update

.python-version

@@ -0,0 +1 @@
+3.12

Makefile

@@ -0,0 +1,4 @@
+.PHONY: pre-deploy
+
+pre-deploy:
+	uv export --no-hashes --frozen > requirements.txt

app.py

@@ -0,0 +1,198 @@
+from __future__ import annotations
+
+import functools
+import os
+import sys
+from pathlib import Path
+from typing import Optional
+
+import gradio as gr
+import nltk
+
+
+def _resolve_open_provence_paths() -> Path:
+    """Locate the open_provence repository so we can import bundled utilities."""
+
+    env_hint = os.getenv("OPEN_PROVENCE_REPO")
+    if env_hint:
+        candidate = Path(env_hint).expanduser().resolve()
+        if (candidate / "scripts" / "open_provence_infer.py").exists():
+            return candidate
+
+    cwd = Path(__file__).resolve().parent
+    for parent in [cwd, *cwd.parents]:
+        candidate = parent / "open_provence"
+        if (candidate / "scripts" / "open_provence_infer.py").exists():
+            return candidate
+
+    default_candidate = Path("/home/hotchpotch/src/github.com/hotchpotch/open_provence").resolve()
+    if (default_candidate / "scripts" / "open_provence_infer.py").exists():
+        return default_candidate
+
+    raise RuntimeError(
+        "open_provence repository not found. Set OPEN_PROVENCE_REPO to the repository root."
+    )
+
+
+OPEN_PROVENCE_ROOT = _resolve_open_provence_paths()
+if str(OPEN_PROVENCE_ROOT) not in sys.path:
+    sys.path.insert(0, str(OPEN_PROVENCE_ROOT))
+
+
+from scripts.open_provence_infer import (  # type: ignore  # noqa: E402
+    InputRecord,
+    _infer_record,
+    load_model,
+)
+from open_provence.modeling_open_provence_standalone import (  # type: ignore  # noqa: E402
+    OpenProvenceModel,
+    resolve_inference_device,
+)
+
+
+def _ensure_nltk_punkt_resources() -> None:
+    """Download punkt resources on first run so inference never fails."""
+
+    for resource in ("punkt", "punkt_tab"):
+        try:
+            nltk.data.find(f"tokenizers/{resource}")
+        except LookupError:
+            nltk.download(resource, quiet=True)
+
+
+DEFAULT_MODEL = "hotchpotch/open-provence-reranker-japanese-xsmall-v1"
+DEFAULT_THRESHOLD = 0.1
+DEFAULT_BATCH_SIZE = 8
+SPLITTER_CHOICES = ("en", "ja")
+
+
+DEVICE_HINT = os.getenv("OPEN_PROVENCE_INFER_DEVICE")
+INFERENCE_DEVICE = resolve_inference_device(DEVICE_HINT)
+
+_ensure_nltk_punkt_resources()
+
+
+@functools.lru_cache(maxsize=4)
+def _load_model_cached(model_identifier: str) -> OpenProvenceModel:
+    resolved_identifier = model_identifier.strip() or DEFAULT_MODEL
+    return load_model(resolved_identifier, device=INFERENCE_DEVICE)
+
+
+def _format_summary(*, model_name: str, score: Optional[float], compression: Optional[float], duration: Optional[float]) -> str:
+    compression_percent = f"{compression:.1f}%" if isinstance(compression, (int, float)) else "N/A"
+    score_str = f"{score:.4f}" if isinstance(score, (int, float)) else "N/A"
+    duration_str = f"{duration:.2f}s" if isinstance(duration, (int, float)) else "N/A"
+    return (
+        f"**Model**: `{model_name}`\n"
+        f"- Score: {score_str}\n"
+        f"- Compression: {compression_percent}\n"
+        f"- Processing time: {duration_str}"
+    )
+
+
+def run_inference(model_identifier: str, splitter_language: str, question: str, title: str, text: str) -> tuple[str, str]:
+    question_value = (question or "").strip()
+    text_value = (text or "").strip()
+    title_value = (title or "").strip() or None
+
+    if not question_value:
+        raise gr.Error("質問文を入力してください。")
+    if not text_value:
+        raise gr.Error("本文テキストを入力してください。")
+
+    model_name = (model_identifier or DEFAULT_MODEL).strip() or DEFAULT_MODEL
+
+    try:
+        model = _load_model_cached(model_name)
+    except Exception as exc:  # pragma: no cover - user provided paths can fail
+        raise gr.Error(f"モデルの読み込みに失敗しました: {exc}") from exc
+
+    splitter_value = splitter_language if splitter_language in SPLITTER_CHOICES else None
+
+    record = InputRecord(question=question_value, text=text_value, title=title_value)
+    result = _infer_record(
+        model,
+        record,
+        threshold=DEFAULT_THRESHOLD,
+        language=splitter_value,
+        batch_size=DEFAULT_BATCH_SIZE,
+        first_line_as_title=False,
+        debug=False,
+    )
+
+    summary = _format_summary(
+        model_name=model_name,
+        score=result.score,
+        compression=result.compression_rate,
+        duration=result.total_seconds,
+    )
+
+    return summary, result.pruned_text
+
+
+def build_interface() -> gr.Blocks:
+    splitter_default = "en"
+
+    with gr.Blocks(title="Open Provence Gradio Demo") as demo:
+        gr.Markdown(
+            """
+            ## Open Provence Reranker Demo
+
+            デフォルトモデル: `hotchpotch/open-provence-reranker-japanese-xsmall-v1`  
+            CPUで動かすための最軽量モデル。英語と日本語、両方に対応しています。
+            """
+        )
+
+        with gr.Group():
+            model_input = gr.Textbox(
+                label="Model identifier",
+                value=DEFAULT_MODEL,
+                placeholder="例: hotchpotch/...",
+                info="CPUで動かすための最軽量モデル。英語と日本語、両方に対応しています。",
+            )
+            splitter_radio = gr.Radio(
+                label="Sentence splitter language",
+                choices=list(SPLITTER_CHOICES),
+                value=splitter_default,
+                info=(
+                    "文分割時に使用する言語を固定します。モデル付属の分割器を使いたい場合は"
+                    "事前に CLI で --splitter-lang を設定する構成を参照してください。"
+                ),
+            )
+
+        with gr.Group():
+            question_input = gr.Textbox(
+                label="Question",
+                placeholder="モデルに答えてほしい質問を入力してください。",
+                lines=2,
+            )
+            title_input = gr.Textbox(
+                label="Title (optional)",
+                placeholder="タイトルがあれば入力してください。",
+                lines=1,
+            )
+            text_input = gr.Textbox(
+                label="Text",
+                placeholder="要約・抽出したい本文を入力してください。",
+                lines=12,
+            )
+
+        run_button = gr.Button("Run inference", variant="primary")
+
+        summary_output = gr.Markdown(label="Summary")
+        pruned_output = gr.Textbox(label="Pruned Text", lines=12)
+
+        run_button.click(
+            fn=run_inference,
+            inputs=[model_input, splitter_radio, question_input, title_input, text_input],
+            outputs=[summary_output, pruned_output],
+        )
+
+    return demo
+
+
+demo = build_interface()
+
+
+if __name__ == "__main__":
+    demo.launch()

