agentforge-openai 0.2.3__py3-none-any.whl

agentforge_openai/__init__.py

@@ -0,0 +1,16 @@
+"""`agentforge-openai` — OpenAI LLM + embedding provider."""
+
+from __future__ import annotations
+
+from agentforge_openai._runner import OpenAIRunner
+from agentforge_openai.client import OpenAIClient
+from agentforge_openai.embedding import OpenAIEmbeddingClient
+
+__version__ = "0.2.3"
+
+__all__ = [
+    "OpenAIClient",
+    "OpenAIEmbeddingClient",
+    "OpenAIRunner",
+    "__version__",
+]

agentforge_openai/_inmem_runner.py

@@ -0,0 +1,141 @@
+"""In-memory `OpenAIRunner` for unit tests."""
+
+from __future__ import annotations
+
+from collections.abc import AsyncIterator
+from dataclasses import dataclass, field
+from typing import Any
+
+
+@dataclass
+class _ChatCall:
+    model: str
+    messages: list[dict[str, Any]]
+    tools: list[dict[str, Any]] | None
+    timeout_s: float
+    extra: dict[str, Any] = field(default_factory=dict)
+
+
+@dataclass
+class _EmbedCall:
+    model: str
+    inputs: list[str]
+    timeout_s: float
+    dimensions: int | None
+
+
+def _default_chat_response() -> dict[str, Any]:
+    return {
+        "id": "chatcmpl_test",
+        "model": "gpt-test",
+        "choices": [
+            {
+                "index": 0,
+                "message": {"role": "assistant", "content": ""},
+                "finish_reason": "stop",
+            },
+        ],
+        "usage": {"prompt_tokens": 0, "completion_tokens": 0, "total_tokens": 0},
+    }
+
+
+def _default_embedding_response(model: str, inputs: list[str], dims: int) -> dict[str, Any]:
+    return {
+        "model": model,
+        "data": [{"index": i, "embedding": [0.0] * dims} for i in range(len(inputs))],
+        "usage": {"prompt_tokens": sum(len(t) for t in inputs), "total_tokens": 0},
+    }
+
+
+class FakeOpenAIRunner:
+    """In-memory recorder for unit tests."""
+
+    def __init__(self) -> None:
+        self._chat_response: dict[str, Any] = _default_chat_response()
+        self._stream_chunks: list[dict[str, Any]] = []
+        self._embedding_dims: int = 1536
+        self._embedding_response: dict[str, Any] | None = None
+        self.chat_calls: list[_ChatCall] = []
+        self.stream_calls: list[_ChatCall] = []
+        self.embedding_calls: list[_EmbedCall] = []
+        self.closed = False
+
+    def set_chat_response(self, response: dict[str, Any]) -> None:
+        self._chat_response = response
+
+    def set_stream_chunks(self, chunks: list[dict[str, Any]]) -> None:
+        self._stream_chunks = list(chunks)
+
+    def set_embedding_dims(self, dims: int) -> None:
+        self._embedding_dims = dims
+
+    def set_embedding_response(self, response: dict[str, Any]) -> None:
+        self._embedding_response = response
+
+    async def chat_completions_create(
+        self,
+        *,
+        model: str,
+        messages: list[dict[str, Any]],
+        tools: list[dict[str, Any]] | None,
+        timeout_s: float,
+        extra: dict[str, Any] | None = None,
+    ) -> dict[str, Any]:
+        self.chat_calls.append(
+            _ChatCall(
+                model=model,
+                messages=[dict(m) for m in messages],
+                tools=[dict(t) for t in tools] if tools else None,
+                timeout_s=timeout_s,
+                extra=dict(extra or {}),
+            ),
+        )
+        return self._chat_response
+
+    async def chat_completions_stream(
+        self,
+        *,
+        model: str,
+        messages: list[dict[str, Any]],
+        tools: list[dict[str, Any]] | None,
+        timeout_s: float,
+        extra: dict[str, Any] | None = None,
+    ) -> AsyncIterator[dict[str, Any]]:
+        self.stream_calls.append(
+            _ChatCall(
+                model=model,
+                messages=[dict(m) for m in messages],
+                tools=[dict(t) for t in tools] if tools else None,
+                timeout_s=timeout_s,
+                extra=dict(extra or {}),
+            ),
+        )
+        return _stream_iter(self._stream_chunks)
+
+    async def embeddings_create(
+        self,
+        *,
+        model: str,
+        inputs: list[str],
+        timeout_s: float,
+        dimensions: int | None = None,
+    ) -> dict[str, Any]:
+        self.embedding_calls.append(
+            _EmbedCall(
+                model=model, inputs=list(inputs), timeout_s=timeout_s, dimensions=dimensions
+            ),
+        )
+        if self._embedding_response is not None:
+            return self._embedding_response
+        return _default_embedding_response(model, inputs, dimensions or self._embedding_dims)
+
+    async def close(self) -> None:
+        self.closed = True
+
+
+async def _stream_iter(chunks: list[dict[str, Any]]) -> AsyncIterator[dict[str, Any]]:
+    for c in chunks:
+        yield c
+
+
+__all__ = ["FakeOpenAIRunner"]

