rag-nextjs 1.0.0__py3-none-any.whl

rag_nextjs/__init__.py

@@ -0,0 +1 @@
+"""Next.js 公式ドキュメントを next-devtools-mcp 経由で取得する RAG 用パッケージ。"""

rag_nextjs/__main__.py

@@ -0,0 +1,4 @@
+from rag_nextjs.cli import main
+
+if __name__ == "__main__":
+    raise SystemExit(main())

rag_nextjs/answer.py

@@ -0,0 +1,351 @@
+from __future__ import annotations
+
+import os
+import sys
+from collections.abc import Callable, Iterator, Mapping
+from typing import Any
+
+from mcp import ClientSession
+from mcp.client.stdio import stdio_client
+from openai import OpenAI
+
+from rag_nextjs.console_status import ConsoleProgress, run_sync_blocking_with_spinner
+from rag_nextjs.indexing import rank_doc_paths
+from rag_nextjs.mcp_docs import (
+    fetch_nextjs_docs_page,
+    format_fetched_pages,
+    next_devtools_stdio_params,
+    read_llms_index,
+)
+
+
+def _env_bool(name: str, default: bool) -> bool:
+    raw = os.getenv(name)
+    if raw is None or raw.strip() == "":
+        return default
+    return raw.strip().lower() in {"1", "true", "yes", "on"}
+
+
+def _client_http_timeout() -> Any:
+    """HTTP タイムアウト。巨大モデルや遅いゲートウェイ向けに環境変数で延長可能。"""
+    import httpx
+
+    raw = (
+        os.getenv("OPENAI_TIMEOUT")
+        or os.getenv("HTTP_TIMEOUT")
+        or os.getenv("CHAT_HTTP_TIMEOUT_SECONDS")
+    )
+    seconds = float(raw) if raw else 600.0
+    connect_raw = os.getenv("CHAT_HTTP_CONNECT_SECONDS")
+    connect = float(connect_raw) if connect_raw else 30.0
+    return httpx.Timeout(timeout=seconds, connect=connect)
+
+
+def _get_client() -> OpenAI:
+    base = os.getenv("OPENAI_BASE_URL") or os.getenv("NVIDIA_OPENAI_BASE_URL")
+    key = os.getenv("OPENAI_API_KEY") or os.getenv("NVIDIA_API_KEY")
+    if not key:
+        raise RuntimeError("OPENAI_API_KEY または NVIDIA_API_KEY を .env に設定してください")
+    kwargs: dict[str, Any] = {"api_key": key, "timeout": _client_http_timeout()}
+    if base:
+        kwargs["base_url"] = base
+    return OpenAI(**kwargs)
+
+
+# OpenAI 互換 API が delta / message に載せる推論テキスト系フィールド（既定ではstdoutに出さない）
+_DELTA_REASONING_KEYS: tuple[str, ...] = (
+    "reasoning_content",
+    "reasoning",
+    "thinking",
+    "thought",
+)
+
+# content が配列のとき、type で推論ブロックとみなして除外する値
+_CONTENT_REASONING_TYPES: frozenset[str] = frozenset({"reasoning", "thinking", "thought"})
+
+
+def _get_str_field(obj: Any, key: str) -> str | None:
+    if obj is None:
+        return None
+    if isinstance(obj, Mapping):
+        val = obj.get(key)
+    else:
+        val = getattr(obj, key, None)
+    if isinstance(val, str) and val:
+        return val
+    return None
+
+
+def _stringify_content_field(content: Any, *, include_reasoning: bool) -> str | None:
+    """message / delta の content が str だけでなく list[dict] 形式のときに連結して返す。"""
+    if content is None:
+        return None
+    if isinstance(content, str):
+        return content if content.strip() else None
+    if isinstance(content, list):
+        parts: list[str] = []
+        for block in content:
+            if isinstance(block, str) and block:
+                parts.append(block)
+                continue
+            if isinstance(block, Mapping):
+                bt = block.get("type")
+                if (
+                    not include_reasoning
+                    and isinstance(bt, str)
+                    and bt.strip().lower() in _CONTENT_REASONING_TYPES
+                ):
+                    continue
+                t = block.get("text")
+                if isinstance(t, str) and t:
+                    parts.append(t)
+                    continue
+                if block.get("type") == "text":
+                    nested = block.get("text")
+                    if isinstance(nested, str) and nested:
+                        parts.append(nested)
+        return "".join(parts) if parts else None
+    return None
+
+
+def _iter_delta_text_parts(delta: Any, *, include_reasoning: bool) -> Iterator[str]:
+    """ストリーミング chunk の delta から、表示すべき文字列を順に取り出す。"""
+    if delta is None:
+        return
+    if include_reasoning:
+        for key in _DELTA_REASONING_KEYS:
+            s = _get_str_field(delta, key)
+            if s:
+                yield s
+    ref = _get_str_field(delta, "refusal")
+    if ref:
+        yield ref
+    raw_c = getattr(delta, "content", None) if not isinstance(delta, Mapping) else delta.get("content")
+    c = _stringify_content_field(raw_c, include_reasoning=include_reasoning)
+    if c:
+        yield c
+
+
+def _iter_message_text_parts(message: Any, *, include_reasoning: bool) -> Iterator[str]:
+    """非ストリーミング応答の message から表示用テキストを取り出す。"""
+    if message is None:
+        return
+    if include_reasoning:
+        for key in _DELTA_REASONING_KEYS:
+            s = _get_str_field(message, key)
+            if s:
+                yield s
+    ref = _get_str_field(message, "refusal")
+    if ref:
+        yield ref
+    raw_c = getattr(message, "content", None) if not isinstance(message, Mapping) else message.get(
+        "content"
+    )
+    c = _stringify_content_field(raw_c, include_reasoning=include_reasoning)
+    if c:
+        yield c
+
+
+def _iter_chunk_text_parts(chunk: Any, *, include_reasoning: bool) -> Iterator[str]:
+    """ChatCompletionChunk 互換オブジェクトからテキスト断片を取り出す。"""
+    choices = getattr(chunk, "choices", None) or []
+    for choice in choices:
+        delta = getattr(choice, "delta", None)
+        yield from _iter_delta_text_parts(delta, include_reasoning=include_reasoning)
+        # 一部プロキシはストリーム終端付近で message に本文だけ載せる
+        msg = getattr(choice, "message", None)
+        yield from _iter_message_text_parts(msg, include_reasoning=include_reasoning)
+
+
+def _model_name() -> str:
+    return os.getenv("CHAT_MODEL", "nvidia/nemotron-3-nano-30b-a3b")
+
+
+def _build_messages(question: str, context: str) -> list[dict[str, str]]:
+    system = (
+        "あなたは Next.js の公式ドキュメント（与えられたコンテキスト）だけを根拠に回答するアシスタントです。"
