(self, messages: list[dict[str, str]], options: dict[str, Any]) -> dict[str, Any]:
+        model = options.get("model", self.model)
+
+        model_config = self._get_model_config("moonshot", model)
+        tokens_param = model_config["tokens_param"]
+        supports_temperature = model_config["supports_temperature"]
+
+        self._validate_model_capabilities(
+            "moonshot",
+            model,
+            using_json_schema=bool(options.get("json_schema")),
+        )
+
+        opts = {"temperature": 1.0, "max_tokens": 512, **options}
+        MoonshotDriver._clamp_temperature(opts)
+
+        data: dict[str, Any] = {
+            "model": model,
+            "messages": messages,
+        }
+        data[tokens_param] = opts.get("max_tokens", 512)
+
+        if supports_temperature and "temperature" in opts:
+            data["temperature"] = opts["temperature"]
+
+        if options.get("json_mode"):
+            json_schema = options.get("json_schema")
+            if json_schema:
+                schema_copy = prepare_strict_schema(json_schema)
+                data["response_format"] = {
+                    "type": "json_schema",
+                    "json_schema": {
+                        "name": "extraction",
+                        "strict": True,
+                        "schema": schema_copy,
+                    },
+                }
+            else:
+                data["response_format"] = {"type": "json_object"}
+
+        async with httpx.AsyncClient() as client:
+            try:
+                response = await client.post(
+                    f"{self.base_url}/chat/completions",
+                    headers=self.headers,
+                    json=data,
+                    timeout=120,
+                )
+                response.raise_for_status()
+                resp = response.json()
+            except httpx.HTTPStatusError as e:
+                error_msg = f"Moonshot API request failed: {e!s}"
+                raise RuntimeError(error_msg) from e
+            except Exception as e:
+                raise RuntimeError(f"Moonshot API request failed: {e!s}") from e
+
+        usage = resp.get("usage", {})
+        prompt_tokens = usage.get("prompt_tokens", 0)
+        completion_tokens = usage.get("completion_tokens", 0)
+        total_tokens = usage.get("total_tokens", 0)
+
+        total_cost = self._calculate_cost("moonshot", model, prompt_tokens, completion_tokens)
+
+        meta = {
+            "prompt_tokens": prompt_tokens,
+            "completion_tokens": completion_tokens,
+            "total_tokens": total_tokens,
+            "cost": round(total_cost, 6),
+            "raw_response": resp,
+            "model_name": model,
+        }
+
+        text = resp["choices"][0]["message"]["content"]
+        return {"text": text, "meta": meta}
+
+    # ------------------------------------------------------------------
+    # Tool use
+    # ------------------------------------------------------------------
+
+    async def generate_messages_with_tools(
+        self,
+        messages: list[dict[str, Any]],
+        tools: list[dict[str, Any]],
+        options: dict[str, Any],
+    ) -> dict[str, Any]:
+        """Generate a response that may include tool calls."""
+        model = options.get("model", self.model)
+        model_config = self._get_model_config("moonshot", model)
+        tokens_param = model_config["tokens_param"]
+        supports_temperature = model_config["supports_temperature"]
+
+        self._validate_model_capabilities("moonshot", model, using_tool_use=True)
+
+        opts = {"temperature": 1.0, "max_tokens": 512, **options}
+        MoonshotDriver._clamp_temperature(opts)
+
+        data: dict[str, Any] = {
+            "model": model,
+            "messages": messages,
+            "tools": tools,
+        }
+        data[tokens_param] = opts.get("max_tokens", 512)
+
+        if supports_temperature and "temperature" in opts:
+            data["temperature"] = opts["temperature"]
+
+        if "tool_choice" in options:
+            data["tool_choice"] = options["tool_choice"]
+
+        MoonshotDriver._sanitize_tool_choice(data)
+
+        async with httpx.AsyncClient() as client:
+            try:
+                response = await client.post(
+                    f"{self.base_url}/chat/completions",
+                    headers=self.headers,
+                    json=data,
+                    timeout=120,
+                )
+                response.raise_for_status()
+                resp = response.json()
+            except httpx.HTTPStatusError as e:
+                error_msg = f"Moonshot API request failed: {e!s}"
+                raise RuntimeError(error_msg) from e
+            except Exception as e:
+                raise RuntimeError(f"Moonshot API request failed: {e!s}") from e
+
+        usage = resp.get("usage", {})
+        prompt_tokens = usage.get("prompt_tokens", 0)
+        completion_tokens = usage.get("completion_tokens", 0)
+        total_tokens = usage.get("total_tokens", 0)
+        total_cost = self._calculate_cost("moonshot", model, prompt_tokens, completion_tokens)
+
+        meta = {
+            "prompt_tokens": prompt_tokens,
+            "completion_tokens": completion_tokens,
+            "total_tokens": total_tokens,
+            "cost": round(total_cost, 6),
+            "raw_response": resp,
+            "model_name": model,
+        }
+
+        choice = resp["choices"][0]
+        text = choice["message"].get("content") or ""
+        stop_reason = choice.get("finish_reason")
+
+        tool_calls_out: list[dict[str, Any]] = []
+        for tc in choice["message"].get("tool_calls", []):
+            try:
+                args = json.loads(tc["function"]["arguments"])
+            except (json.JSONDecodeError, TypeError):
+                args = {}
+            tool_calls_out.append(
+                {
+                    "id": tc["id"],
+                    "name": tc["function"]["name"],
+                    "arguments": args,
+                }
+            )
+
+        return {
+            "text": text,
+            "meta": meta,
+            "tool_calls": tool_calls_out,
+            "stop_reason": stop_reason,
+        }
+
+    # ------------------------------------------------------------------
+    # Streaming
+    # ------------------------------------------------------------------
+
+    async def generate_messages_stream(
+        self,
+        messages: list[dict[str, Any]],
+        options: dict[str, Any],
+    ) -> AsyncIterator[dict[str, Any]]:
+        """Yield response chunks via Moonshot streaming API."""
+        model = options.get("model", self.model)
+        model_config = self._get_model_config("moonshot", model)
+        tokens_param = model_config["tokens_param"]
+        supports_temperature = model_config["supports_temperature"]
+
+        opts = {"temperature": 1.0, "max_tokens": 512, **options}
+        MoonshotDriver._clamp_temperature(opts)
+
+        data: dict[str, Any] = {
+            "model": model,
+            "messages": messages,
+            "stream": True,
+            "stream_options": {"include_usage": True},
+        }
+        data[tokens_param] = opts.get("max_tokens", 512)
+
+        if supports_temperature and "temperature" in opts:
+            data["temperature"] = opts["temperature"]
+
+        full_text = ""
+        prompt_tokens = 0
+        completion_tokens = 0
+
+        async with (
+            httpx.AsyncClient() as client,
+            client.stream(
+                "POST",
+                f"{self.base_url}/chat/completions",
+                headers=self.headers,
+                json=data,
+                timeout=120,
+            ) as response,
+        ):
+            response.raise_for_status()
+            async for line in response.aiter_lines():
+                if not line or not line.startswith("data: "):
+                    continue
+                payload = line[len("data: ") :]
+                if payload.strip() == "[DONE]":
+                    break
+                try:
+                    chunk = json.loads(payload)
+                except json.JSONDecodeError:
+                    continue
+
+                usage = chunk.get("usage")
+                if usage:
+                    prompt_tokens = usage.get("prompt_tokens", 0)
+                    completion_tokens = usage.get("completion_tokens", 0)
+
+                choices = chunk.get("choices", [])
+                if choices:
+                    delta = choices[0].get("delta", {})
+                    content = delta.get("content", "")
+                    if content:
+                        full_text += content
+                        yield {"type": "delta", "text": content}
+
+        total_tokens = prompt_tokens + completion_tokens
+        total_cost = self._calculate_cost("moonshot", model, prompt_tokens, completion_tokens)
+
+        yield {
+            "type": "done",
+            "text": full_text,
+            "meta": {
+                "prompt_tokens": prompt_tokens,
+                "completion_tokens": completion_tokens,
+                "total_tokens": total_tokens,
+                "cost": round(total_cost, 6),
+                "raw_response": {},
+                "model_name": model,
+            },
+        }