agentforge_openai/_pricing.py

@@ -0,0 +1,66 @@
+"""Per-model price table for OpenAI models.
+
+Prices are USD per 1M tokens, snapshotted 2026-05-14 from
+OpenAI's public pricing page. Embedding models use the same
+shape but only `input` is meaningful.
+
+A model not in the table returns `0.0` and we log a debug
+message — callers see `cost_usd=0.0` rather than crashing.
+"""
+
+from __future__ import annotations
+
+import logging
+
+log = logging.getLogger(__name__)
+
+_CHAT_PRICES: dict[str, tuple[float, float]] = {
+    # Each entry: input / output USD per 1M tokens.
+    "gpt-4o": (2.50, 10.0),
+    "gpt-4o-mini": (0.15, 0.60),
+    "gpt-4.1": (3.0, 12.0),
+    "gpt-4.1-mini": (0.40, 1.60),
+    "gpt-4.1-nano": (0.10, 0.40),
+    "o3": (10.0, 40.0),
+    "o3-mini": (1.10, 4.40),
+    "o4-mini": (1.10, 4.40),
+}
+
+_EMBED_PRICES: dict[str, float] = {
+    # Each entry: USD per 1M input tokens (embeddings have no output).
+    "text-embedding-3-small": 0.02,
+    "text-embedding-3-large": 0.13,
+    "text-embedding-ada-002": 0.10,
+}
+
+
+def chat_cost_usd(model: str, *, input_tokens: int, output_tokens: int) -> float:
+    canonical = _canonical_chat(model)
+    prices = _CHAT_PRICES.get(canonical)
+    if prices is None:
+        log.debug("agentforge-openai: no chat-price entry for %r; cost_usd=0.0", model)
+        return 0.0
+    in_p, out_p = prices
+    return (input_tokens / 1_000_000) * in_p + (output_tokens / 1_000_000) * out_p
+
+
+def embedding_cost_usd(model: str, *, input_tokens: int) -> float:
+    price = _EMBED_PRICES.get(model)
+    if price is None:
+        log.debug("agentforge-openai: no embedding-price entry for %r; cost_usd=0.0", model)
+        return 0.0
+    return (input_tokens / 1_000_000) * price
+
+
+def _canonical_chat(model: str) -> str:
+    """Strip dated suffixes (`-2026-03-01`, `-20260301`) for lookup."""
+    # OpenAI uses `-YYYY-MM-DD` style suffixes; chop after the model
+    # family if a date-like tail is present.
+    for suffix_prefix in ("-2025", "-2026", "-2027", "-2028", "-2029", "-2030"):
+        idx = model.find(suffix_prefix)
+        if idx > 0:
+            return model[:idx]
+    return model
+
+
+__all__ = ["chat_cost_usd", "embedding_cost_usd"]