+        "コンテキストに無い内容は推測せず、その旨を述べてください。"
+        "コード例は公式に沿った形で示し、必要なら path（ドキュメント上のパス）を引用してください。"
+    )
+    user = f"## 参照した公式ドキュメント\n\n{context}\n\n## 質問\n\n{question}"
+    return [
+        {"role": "system", "content": system},
+        {"role": "user", "content": user},
+    ]
+
+
+async def gather_context(
+    session: ClientSession,
+    question: str,
+    top_k: int,
+    *,
+    on_phase: Callable[[str, str], None] | None = None,
+) -> tuple[str, list[str]]:
+    if on_phase:
+        on_phase("インデックスを読み込み中", "")
+    index = await read_llms_index(session)
+    paths = rank_doc_paths(question, index, top_k=top_k)
+    if not paths:
+        return "", []
+
+    pages: list[dict[str, Any]] = []
+    for i, path in enumerate(paths, 1):
+        if on_phase:
+            on_phase("ドキュメント本文を取得中", f"{i}/{len(paths)}  {path}")
+        try:
+            pages.append(await fetch_nextjs_docs_page(session, path))
+        except Exception as exc:  # noqa: BLE001
+            pages.append({"path": path, "url": "", "content": f"[取得エラー: {exc}]"})
+
+    return format_fetched_pages(pages), paths
+
+
+def _extra_body_from_env() -> dict[str, Any]:
+    raw_extra = os.getenv("CHAT_EXTRA_BODY")
+    if not raw_extra:
+        return {}
+    import json
+
+    return json.loads(raw_extra)
+
+
+def _build_chat_completion_kwargs(question: str, context: str, *, stream: bool) -> dict[str, Any]:
+    model = _model_name()
+    messages = _build_messages(question, context)
+    extra = _extra_body_from_env()
+
+    kwargs: dict[str, Any] = {
+        "model": model,
+        "messages": messages,
+        "temperature": float(os.getenv("CHAT_TEMPERATURE", "0.2")),
+        "stream": stream,
+    }
+
+    mt = os.getenv("CHAT_MAX_TOKENS")
+    if mt and mt.strip():
+        kwargs["max_tokens"] = int(mt.strip())
+
+    if extra:
+        kwargs["extra_body"] = extra
+
+    if stream and _env_bool("CHAT_STREAM_INCLUDE_USAGE", False):
+        kwargs["stream_options"] = {"include_usage": True}
+
+    return kwargs
+
+
+def _non_stream_completion(
+    client: OpenAI,
+    kwargs: dict[str, Any],
+    *,
+    include_reasoning: bool,
+) -> Iterator[str]:
+    kw = dict(kwargs)
+    kw["stream"] = False
+    kw.pop("stream_options", None)
+    resp = client.chat.completions.create(**kw)
+    choices = getattr(resp, "choices", None) or []
+    if not choices:
+        yield "[エラー] API 応答に choices がありません。"
+        return
+    msg = getattr(choices[0], "message", None)
+    parts = list(_iter_message_text_parts(msg, include_reasoning=include_reasoning))
+    if parts:
+        for p in parts:
+            yield p
+        return
+    first = choices[0]
+    yield (
+        "[エラー] メッセージ本文を取得できませんでした。"
+        f" finish_reason={getattr(first, 'finish_reason', None)!r}"
+    )
+
+
+def stream_answer(question: str, context: str) -> Iterator[str]:
+    """OpenAI Chat Completions 互換エンドポイント向け。ストリーム非対応・空ストリームのモデルは非ストリームにフォールバック。"""
+    client = _get_client()
+    use_stream = _env_bool("CHAT_STREAM", True)
+    fb = _env_bool("CHAT_STREAM_FALLBACK", True)
+    include_reasoning = _env_bool("CHAT_SHOW_REASONING", False)
+
+    if not use_stream:
+        kwargs = _build_chat_completion_kwargs(question, context, stream=False)
+        yield from _non_stream_completion(client, kwargs, include_reasoning=include_reasoning)
+        return
+
+    kwargs_stream = _build_chat_completion_kwargs(question, context, stream=True)
+    stream = client.chat.completions.create(**kwargs_stream)
+
+    got_text = False
+    try:
+        for chunk in stream:
+            for piece in _iter_chunk_text_parts(chunk, include_reasoning=include_reasoning):
+                got_text = True
+                yield piece
+    finally:
+        close = getattr(stream, "close", None)
+        if callable(close):
+            close()
+
+    if not got_text and fb:
+        kwargs_ns = _build_chat_completion_kwargs(question, context, stream=False)
+        yield from _non_stream_completion(client, kwargs_ns, include_reasoning=include_reasoning)
+
+
+async def run_rag_cli(question: str, top_k: int, *, quiet: bool = False) -> None:
+    params = next_devtools_stdio_params()
+
+    quiet = quiet or _env_bool("RAG_QUIET", False)
+    progress = ConsoleProgress(enabled=not quiet)
+
+    context = ""
+    paths: list[str] = []
+
+    try:
+        progress.start("next-devtools-mcp を起動・接続中 …")
+
+        async with stdio_client(params) as (read, write):
+            async with ClientSession(read, write) as session:
+                await session.initialize()
+                progress.phase("next-devtools-mcp と通信中 …", "")
+                try:
+                    await session.call_tool("init", {"project_path": os.getcwd()})
+                except Exception:
+                    pass
+
+                context, paths = await gather_context(
+                    session,
+                    question,
+                    top_k,
+                    on_phase=lambda main, detail="": progress.phase(main, detail),
+                )
+    finally:
+        progress.stop_clear()
+
+    # MCP 子プロセスを閉じてから LLM ストリーム（空 choices チャンク等の異常時にセッションを巻き込まない）
+    if paths:
+        print("取得したドキュメント path:", ", ".join(paths))
+        print()
+    if not context.strip():
+        print("関連ドキュメントを取得できませんでした。質問の表現を変えて再度お試しください。")
+        return
+
+    iterator = iter(stream_answer(question, context))
+    first = run_sync_blocking_with_spinner(
+        lambda: next(iterator, None),
+        out=sys.stderr,
+        main="モデル応答を待機中",
+        detail=_model_name(),
+        enabled=(not quiet),
+    )
+    if first is None:
+        print()
+        return
+
+    sys.stdout.write(first)
+    sys.stdout.flush()
+    for piece in iterator:
+        sys.stdout.write(piece)
+        sys.stdout.flush()
+    sys.stdout.write("\n")
+    sys.stdout.flush()