prompture/drivers/async_openai_driver.py

@@ -13,7 +13,7 @@ except Exception:
     AsyncOpenAI = None
 
 from ..async_driver import AsyncDriver
-from ..cost_mixin import CostMixin
+from ..cost_mixin import CostMixin, prepare_strict_schema
 from .openai_driver import OpenAIDriver
 
 
@@ -80,12 +80,13 @@ class AsyncOpenAIDriver(CostMixin, AsyncDriver):
         if options.get("json_mode"):
             json_schema = options.get("json_schema")
             if json_schema:
+                schema_copy = prepare_strict_schema(json_schema)
                 kwargs["response_format"] = {
                     "type": "json_schema",
                     "json_schema": {
                         "name": "extraction",
                         "strict": True,
-                        "schema": json_schema,
+                        "schema": schema_copy,
                     },
                 }
             else:

prompture/drivers/async_openrouter_driver.py

@@ -2,23 +2,28 @@
 
 from __future__ import annotations
 
+import json
 import os
+from collections.abc import AsyncIterator
 from typing import Any
 
 import httpx
 
 from ..async_driver import AsyncDriver
-from ..cost_mixin import CostMixin
+from ..cost_mixin import CostMixin, prepare_strict_schema
 from .openrouter_driver import OpenRouterDriver
 
 
 class AsyncOpenRouterDriver(CostMixin, AsyncDriver):
     supports_json_mode = True
+    supports_json_schema = True
+    supports_tool_use = True
+    supports_streaming = True
     supports_vision = True
 