agentforge_openai/_runner.py

@@ -0,0 +1,133 @@
+"""OpenAI runner Protocol + production SDK wrapper.
+
+Protocol abstracts the three SDK calls we make (chat completions,
+chat completions stream, embeddings) so unit tests don't need
+`openai` in the dev venv. Production `_OpenAISDKRunner` wraps
+`openai.AsyncOpenAI` under `# pragma: no cover`.
+"""
+
+from __future__ import annotations
+
+from typing import Any, Protocol, cast
+
+
+class OpenAIRunner(Protocol):
+    """Lifecycle Protocol for OpenAI API calls."""
+
+    async def chat_completions_create(
+        self,
+        *,
+        model: str,
+        messages: list[dict[str, Any]],
+        tools: list[dict[str, Any]] | None,
+        timeout_s: float,
+        extra: dict[str, Any] | None = None,
+    ) -> dict[str, Any]:  # pragma: no cover
+        """Issue a non-streaming chat.completions.create call."""
+        ...
+
+    async def chat_completions_stream(
+        self,
+        *,
+        model: str,
+        messages: list[dict[str, Any]],
+        tools: list[dict[str, Any]] | None,
+        timeout_s: float,
+        extra: dict[str, Any] | None = None,
+    ) -> Any:  # pragma: no cover
+        """Open a streaming chat.completions.create call.
+
+        Returns an async iterator over per-event chunks.
+        """
+        ...
+
+    async def embeddings_create(
+        self,
+        *,
+        model: str,
+        inputs: list[str],
+        timeout_s: float,
+        dimensions: int | None = None,
+    ) -> dict[str, Any]:  # pragma: no cover
+        """Issue an embeddings.create call."""
+        ...
+
+    async def close(self) -> None:  # pragma: no cover
+        """Release the underlying HTTP client."""
+        ...
+
+
+class _OpenAISDKRunner:  # pragma: no cover — exercised only with `-m live`.
+    """Production runner wrapping ``openai.AsyncOpenAI``."""
+
+    def __init__(self, client: Any) -> None:
+        self._client = client
+
+    async def chat_completions_create(
+        self,
+        *,
+        model: str,
+        messages: list[dict[str, Any]],
+        tools: list[dict[str, Any]] | None,
+        timeout_s: float,
+        extra: dict[str, Any] | None = None,
+    ) -> dict[str, Any]:
+        kwargs: dict[str, Any] = {
+            "model": model,
+            "messages": messages,
+            "timeout": timeout_s,
+        }
+        if tools:
+            kwargs["tools"] = tools
+        if extra:
+            kwargs.update(extra)
+        result = await self._client.chat.completions.create(**kwargs)
+        return cast("dict[str, Any]", result.model_dump(mode="python", exclude_none=False))
+
+    async def chat_completions_stream(
+        self,
+        *,
+        model: str,
+        messages: list[dict[str, Any]],
+        tools: list[dict[str, Any]] | None,
+        timeout_s: float,
+        extra: dict[str, Any] | None = None,
+    ) -> Any:
+        kwargs: dict[str, Any] = {
+            "model": model,
+            "messages": messages,
+            "stream": True,
+            "timeout": timeout_s,
+            "stream_options": {"include_usage": True},
+        }
+        if tools:
+            kwargs["tools"] = tools
+        if extra:
+            kwargs.update(extra)
+        return await self._client.chat.completions.create(**kwargs)
+
+    async def embeddings_create(
+        self,
+        *,
+        model: str,
+        inputs: list[str],
+        timeout_s: float,
+        dimensions: int | None = None,
+    ) -> dict[str, Any]:
+        kwargs: dict[str, Any] = {
+            "model": model,
+            "input": inputs,
+            "timeout": timeout_s,
+        }
+        if dimensions is not None:
+            kwargs["dimensions"] = dimensions
+        result = await self._client.embeddings.create(**kwargs)
+        return cast("dict[str, Any]", result.model_dump(mode="python", exclude_none=False))
+
+    async def close(self) -> None:
+        close = getattr(self._client, "close", None)
+        if callable(close):
+            await close()
+
+
+__all__ = ["OpenAIRunner"]