rag_nextjs/cli.py

@@ -0,0 +1,47 @@
+"""
+Next.js 公式ドキュメント RAG のエントリポイント（コンソールスクリプト / python -m と共有）。
+"""
+
+from __future__ import annotations
+
+import argparse
+import asyncio
+import os
+import sys
+
+from dotenv import load_dotenv
+
+
+def main() -> int:
+    load_dotenv()
+    if not os.getenv("NVIDIA_API_KEY") and not os.getenv("OPENAI_API_KEY"):
+        print(
+            "エラー: NVIDIA_API_KEY または OPENAI_API_KEY を .env に設定してください。",
+            file=sys.stderr,
+        )
+        return 1
+
+    parser = argparse.ArgumentParser(description="Next.js MCP 公式ドキュメント RAG")
+    parser.add_argument("question", help="Next.js についての質問")
+    parser.add_argument(
+        "--top-k",
+        type=int,
+        default=int(os.getenv("RAG_TOP_K", "4")),
+        help="取得するドキュメントページ数（既定 4）",
+    )
+    parser.add_argument(
+        "-q",
+        "--quiet",
+        action="store_true",
+        help="進捗・スピナーを出さない（スクリプトからの利用向け）",
+    )
+    args = parser.parse_args()
+
+    from rag_nextjs.answer import run_rag_cli
+
+    asyncio.run(run_rag_cli(args.question, top_k=args.top_k, quiet=args.quiet))
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