     MODEL_PRICING = OpenRouterDriver.MODEL_PRICING
 
-    def __init__(self, api_key: str | None = None, model: str = "openai/gpt-3.5-turbo"):
+    def __init__(self, api_key: str | None = None, model: str = "openai/gpt-4o-mini"):
         self.api_key = api_key or os.getenv("OPENROUTER_API_KEY")
         if not self.api_key:
             raise ValueError("OpenRouter API key not found. Set OPENROUTER_API_KEY env var.")
@@ -51,6 +56,13 @@ class AsyncOpenRouterDriver(CostMixin, AsyncDriver):
         tokens_param = model_config["tokens_param"]
         supports_temperature = model_config["supports_temperature"]
 
+        # Validate capabilities against models.dev metadata
+        self._validate_model_capabilities(
+            "openrouter",
+            model,
+            using_json_schema=bool(options.get("json_schema")),
+        )
+
         opts = {"temperature": 1.0, "max_tokens": 512, **options}
 
         data = {
@@ -64,7 +76,19 @@ class AsyncOpenRouterDriver(CostMixin, AsyncDriver):
 
         # Native JSON mode support
         if options.get("json_mode"):
-            data["response_format"] = {"type": "json_object"}
+            json_schema = options.get("json_schema")
+            if json_schema:
+                schema_copy = prepare_strict_schema(json_schema)
+                data["response_format"] = {
+                    "type": "json_schema",
+                    "json_schema": {
+                        "name": "extraction",
+                        "strict": True,
+                        "schema": schema_copy,
+                    },
+                }
+            else:
+                data["response_format"] = {"type": "json_object"}
 
         async with httpx.AsyncClient() as client:
             try:
@@ -100,3 +124,168 @@ class AsyncOpenRouterDriver(CostMixin, AsyncDriver):
 