agentforge_openai/client.py

@@ -0,0 +1,356 @@
+"""`OpenAIClient` — `LLMClient` over OpenAI's chat.completions API.
+
+Maps framework shapes onto OpenAI's request/response model:
+
+- System prompt → first `{"role": "system", ...}` element in messages.
+- Messages → `[{"role": ..., "content": ...}]`.
+  Tool result messages encode with `role="tool"` + `tool_call_id`.
+- Tools → `[{"type": "function", "function": {...}}]`.
+
+Finish reasons normalised:
+  stop / length / tool_calls / function_call / content_filter →
+  end_turn / max_tokens / tool_use / tool_use / other
+"""
+
+from __future__ import annotations
+
+from collections.abc import AsyncIterator
+from typing import TYPE_CHECKING, Any
+
+from agentforge_core.contracts.llm import LLMClient
+from agentforge_core.production.exceptions import ModuleError
+from agentforge_core.resolver import register_provider
+from agentforge_core.values.messages import (
+    LLMResponse,
+    Message,
+    StopReason,
+    StreamChunk,
+    TokenUsage,
+    ToolCall,
+    ToolSpec,
+)
+
+from agentforge_openai._pricing import chat_cost_usd
+
+if TYPE_CHECKING:
+    from agentforge_openai._runner import OpenAIRunner
+
+
+_PROVIDER_NAME = "openai"
+_DEFAULT_TIMEOUT_SECONDS = 60.0
+
+_FINISH_REASON_MAP: dict[str, StopReason] = {
+    "stop": "end_turn",
+    "length": "max_tokens",
+    "tool_calls": "tool_use",
+    "function_call": "tool_use",
+    "content_filter": "other",
+}
+
+_VISION_MODELS = frozenset({"gpt-4o", "gpt-4o-mini", "gpt-4.1", "gpt-4.1-mini", "gpt-4.1-nano"})
+
+
+@register_provider("openai")
+class OpenAIClient(LLMClient):
+    """`LLMClient` over OpenAI's chat.completions API."""
+
+    def __init__(
+        self,
+        *,
+        model_id: str,
+        runner: OpenAIRunner | None = None,
+        api_key: str | None = None,
+        base_url: str | None = None,
+        organization: str | None = None,
+        timeout_seconds: float = _DEFAULT_TIMEOUT_SECONDS,
+        json_mode: bool = False,
+    ) -> None:
+        if not model_id:
+            raise ValueError("model_id must be a non-empty string")
+        if timeout_seconds <= 0:
+            raise ValueError("timeout_seconds must be > 0")
+        self._runner = (
+            runner
+            if runner is not None
+            else _build_sdk_runner(
+                api_key=api_key,
+                base_url=base_url,
+                organization=organization,
+            )
+        )
+        self._model = model_id
+        self._timeout_seconds = timeout_seconds
+        self._json_mode = json_mode
+
+    @classmethod
+    def from_config(
+        cls,
+        *,
+        model: str,
+        api_key: str | None = None,
+        base_url: str | None = None,
+        organization: str | None = None,
+        timeout_seconds: float = _DEFAULT_TIMEOUT_SECONDS,
+        json_mode: bool = False,
+    ) -> OpenAIClient:  # pragma: no cover — exercised only with `-m live`.
+        """Build an `OpenAIClient` backed by a real `openai.AsyncOpenAI`."""
+        return cls(
+            model_id=model,
+            api_key=api_key,
+            base_url=base_url,
+            organization=organization,
+            timeout_seconds=timeout_seconds,
+            json_mode=json_mode,
+        )
+
+    def capabilities(self) -> set[str]:
+        caps = {"tools", "json_mode", "streaming"}
+        if _canonical_chat(self._model) in _VISION_MODELS:
+            caps.add("vision")
+        return caps
+
+    async def call(
+        self,
+        system: str,
+        messages: list[Message],
+        tools: list[ToolSpec] | None = None,
+    ) -> LLMResponse:
+        raw = await self._runner.chat_completions_create(
+            model=self._model,
+            messages=_build_messages(system, messages),
+            tools=_tools_to_openai(tools),
+            timeout_s=self._timeout_seconds,
+            extra=self._extra_kwargs(),
+        )
+        return self._normalise_response(raw)
+