rag_nextjs/console_status.py

@@ -0,0 +1,171 @@
+"""
+ターミナル用スピナーと進捗表示（回答の標準出力とは分離し stderr に表示）。
+TTY でない場合はアニメーションの代わりに行単位のメッセージを出します。
+"""
+
+from __future__ import annotations
+
+import os
+import shutil
+import sys
+import threading
+from collections.abc import Callable
+from typing import IO, TypeVar
+
+_SPIN_FRAMES_ASCII = "|/-\\"
+_SPIN_FRAMES_UNICODE = "⠋⠙⠹⠸⠼⠴⠦⠇⠏"
+
+
+def _env_bool(name: str, default: bool) -> bool:
+    raw = os.getenv(name)
+    if raw is None or raw.strip() == "":
+        return default
+    return raw.strip().lower() in {"1", "true", "yes", "on"}
+
+
+def _use_unicode_spinner(stream: IO[str]) -> bool:
+    enc = getattr(stream, "encoding", None)
+    if isinstance(enc, str):
+        n = enc.lower().replace("_", "")
+        if n in {"utf-8", "utf8"}:
+            return True
+    return False
+
+
+def _format_status_line(prefix: str, main: str, detail: str) -> str:
+    cols = shutil.get_terminal_size(fallback=(88, 24)).columns
+    width = max(40, cols - 2)
+    body = f"{prefix}  {main}"
+    if detail.strip():
+        body = f"{body}  ·  {detail.strip()}"
+    if len(body) > width:
+        body = body[: width - 1] + "…"
+    return body
+
+
+class _SpinnerRunner(threading.Thread):
+    def __init__(self, out: IO[str], frames: str) -> None:
+        super().__init__(daemon=True)
+        self._out = out
+        self._frames = frames
+        self._stop = threading.Event()
+        self._lock = threading.Lock()
+        self._main = ""
+        self._detail = ""
+
+    def configure(self, main: str, detail: str = "") -> None:
+        with self._lock:
+            self._main = main
+            self._detail = detail
+
+    def halt(self) -> None:
+        self._stop.set()
+
+    def run(self) -> None:
+        i = 0
+        while not self._stop.wait(0.09):
+            with self._lock:
+                main, detail = self._main, self._detail
+            symbol = self._frames[i % len(self._frames)]
+            i += 1
+            line = _format_status_line(symbol, main, detail)
+            self._out.write(f"\r{line}\x1b[K")
+            self._out.flush()
+
+    def wipe(self) -> None:
+        self._out.write("\r\x1b[K")
+        self._out.flush()
+
+
+class ConsoleProgress:
+    """MCP・取得フェーズ用プログレス表示。"""
+
+    def __init__(self, *, enabled: bool, stream: IO[str] | None = None) -> None:
+        self._out: IO[str] = stream if stream is not None else sys.stderr
+        is_tty = getattr(self._out, "isatty", lambda: False)()
+        spinner_on = _env_bool("RAG_SPINNER", True)
+        force = _env_bool("RAG_SPINNER_FORCE", False)
+        frames = _SPIN_FRAMES_UNICODE if _use_unicode_spinner(self._out) else _SPIN_FRAMES_ASCII
+        self._enabled = enabled
+        self._animate = enabled and spinner_on and (is_tty or force)
+        self._runner: _SpinnerRunner | None = _SpinnerRunner(self._out, frames) if self._animate else None
+        self._started = False
+        self._last_plain: tuple[str, str] | None = None
+
+    def start(self, main: str, detail: str = "") -> None:
+        self.phase(main, detail)
+        if not self._enabled or self._started:
+            return
+        self._started = True
+        if self._runner:
+            self._runner.configure(main, detail)
+            self._runner.start()
+
+    def phase(self, main: str, detail: str = "") -> None:
+        if not self._enabled:
+            return
+        if self._runner and self._started and self._runner.is_alive():
+            self._runner.configure(main, detail)
+            return
+        if not self._animate:
+            key = (main, detail)
+            if key != self._last_plain:
+                self._last_plain = key
+                suffix = f" ({detail})" if detail.strip() else ""
+                self._out.write(f"[nextjs-rag]{suffix} {main}\n")
+                self._out.flush()
+
+    def stop_clear(self) -> None:
+        if not self._enabled:
+            return
+        if self._runner and self._started:
+            self._runner.halt()
+            self._runner.join(timeout=3.0)
+            self._runner.wipe()
+            self._started = False
+
+
+T = TypeVar("T")
+
+
+def run_sync_blocking_with_spinner(
+    fn: Callable[[], T],
+    *,
+    out: IO[str],
+    main: str,
+    detail: str = "",
+    enabled: bool,
+) -> T:
+    """同期ブロッキング処理（例: 最初のチャンクまで）を別スレッドのスピナー付きで包む。"""
+    if not enabled:
+        return fn()
+
+    is_tty = getattr(out, "isatty", lambda: False)()
+    spinner_on = _env_bool("RAG_SPINNER", True)
+    force = _env_bool("RAG_SPINNER_FORCE", False)
+    if not spinner_on or (not is_tty and not force):
+        out.write(f"[nextjs-rag] {main}{(' · ' + detail) if detail.strip() else ''}\n")
+        out.flush()
+        return fn()
+
+    frames = _SPIN_FRAMES_UNICODE if _use_unicode_spinner(out) else _SPIN_FRAMES_ASCII
+    stop = threading.Event()
+
+    def worker() -> None:
+        i = 0
+        while not stop.wait(0.09):
+            symbol = frames[i % len(frames)]
+            i += 1
+            line = _format_status_line(symbol, main, detail)
+            out.write(f"\r{line}\x1b[K")
+            out.flush()
+
+    th = threading.Thread(target=worker, daemon=True)
+    th.start()
+    try:
+        return fn()
+    finally:
+        stop.set()
+        th.join(timeout=3.0)
+        out.write("\r\x1b[K")
+        out.flush()