         text = resp["choices"][0]["message"]["content"]
         return {"text": text, "meta": meta}
+
+    # ------------------------------------------------------------------
+    # Tool use
+    # ------------------------------------------------------------------
+
+    async def generate_messages_with_tools(
+        self,
+        messages: list[dict[str, Any]],
+        tools: list[dict[str, Any]],
+        options: dict[str, Any],
+    ) -> dict[str, Any]:
+        """Generate a response that may include tool calls."""
+        model = options.get("model", self.model)
+        model_config = self._get_model_config("openrouter", model)
+        tokens_param = model_config["tokens_param"]
+        supports_temperature = model_config["supports_temperature"]
+
+        self._validate_model_capabilities("openrouter", model, using_tool_use=True)
+
+        opts = {"temperature": 1.0, "max_tokens": 512, **options}
+
+        data: dict[str, Any] = {
+            "model": model,
+            "messages": messages,
+            "tools": tools,
+        }
+        data[tokens_param] = opts.get("max_tokens", 512)
+
+        if supports_temperature and "temperature" in opts:
+            data["temperature"] = opts["temperature"]
+
+        async with httpx.AsyncClient() as client:
+            try:
+                response = await client.post(
+                    f"{self.base_url}/chat/completions",
+                    headers=self.headers,
+                    json=data,
+                    timeout=120,
+                )
+                response.raise_for_status()
+                resp = response.json()
+            except httpx.HTTPStatusError as e:
+                error_msg = f"OpenRouter API request failed: {e!s}"
+                raise RuntimeError(error_msg) from e
+            except Exception as e:
+                raise RuntimeError(f"OpenRouter API request failed: {e!s}") from e
+
+        usage = resp.get("usage", {})
+        prompt_tokens = usage.get("prompt_tokens", 0)
+        completion_tokens = usage.get("completion_tokens", 0)
+        total_tokens = usage.get("total_tokens", 0)
+        total_cost = self._calculate_cost("openrouter", model, prompt_tokens, completion_tokens)
+
+        meta = {
+            "prompt_tokens": prompt_tokens,
+            "completion_tokens": completion_tokens,
+            "total_tokens": total_tokens,
+            "cost": round(total_cost, 6),
+            "raw_response": resp,
+            "model_name": model,
+        }
+
+        choice = resp["choices"][0]
+        text = choice["message"].get("content") or ""
+        stop_reason = choice.get("finish_reason")
+
+        tool_calls_out: list[dict[str, Any]] = []
+        for tc in choice["message"].get("tool_calls", []):
+            try:
+                args = json.loads(tc["function"]["arguments"])
+            except (json.JSONDecodeError, TypeError):
+                args = {}
+            tool_calls_out.append({
+                "id": tc["id"],
+                "name": tc["function"]["name"],
+                "arguments": args,
+            })
+
+        return {
+            "text": text,
+            "meta": meta,
+            "tool_calls": tool_calls_out,
+            "stop_reason": stop_reason,
+        }
+
+    # ------------------------------------------------------------------
+    # Streaming
+    # ------------------------------------------------------------------
+
+    async def generate_messages_stream(
+        self,
+        messages: list[dict[str, Any]],
+        options: dict[str, Any],
+    ) -> AsyncIterator[dict[str, Any]]:
+        """Yield response chunks via OpenRouter streaming API."""
+        model = options.get("model", self.model)
+        model_config = self._get_model_config("openrouter", model)
+        tokens_param = model_config["tokens_param"]
+        supports_temperature = model_config["supports_temperature"]
+
+        opts = {"temperature": 1.0, "max_tokens": 512, **options}
+
+        data: dict[str, Any] = {
+            "model": model,
+            "messages": messages,
+            "stream": True,
+            "stream_options": {"include_usage": True},
+        }
+        data[tokens_param] = opts.get("max_tokens", 512)
+
+        if supports_temperature and "temperature" in opts:
+            data["temperature"] = opts["temperature"]
+
+        full_text = ""
+        prompt_tokens = 0
+        completion_tokens = 0
+
+        async with httpx.AsyncClient() as client, client.stream(
+            "POST",
+            f"{self.base_url}/chat/completions",
+            headers=self.headers,
+            json=data,
+            timeout=120,
+        ) as response:
+            response.raise_for_status()
+            async for line in response.aiter_lines():
+                if not line or not line.startswith("data: "):
+                    continue
+                payload = line[len("data: "):]
+                if payload.strip() == "[DONE]":
+                    break
+                try:
+                    chunk = json.loads(payload)
+                except json.JSONDecodeError:
+                    continue
+
+                # Usage comes in the final chunk
+                usage = chunk.get("usage")
+                if usage:
+                    prompt_tokens = usage.get("prompt_tokens", 0)
+                    completion_tokens = usage.get("completion_tokens", 0)
+
+                choices = chunk.get("choices", [])
+                if choices:
+                    delta = choices[0].get("delta", {})
+                    content = delta.get("content", "")
+                    if content:
+                        full_text += content
+                        yield {"type": "delta", "text": content}
+
+        total_tokens = prompt_tokens + completion_tokens
+        total_cost = self._calculate_cost("openrouter", model, prompt_tokens, completion_tokens)
+
+        yield {
+            "type": "done",
+            "text": full_text,
+            "meta": {
+                "prompt_tokens": prompt_tokens,
+                "completion_tokens": completion_tokens,
+                "total_tokens": total_tokens,
+                "cost": round(total_cost, 6),
+                "raw_response": {},
+                "model_name": model,
+            },
+        }

prompture/drivers/async_registry.py

@@ -22,9 +22,12 @@ from .async_grok_driver import AsyncGrokDriver
 from .async_groq_driver import AsyncGroqDriver
 from .async_lmstudio_driver import AsyncLMStudioDriver
 from .async_local_http_driver import AsyncLocalHTTPDriver
+from .async_modelscope_driver import AsyncModelScopeDriver
+from .async_moonshot_driver import AsyncMoonshotDriver
 from .async_ollama_driver import AsyncOllamaDriver
 from .async_openai_driver import AsyncOpenAIDriver
 from .async_openrouter_driver import AsyncOpenRouterDriver
+from .async_zai_driver import AsyncZaiDriver
 from .registry import (
     _get_async_registry,
     get_async_driver_factory,
@@ -90,6 +93,33 @@ register_async_driver(
     lambda model=None: AsyncGrokDriver(api_key=settings.grok_api_key, model=model or settings.grok_model),
     overwrite=True,
 )
+register_async_driver(
+    "moonshot",
+    lambda model=None: AsyncMoonshotDriver(
+        api_key=settings.moonshot_api_key,
+        model=model or settings.moonshot_model,
+        endpoint=settings.moonshot_endpoint,
+    ),
+    overwrite=True,
+)
+register_async_driver(
+    "modelscope",
+    lambda model=None: AsyncModelScopeDriver(
+        api_key=settings.modelscope_api_key,
+        model=model or settings.modelscope_model,
+        endpoint=settings.modelscope_endpoint,
+    ),
+    overwrite=True,
+)
+register_async_driver(
+    "zai",
+    lambda model=None: AsyncZaiDriver(
+        api_key=settings.zhipu_api_key,
+        model=model or settings.zhipu_model,
+        endpoint=settings.zhipu_endpoint,
+    ),
+    overwrite=True,
+)
 register_async_driver(
     "airllm",
     lambda model=None: AsyncAirLLMDriver(