+    def stream(
+        self,
+        system: str,
+        messages: list[Message],
+        tools: list[ToolSpec] | None = None,
+    ) -> AsyncIterator[StreamChunk]:
+        return self._stream_request(system=system, messages=messages, tools=tools)
+
+    async def close(self) -> None:
+        await self._runner.close()
+
+    # ------------------------------------------------------------------
+    # Internal
+    # ------------------------------------------------------------------
+
+    def _extra_kwargs(self) -> dict[str, Any] | None:
+        if self._json_mode:
+            return {"response_format": {"type": "json_object"}}
+        return None
+
+    async def _stream_request(
+        self,
+        *,
+        system: str,
+        messages: list[Message],
+        tools: list[ToolSpec] | None,
+    ) -> AsyncIterator[StreamChunk]:
+        stream = await self._runner.chat_completions_stream(
+            model=self._model,
+            messages=_build_messages(system, messages),
+            tools=_tools_to_openai(tools),
+            timeout_s=self._timeout_seconds,
+            extra=self._extra_kwargs(),
+        )
+        tool_buffers: dict[int, dict[str, Any]] = {}
+        stop_reason: StopReason = "end_turn"
+        usage = TokenUsage(input_tokens=0, output_tokens=0)
+
+        async for event in stream:
+            for choice in event.get("choices", []) or []:
+                delta = choice.get("delta") or {}
+                text = delta.get("content")
+                if text:
+                    yield StreamChunk(kind="text", delta=text)
+                for tc in delta.get("tool_calls") or []:
+                    idx = int(tc.get("index", 0))
+                    buf = tool_buffers.setdefault(
+                        idx,
+                        {"id": "", "name": "", "input_json": ""},
+                    )
+                    if tc.get("id"):
+                        buf["id"] = tc["id"]
+                    fn = tc.get("function") or {}
+                    if fn.get("name"):
+                        buf["name"] = fn["name"]
+                    if fn.get("arguments"):
+                        buf["input_json"] += fn["arguments"]
+                finish_raw = choice.get("finish_reason")
+                if finish_raw:
+                    stop_reason = _FINISH_REASON_MAP.get(finish_raw, "other")
+            usage_raw = event.get("usage")
+            if usage_raw:
+                usage = TokenUsage(
+                    input_tokens=int(usage_raw.get("prompt_tokens", 0)),
+                    output_tokens=int(usage_raw.get("completion_tokens", 0)),
+                )
+
+        for buf in tool_buffers.values():
+            import json as _json  # noqa: PLC0415
+
+            try:
+                args = _json.loads(buf["input_json"]) if buf["input_json"] else {}
+            except _json.JSONDecodeError:
+                args = {}
+            if not isinstance(args, dict):
+                args = {}
+            yield StreamChunk(
+                kind="tool_call",
+                tool_call=ToolCall(id=buf["id"], name=buf["name"], arguments=args),
+            )
+
+        cost = chat_cost_usd(
+            self._model,
+            input_tokens=usage.input_tokens,
+            output_tokens=usage.output_tokens,
+        )
+        yield StreamChunk(
+            kind="stop",
+            stop_reason=stop_reason,
+            usage=usage,
+            cost_usd=cost,
+        )
+
+    def _normalise_response(self, raw: dict[str, Any]) -> LLMResponse:
+        choices = raw.get("choices") or []
+        choice0 = choices[0] if choices else {}
+        msg = choice0.get("message") or {}
+
+        content = msg.get("content") or ""
+        if isinstance(content, list):
+            # gpt-4o may return content as a list of parts; flatten text parts.
+            content = "".join(
+                part.get("text", "")
+                for part in content