rag_nextjs/indexing.py

@@ -0,0 +1,49 @@
+from __future__ import annotations
+
+import re
+from collections import Counter
+
+# llms.txt 内の公式ドキュメント URL → nextjs_docs の path 引数
+_DOC_PATH = re.compile(r"https://nextjs\.org(/docs[/a-zA-Z0-9._\-]+)")
+
+
+def tokenize(text: str) -> list[str]:
+    return re.findall(r"[a-z0-9][a-z0-9_+\-/]*|[\u3040-\u30ff\u3400-\u9fff]+", text.lower())
+
+
+def _line_score(query_counter: Counter[str], line: str) -> float:
+    line_counter = Counter(tokenize(line))
+    if not line_counter:
+        return 0.0
+    dot = sum(line_counter[w] * query_counter.get(w, 0) for w in line_counter)
+    return dot / (sum(line_counter.values()) ** 0.5 + 1e-9)
+
+
+def extract_doc_paths_from_index(llms_index: str) -> list[tuple[str, str]]:
+    """各行について (行テキスト, 最初の /docs パス) を返す。"""
+    rows: list[tuple[str, str]] = []
+    for line in llms_index.splitlines():
+        m = _DOC_PATH.search(line)
+        if m:
+            rows.append((line, m.group(1)))
+    return rows
+
+
+def rank_doc_paths(query: str, llms_index: str, top_k: int) -> list[str]:
+    """クエリとインデックス行の簡易スコアリングで、重複のない path を最大 top_k 件選ぶ。"""
+    q = Counter(tokenize(query))
+    scored: list[tuple[float, str, str]] = []
+    for line, path in extract_doc_paths_from_index(llms_index):
+        scored.append((_line_score(q, line), path, line))
+
+    scored.sort(key=lambda x: x[0], reverse=True)
+    seen: set[str] = set()
+    out: list[str] = []
+    for _score, path, _line in scored:
+        if path in seen:
+            continue
+        seen.add(path)
+        out.append(path)
+        if len(out) >= top_k:
+            break
+    return out

rag_nextjs/mcp_docs.py

@@ -0,0 +1,71 @@
+from __future__ import annotations
+
+import json
+import os
+import shlex
+from typing import Any
+
+from pydantic import AnyUrl
+
+from mcp import StdioServerParameters
+from mcp.client.stdio import stdio_client
+
+
+def next_devtools_stdio_params() -> StdioServerParameters:
+    """環境変数 NEXTJS_MCP_COMMAND（既定: npx -y next-devtools-mcp）から起動コマンドを組み立てる。"""
+    raw = os.environ.get("NEXTJS_MCP_COMMAND", "npx -y next-devtools-mcp")
+    parts = shlex.split(raw)
+    if not parts:
+        raise ValueError("NEXTJS_MCP_COMMAND が空です")
+    return StdioServerParameters(command=parts[0], args=parts[1:])
+
+
+async def read_llms_index(session: ClientSession) -> str:
+    result = await session.read_resource(AnyUrl("nextjs-docs://llms-index"))
+    return "".join(c.text or "" for c in result.contents if c.text is not None)
+
+
+def _tool_text_payload(result: Any) -> str:
+    parts: list[str] = []
+    for block in getattr(result, "content", ()) or []:
+        if getattr(block, "type", None) == "text" and block.text:
+            parts.append(block.text)
+    return "\n".join(parts)
+
+
+async def fetch_nextjs_docs_page(session: ClientSession, path: str, anchor: str | None = None) -> dict[str, Any]:
+    args: dict[str, Any] = {"path": path}
+    if anchor:
+        args["anchor"] = anchor
+    result = await session.call_tool("nextjs_docs", args)
+    if result.isError:
+        payload = _tool_text_payload(result)
+        raise RuntimeError(f"nextjs_docs 失敗 path={path}: {payload}")
+
+    payload = _tool_text_payload(result)
+    data = json.loads(payload)
+    return data
+
+
+def format_fetched_pages(pages: list[dict[str, Any]], max_chars: int = 120_000) -> str:
+    chunks: list[str] = []
+    total = 0
+    for p in pages:
+        block = (
+            f"### path: {p.get('path')}\n"
+            f"url: {p.get('url')}\n\n"
+            f"{p.get('content', '')}"
+        )
+        if total + len(block) > max_chars:
+            remain = max_chars - total
+            if remain > 500:
+                chunks.append(block[:remain] + "\n\n[truncated]")
+            break
+        chunks.append(block)
+        total += len(block)
+    return "\n\n---\n\n".join(chunks)
+
+
+def next_devtools_client():
+    """stdio で next-devtools-mcp 子プロセスを起動する `stdio_client` 用コンテキストマネージャ。"""
+    return stdio_client(next_devtools_stdio_params())

rag_nextjs-1.0.0.dist-info/METADATA

@@ -0,0 +1,146 @@
+Metadata-Version: 2.4
+Name: rag-nextjs
+Version: 1.0.0
+Summary: Next.js 公式ドキュメント RAG CLI (next-devtools-mcp)
+Author-email: econanringo <151092188+econanringo@users.noreply.github.com>
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: openai>=1.0.0
+Requires-Dist: python-dotenv>=1.0.0
+Requires-Dist: mcp>=1.27.0
+Dynamic: license-file
+
+# Next.js 公式ドキュメント RAG（next-devtools-mcp）
+
+[公式の `next-devtools-mcp`](https://www.npmjs.com/package/next-devtools-mcp) を経由して Next.js の公式ドキュメントを取得し、その内容だけを根拠に質問に答える Python の CLI です。MCP の `nextjs_docs` と `nextjs-docs://llms-index` を使うため、検索対象は常に公式が配布するインデックス／本文に沿います。
+
+## 前提条件
+
+- **Python 3.10+**（プロジェクトでは 3.14 付近で動作確認）
+- **Node.js と `npx`**（MCP サーバーを `npx -y next-devtools-mcp` で起動するため）
+- **OpenAI 互換の Chat API** に渡せる API キー（例: [NVIDIA Integrate API](https://build.nvidia.com/)）
+
+## セットアップ
+
+```bash
+cd /path/to/rag
+python3 -m venv venv
+source venv/bin/activate   # Windows: venv\Scripts\activate
+pip install -r requirements.txt
+
+# （推奨）パスを通さなくても `rag-nextjs` で起動できるようにする
+pip install -e .
+```
+
+環境変数用に `.env` を用意します。
+
+```bash
+cp .env.example .env
+# .env を編集して API キーなどを設定
+```
+
+## 環境変数
+
+| 変数 | 必須 | 説明 |
+|------|------|------|
+| `NVIDIA_API_KEY` または `OPENAI_API_KEY` | はい | Chat API 用のキー |
+| `NVIDIA_OPENAI_BASE_URL` または `OPENAI_BASE_URL` | 推奨 | 互換 API のベース URL（NVIDIA 利用時は `.env.example` の URL を参照） |
+| `CHAT_MODEL` | いいえ | 既定: `nvidia/nemotron-3-nano-30b-a3b` |
+| `CHAT_TEMPERATURE` | いいえ | 既定: `0.2` |
+| `CHAT_EXTRA_BODY` | いいえ | `extra_body` に渡す JSON（例: Nemotron の推論オプション）。1 行の JSON 文字列 |
+| `CHAT_STREAM` | いいえ | `1` / `true` / `yes` / `on` でストリーミング（既定）。`0` など falsy で常に **非ストリーミング** のみ送信 |
+| `CHAT_STREAM_FALLBACK` | いいえ | 既定オン。ストリームを最後まで読んでも表示用テキストが一度も出ないとき、同じ入力で **非ストリーミングを 1 回** 試す |
+| `CHAT_SHOW_REASONING` | いいえ | 既定オフ。オンにすると `reasoning_content` 等の **推論テキストも標準出力**に流す。推論付きモデルの通常利用ではオフのままがよい |
+| `CHAT_MAX_TOKENS` | いいえ | 設定時のみ、`max_tokens` としてリクエストに付与 |
+| `CHAT_STREAM_INCLUDE_USAGE` | いいえ | オン時、ストリーミング要求に `stream_options: {"include_usage": true}` を付与（使用量を最終チャンクで受け取る用途） |
+| `CHAT_HTTP_TIMEOUT_SECONDS` | いいえ | 読み取り込みを含む HTTP タイムアウト（秒）。既定: `600` |
+| `OPENAI_TIMEOUT` または `HTTP_TIMEOUT` | いいえ | 上記と同目的。評価は **`OPENAI_TIMEOUT` → `HTTP_TIMEOUT` → `CHAT_HTTP_TIMEOUT_SECONDS`** の先勝ち |
+| `CHAT_HTTP_CONNECT_SECONDS` | いいえ | 接続確立までのタイムアウト（秒）。既定: `30`（遅い TLS でも切れにくくするため） |
+| `NEXTJS_MCP_COMMAND` | いいえ | MCP 起動コマンド（シェルと同様に空白区切り）。既定: `npx -y next-devtools-mcp` |
+| `RAG_TOP_K` | いいえ | `--top-k` の既定値（ページ取得数）。既定: `4` |
+| `RAG_SPINNER` | いいえ | 進捗行のブラケット／スピナーアニメーション。既定オン（`1`）。`0` でアニメーションを無効にし、必要なときのみプレーンな `[nextjs-rag]` 行になる |
+| `RAG_SPINNER_FORCE` | いいえ | `1` のとき **TTY でなくても** スピナーを描画（パイプ先が端末など特殊なとき向け）。既定オフ |
+| `RAG_QUIET` | いいえ | `1` のとき **進捗表示をすべて出さない**（`-q` と同様の効果）。コマンドラインの `-q` と併用すると、どちらかが真なら抑止されます |
+
+### Chat API（モデル差の吸収）
+
+OpenAI の Chat Completions 互換エンドポイントを前提にしています。プロバイダやモデルによって次のような差があります。
+
+- **ストリーム本文の欠落** … ゲートウェイ側でストリーミングの `delta` に本文が載らない場合がある → 既定では `CHAT_STREAM_FALLBACK` により **非ストリームで再試行**。
+- **推論フィールドの名称** … `reasoning_content` 以外に `reasoning` / `thinking` / `thought` などで返す API がある → **既定ではユーザー向け回答の `content`（と `refusal`）だけ**を標準出力へ出し、推論は出しません。デバッグで推論も見たいときは `CHAT_SHOW_REASONING=1` を指定してください。