+                if isinstance(part, dict) and part.get("type") == "text"
+            )
+
+        tool_calls: list[ToolCall] = []
+        for tc in msg.get("tool_calls") or []:
+            fn = tc.get("function") or {}
+            args_raw = fn.get("arguments") or "{}"
+            import json as _json  # noqa: PLC0415
+
+            try:
+                args = _json.loads(args_raw)
+            except _json.JSONDecodeError:
+                args = {}
+            if not isinstance(args, dict):
+                args = {}
+            tool_calls.append(
+                ToolCall(id=tc.get("id", ""), name=fn.get("name", ""), arguments=args)
+            )
+
+        usage_raw = raw.get("usage") or {}
+        usage = TokenUsage(
+            input_tokens=int(usage_raw.get("prompt_tokens", 0)),
+            output_tokens=int(usage_raw.get("completion_tokens", 0)),
+        )
+
+        finish_raw = choice0.get("finish_reason") or "stop"
+        stop_reason: StopReason = _FINISH_REASON_MAP.get(finish_raw, "other")
+
+        cost = chat_cost_usd(
+            self._model,
+            input_tokens=usage.input_tokens,
+            output_tokens=usage.output_tokens,
+        )
+
+        return LLMResponse(
+            content=content,
+            tool_calls=tuple(tool_calls),
+            stop_reason=stop_reason,
+            usage=usage,
+            cost_usd=cost,
+            model=raw.get("model", self._model),
+            provider=_PROVIDER_NAME,
+        )
+
+
+# ----------------------------------------------------------------------
+# Helpers
+# ----------------------------------------------------------------------
+
+
+def _build_messages(system: str, messages: list[Message]) -> list[dict[str, Any]]:
+    """Build the OpenAI messages list. System prompt becomes the first
+    element when non-empty."""
+    out: list[dict[str, Any]] = []
+    if system:
+        out.append({"role": "system", "content": system})
+    out.extend(_message_to_openai(m) for m in messages)
+    return out
+
+
+def _message_to_openai(message: Message) -> dict[str, Any]:
+    if message.role == "tool":
+        return {
+            "role": "tool",
+            "tool_call_id": message.tool_call_id or "",
+            "content": message.content,
+        }
+    if message.role == "system":
+        # Should be hoisted by the caller; defensive passthrough.
+        return {"role": "system", "content": message.content}
+    return {"role": message.role, "content": message.content}
+
+
+def _tools_to_openai(tools: list[ToolSpec] | None) -> list[dict[str, Any]] | None:
+    if not tools:
+        return None
+    return [
+        {
+            "type": "function",
+            "function": {
+                "name": t.name,
+                "description": t.description,
+                "parameters": t.schema_,
+            },
+        }
+        for t in tools
+    ]
+
+
+def _canonical_chat(model: str) -> str:
+    for suffix_prefix in ("-2025", "-2026", "-2027", "-2028", "-2029", "-2030"):
+        idx = model.find(suffix_prefix)
+        if idx > 0:
+            return model[:idx]
+    return model
+
+
+def _build_sdk_runner(  # pragma: no cover — `-m live` only.
+    *,
+    api_key: str | None,
+    base_url: str | None,
+    organization: str | None,
+) -> OpenAIRunner:
+    try:
+        import openai  # noqa: PLC0415
+    except ImportError as exc:
+        msg = (
+            "openai is not installed. Install via "
+            "`pip install agentforge-openai[openai]` to use the production runner."
+        )
+        raise ModuleError(msg) from exc
+
+    from agentforge_openai._runner import _OpenAISDKRunner  # noqa: PLC0415
+
+    kwargs: dict[str, Any] = {}
+    if api_key:
+        kwargs["api_key"] = api_key
+    if base_url:
+        kwargs["base_url"] = base_url
+    if organization:
+        kwargs["organization"] = organization
+    client = openai.AsyncOpenAI(**kwargs)
+    return _OpenAISDKRunner(client)
+
+
+__all__ = ["OpenAIClient"]