+- **`content` の形** … 文字列だけでなく、`type: "text"` のブロックの配列で返す実装に対応します。
+
+大規模モデル（例: `nvidia/nemotron-3-super-120b-a12b`）は **最初のトークンまで数十秒〜数分かかる**ことがあります。ログに path が出たあと無言に見える場合は、しばらく待つか、`CHAT_HTTP_TIMEOUT_SECONDS` をさらに延ばしてください。ストリームが不安定なら `CHAT_STREAM=0` で非ストリーム固定にすると確実です。
+
+## 使い方
+
+開発用にリポジトリへ **editable install** した場合（`pip install -e .`）、どこからでも次のように起動できます。
+
+```bash
+rag-nextjs "App Router で Server Actions を使うには？"
+```
+
+リポジトリ直下では従来どおり次も同じです（`main.py` に shebang があるため `./main.py` も可）。
+
+```bash
+python main.py "App Router で Server Actions を使うには？"
+python -m rag_nextjs "App Router で Server Actions を使うには？"
+./main.py --help
+```
+
+取得するドキュメントページ数を増やす:
+
+```bash
+python main.py --top-k 5 "キャッシュの revalidate について"
+rag-nextjs --top-k 5 "キャッシュの revalidate について"
+```
+
+ヘルプ:
+
+```bash
+rag-nextjs --help
+python -m rag_nextjs --help
+python main.py --help
+```
+
+進捗は **標準エラー出力** に表示されます（処理中は `⠋ …` のようなアニメーションと、フェーズごとの説明。**ドキュメント取得中は何件目・どの path か**もここで更新されます。**モデル応答の初回トークン待ち**でも同様にスピナーを出します）。
+
+**標準出力**には、`取得したドキュメント path: …` と空行のあと、続けて **回答本文**がストリーミングされます。回答だけファイルに残したいときは `2>/dev/null` で stderr を捨てるか、path 行を除去する必要があります。
+
+```bash
+rag-nextjs "質問…" 2>/dev/null > answer.txt
+```
+
+ログを抑えたいときは `-q` / `--quiet` または環境変数 `RAG_QUIET=1` を指定してください。回答は **できる限りストリーミング** され、非対応時は `CHAT_STREAM_FALLBACK` で非ストリーム再試行があります（オフにすると `CHAT_STREAM_FALLBACK=0`）。
+
+## 動作の流れ（概要）
+
+1. 子プロセスで `next-devtools-mcp` を起動し、MCP セッションを確立する。
+2. リソース `nextjs-docs://llms-index` で公式インデックス（`llms.txt` 相当）を読む。
+3. 質問文とインデックス行の単語重なりで、関連しそうなドキュメント path を最大 `top_k` 件選ぶ。
+4. 各 path に対してツール `nextjs_docs` で本文を取得する。
+5. 取得本文をコンテキストとして、設定した Chat モデルに送り、回答を生成する。
+
+LLM は可能なら **ストリーミング** で逐次標準出力に表示し、(5) が本文ゼロで終わった場合のみ **同一プロンプトを非ストリーミングで再実行**します（`CHAT_STREAM_FALLBACK` がオンで、`CHAT_STREAM` がオンのとき）。進捗の見た目は stderr のスピナーが担います。
+
+インデックス説明が英語中心のため、日本語のみの短い質問だと取得 path が外れやすい場合があります。そのときは **`--top-k` を大きくする**か、ドキュメントで使われている用語（Server Actions、Route Handler など）を混ぜてみてください。
+
+この README の対象は **公式サイトのドキュメントを読む**経路です。
+
+- **`nextjs_docs` / `llms-index`** … nextjs.org の公式ドキュメントに基づいて答える（本プロジェクトが利用）。
+- **`nextjs_index` / `nextjs_call`** … ローカルで動いている **Next.js 16 以降の開発サーバー**の MCP（`/_next/mcp`）に接続し、コンパイルエラーやルート情報など**実行中アプリの状態**を調べる用途。
+
+アプリのランタイム診断まで Python から行いたい場合は、別途 dev サーバーを起動したうえで、そのツール群を呼び出す実装が必要になります。
+
+## トラブルシューティング
+
+- **`npx` が見つからない** … Node.js をインストールし、`NEXTJS_MCP_COMMAND` でフルパスの `npx` を指定するか、`PATH` を通してください。
+- **`next-devtools-mcp` の起動に失敗する** … ネットワークで npm レジストリに届くか、`npx -y next-devtools-mcp` を手動で実行できるか確認してください。
+- **関連ドキュメントが取得できない** … 質問を変える、`--top-k` を増やす、英語キーワードを足す。
+- **API エラー** … `.env` のベース URL・モデル名・キーが、そのプロバイダの OpenAI 互換仕様と一致しているか確認してください。
+- **回答がずっと出ない（path のあとで止まる）** … stderr に「モデル応答を待機中…」のスピナーが動いていれば、まだ初回トークン待ちです。大モデルは数十秒〜かかることがあります。標準エラーを見ていないターミナルでは `2>&1 | tee log.txt` などで確認してください。長すぎる場合は `CHAT_HTTP_TIMEOUT_SECONDS` を延ばす。ストリームが空のまま終わる API では `CHAT_STREAM_FALLBACK`（既定オン）で非ストリームに切り替わるはずなので、それでもダメなら `CHAT_STREAM=0` や `CHAT_MAX_TOKENS` / `CHAT_EXTRA_BODY` でプロバイダ必須パラメータを渡す。
+
+## ライセンス
+
+リポジトリ側のライセンスに従ってください。`next-devtools-mcp` 本体は npm パッケージのライセンス（MIT）が適用されます。