agentforge_openai/embedding.py

@@ -0,0 +1,169 @@
+"""`OpenAIEmbeddingClient` — `EmbeddingClient` over OpenAI's
+`embeddings.create()` API.
+
+Supports `text-embedding-3-small` (1536 dims), `text-embedding-3-large`
+(3072 dims), and the legacy `text-embedding-ada-002` (1536 dims).
+The `text-embedding-3-*` models support Matryoshka-style truncation
+via the `dimensions=` API parameter; declaring that capability lets
+callers ask for a shorter vector at construction time.
+"""
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+from agentforge_core.contracts.embedding import EmbeddingClient
+from agentforge_core.production.exceptions import ModuleError
+from agentforge_core.resolver import register_embedding_provider
+from agentforge_core.values.messages import EmbeddingResponse, TokenUsage
+
+from agentforge_openai._pricing import embedding_cost_usd
+
+if TYPE_CHECKING:
+    from agentforge_openai._runner import OpenAIRunner
+
+
+_PROVIDER_NAME = "openai"
+_DEFAULT_TIMEOUT_SECONDS = 30.0
+
+_MODEL_DIMS: dict[str, int] = {
+    "text-embedding-3-small": 1536,
+    "text-embedding-3-large": 3072,
+    "text-embedding-ada-002": 1536,
+}
+
+_MATRYOSHKA_MODELS = frozenset({"text-embedding-3-small", "text-embedding-3-large"})
+
+
+@register_embedding_provider("openai")
+class OpenAIEmbeddingClient(EmbeddingClient):
+    """`EmbeddingClient` over OpenAI's embeddings API."""
+
+    def __init__(
+        self,
+        *,
+        runner: OpenAIRunner,
+        model: str,
+        dimensions: int | None = None,
+        timeout_seconds: float = _DEFAULT_TIMEOUT_SECONDS,
+    ) -> None:
+        if not model:
+            raise ValueError("model must be a non-empty string")
+        if timeout_seconds <= 0:
+            raise ValueError("timeout_seconds must be > 0")
+        native = _MODEL_DIMS.get(model)
+        if dimensions is not None:
+            if dimensions < 1:
+                raise ValueError("dimensions must be >= 1")
+            if native is not None and dimensions > native:
+                raise ValueError(
+                    f"dimensions={dimensions} exceeds native={native} for {model}",
+                )
+            if dimensions != native and model not in _MATRYOSHKA_MODELS:
+                raise ValueError(
+                    f"{model} does not support dimension override (Matryoshka)",
+                )
+        self._runner = runner
+        self._model = model
+        self._dimensions = dimensions if dimensions is not None else native or 1536
+        self._dim_arg: int | None = dimensions if dimensions != native else None
+        self._timeout_seconds = timeout_seconds
+
+    @classmethod
+    def from_config(
+        cls,
+        *,
+        model: str,
+        api_key: str | None = None,
+        base_url: str | None = None,
+        organization: str | None = None,
+        dimensions: int | None = None,
+        timeout_seconds: float = _DEFAULT_TIMEOUT_SECONDS,
+    ) -> OpenAIEmbeddingClient:  # pragma: no cover — exercised only with `-m live`.
+        runner = _build_sdk_runner(
+            api_key=api_key,
+            base_url=base_url,
+            organization=organization,
+        )
+        return cls(
+            runner=runner,
+            model=model,
+            dimensions=dimensions,
+            timeout_seconds=timeout_seconds,
+        )
+
+    def dimensions(self) -> int:
+        return self._dimensions
+
+    def capabilities(self) -> set[str]:
+        if self._model in _MATRYOSHKA_MODELS:
+            return {"matryoshka"}
+        return set()
+
+    async def embed(self, texts: list[str]) -> EmbeddingResponse:
+        if not texts:
+            raise ValueError("texts must contain at least one item")
+        raw = await self._runner.embeddings_create(
+            model=self._model,
+            inputs=texts,
+            timeout_s=self._timeout_seconds,
+            dimensions=self._dim_arg,
+        )
+        data = raw.get("data") or []
+        vectors = tuple(tuple(float(v) for v in item.get("embedding") or []) for item in data)
+        if any(len(v) != self._dimensions for v in vectors):
+            raise ValueError(
+                f"OpenAI returned vector with unexpected dimensionality; "
+                f"expected {self._dimensions}, got "
+                f"{[len(v) for v in vectors]}",
+            )
+        usage_raw = raw.get("usage") or {}
+        usage = TokenUsage(
+            input_tokens=int(usage_raw.get("prompt_tokens", 0)),
+            output_tokens=0,
+        )
+        cost = embedding_cost_usd(self._model, input_tokens=usage.input_tokens)
+        return EmbeddingResponse(
+            vectors=vectors,
+            dimensions=self._dimensions,
+            usage=usage,
+            cost_usd=cost,
+            model=self._model,
+            provider=_PROVIDER_NAME,
+        )
+
+    async def close(self) -> None:
+        await self._runner.close()
+
+
+def _build_sdk_runner(  # pragma: no cover — `-m live` only.
+    *,
+    api_key: str | None,
+    base_url: str | None,
+    organization: str | None,
+) -> OpenAIRunner:
+    try:
+        import openai  # noqa: PLC0415
+    except ImportError as exc:
+        msg = (
+            "openai is not installed. Install via "
+            "`pip install agentforge-openai[openai]` to use the production runner."
+        )
+        raise ModuleError(msg) from exc
+
+    from typing import Any  # noqa: PLC0415
+
+    from agentforge_openai._runner import _OpenAISDKRunner  # noqa: PLC0415
+
+    kwargs: dict[str, Any] = {}
+    if api_key:
+        kwargs["api_key"] = api_key
+    if base_url:
+        kwargs["base_url"] = base_url
+    if organization:
+        kwargs["organization"] = organization
+    client = openai.AsyncOpenAI(**kwargs)
+    return _OpenAISDKRunner(client)
+
+
+__all__ = ["OpenAIEmbeddingClient"]

agentforge_openai-0.2.3.dist-info/METADATA

@@ -0,0 +1,72 @@
+Metadata-Version: 2.4
+Name: agentforge-openai
+Version: 0.2.3
+Summary: OpenAI LLM + embeddings provider for AgentForge — gpt-4o / o-series + text-embedding-3-*
+Project-URL: Homepage, https://github.com/Scaffoldic/agentforge-py
+Project-URL: Repository, https://github.com/Scaffoldic/agentforge-py
+Project-URL: Documentation, https://github.com/Scaffoldic/agentforge-py
+Project-URL: Changelog, https://github.com/Scaffoldic/agentforge-py/blob/main/CHANGELOG.md
+Project-URL: Issues, https://github.com/Scaffoldic/agentforge-py/issues
+Author: The AgentForge Authors
+License-Expression: Apache-2.0
+License-File: LICENSE
+Keywords: agent,ai,embeddings,gpt,llm,openai
+Classifier: Development Status :: 2 - Pre-Alpha
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Typing :: Typed
+Requires-Python: >=3.13
+Requires-Dist: agentforge-core~=0.2.3
+Provides-Extra: openai
+Requires-Dist: openai>=1.50; extra == 'openai'
+Description-Content-Type: text/markdown
+
+# agentforge-openai
+
+OpenAI native LLM + embedding provider for
+[AgentForge](https://github.com/Scaffoldic/agentforge-py).
+
+Implements both `LLMClient` and `EmbeddingClient` against
+OpenAI's `chat.completions.create()` (gpt-4o / o-series) and
+`embeddings.create()` (text-embedding-3-small / -large).
+
+## Install
+
+```bash
+pip install "agentforge-openai[openai]"
+```
+
+The bare `pip install agentforge-openai` install keeps the
+package importable for tests; the production runner needs the
+`openai` SDK, which the `[openai]` extra pulls in.
+
+## Use
+
+```python
+from agentforge import Agent
+from agentforge_openai import OpenAIClient, OpenAIEmbeddingClient
+
+# Chat
+agent = Agent(model="openai:gpt-4o-mini")
+
+# Embeddings (direct construction)
+emb = OpenAIEmbeddingClient.from_config(
+    model="text-embedding-3-small",
+    api_key="sk-...",
+)
+```
+
+## Capabilities
+
+`OpenAIClient.capabilities()` declares: `tools`, `json_mode`,
+`streaming`, `vision` (for `gpt-4o*` variants). Caching and
+extended thinking are not surfaced (OpenAI handles caching
+transparently; o-series internal reasoning is not exposed as a
+budget-able knob through the public API today).
+
+## License
+
+Apache-2.0. See [LICENSE](./LICENSE).

agentforge_openai-0.2.3.dist-info/RECORD

@@ -0,0 +1,12 @@
+agentforge_openai/__init__.py,sha256=0ixKj61sA2M49A5gGw34gKK1dSRbviHY56tlGNNRAVA,389
+agentforge_openai/_inmem_runner.py,sha256=YVUPOA6vb-iNmhtSZxNfEo2BHXyqOcKNg8Yob8cdF4Y,4185
+agentforge_openai/_pricing.py,sha256=05SCZQBefyMEtCispe8272iVZo4MUbtS0D1uU6syFNc,2130
+agentforge_openai/_runner.py,sha256=2Qe1PCiH5tl2RkdjN5Rgh96XxGA1lK4h01GNqlDUtNU,3908
+agentforge_openai/client.py,sha256=KKXHOhgT36avFOA8ehTG64p5R4HX1FIEWA3WcBHYHho,11584
+agentforge_openai/embedding.py,sha256=z3e_ZBs6sJyGSbjZnJH1YfsF1T-vxm8dRriLpwpiQXU,5680
+agentforge_openai/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+agentforge_openai-0.2.3.dist-info/METADATA,sha256=5J1zTOJQ8iDbWMHAfEK40zbEfRCKItoqO0y3tgfNebA,2393
+agentforge_openai-0.2.3.dist-info/WHEEL,sha256=QccIxa26bgl1E6uMy58deGWi-0aeIkkangHcxk2kWfw,87
+agentforge_openai-0.2.3.dist-info/entry_points.txt,sha256=mReMRyQTZw8JHBXBrYf7B2uzGbbugQNZ_M21FOThkKk,137
+agentforge_openai-0.2.3.dist-info/licenses/LICENSE,sha256=z8d0m5b2O9McPEK1xHG_dWgUBT6EfBDz6wA0F7xSPTA,11358
+agentforge_openai-0.2.3.dist-info/RECORD,,

agentforge_openai-0.2.3.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.29.0
+Root-Is-Purelib: true
+Tag: py3-none-any

agentforge_openai-0.2.3.dist-info/entry_points.txt

@@ -0,0 +1,5 @@
+[agentforge.embeddings]
+openai = agentforge_openai:OpenAIEmbeddingClient
+
+[agentforge.providers]
+openai = agentforge_openai:OpenAIClient

agentforge_openai-0.2.3.dist-info/licenses/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.