rag_nextjs-1.0.0.dist-info/RECORD

@@ -0,0 +1,13 @@
+rag_nextjs/__init__.py,sha256=ZJVW17NzZ9mw3vT64QSC9dCbzo1ilkrQTzeGPYs-J8I,108
+rag_nextjs/__main__.py,sha256=x5LayGsEAhfzBvm77EHraFSDOaIMzbgt4wQvRj57KrI,89
+rag_nextjs/answer.py,sha256=WSncmpwQVGm77IF5h3HRETugeAT29hr0vxJZ9C1JtfM,12288
+rag_nextjs/cli.py,sha256=CqzDT_zJte23pH1-3DQ-ix-1sGh399rR9uGSfT0O9bs,1324
+rag_nextjs/console_status.py,sha256=FPjT44aZbvk7Qs8Ycaclgx8K8-s-UjdhwaSysPbvZnI,5440
+rag_nextjs/indexing.py,sha256=hU_D0equKYN9EVpreN-LmWrMiHVm0Y2HNPmo5oS7eS8,1705
+rag_nextjs/mcp_docs.py,sha256=tWvMO__iS5GObkKSKiGMnM249x7nff5awUr59unap-w,2404
+rag_nextjs-1.0.0.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+rag_nextjs-1.0.0.dist-info/METADATA,sha256=UkdBfnarW7A2OJIL7l9lvcW-m7ALcxOvBv6PUJ7EH1Y,11011
+rag_nextjs-1.0.0.dist-info/WHEEL,sha256=aeYiig01lYGDzBgS8HxWXOg3uV61G9ijOsup-k9o1sk,91
+rag_nextjs-1.0.0.dist-info/entry_points.txt,sha256=idgnJ-AioDbACAeA6DDWkaJX09CaYeGkTM67YWchNws,51
+rag_nextjs-1.0.0.dist-info/top_level.txt,sha256=N6YvEt4c5EE6K5Gx9YUlZKdd9Au7B_57NZJsG25bho8,11
+rag_nextjs-1.0.0.dist-info/RECORD,,

rag_nextjs-1.0.0.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (82.0.1)
+Root-Is-Purelib: true
+Tag: py3-none-any
+

rag_nextjs-1.0.0.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+rag-nextjs = rag_nextjs.cli:main

rag_nextjs-1.0.0.dist-info/licenses/LICENSE

@@ -0,0 +1,201 @@
+                                 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.

rag_nextjs-1.0.0.dist-info/top_level.txt

@@ -0,0 +1 @@
+rag_nextjs