pyproject.toml

@@ -0,0 +1,13 @@
+[project]
+name = "open-provence-demo"
+version = "0.1.0"
+description = "Add your description here"
+readme = "README.md"
+requires-python = ">=3.12"
+dependencies = [
+    "fast-bunkai>=0.1.1",
+    "gradio==5.49.1",
+    "nltk>=3.9.2",
+    "torch>=2.9.0",
+    "transformers>=4.57.1",
+]

requirements.txt

@@ -0,0 +1,250 @@
+# This file was autogenerated by uv via the following command:
+#    uv export --no-hashes --frozen
+aiofiles==24.1.0
+    # via gradio
+annotated-doc==0.0.3
+    # via fastapi
+annotated-types==0.7.0
+    # via pydantic
+anyio==4.11.0
+    # via
+    #   gradio
+    #   httpx
+    #   starlette
+audioop-lts==0.2.2 ; python_full_version >= '3.13'
+    # via gradio
+brotli==1.1.0
+    # via gradio
+certifi==2025.10.5
+    # via
+    #   httpcore
+    #   httpx
+    #   requests
+charset-normalizer==3.4.4
+    # via requests
+click==8.3.0
+    # via
+    #   nltk
+    #   typer
+    #   uvicorn
+colorama==0.4.6 ; sys_platform == 'win32'
+    # via
+    #   click
+    #   tqdm
+fast-bunkai==0.1.1
+    # via open-provence-demo
+fastapi==0.120.1
+    # via gradio
+ffmpy==0.6.4
+    # via gradio
+filelock==3.20.0
+    # via
+    #   huggingface-hub
+    #   torch
+    #   transformers
+fsspec==2025.9.0
+    # via
+    #   gradio-client
+    #   huggingface-hub
+    #   torch
+gradio==5.49.1
+    # via open-provence-demo
+gradio-client==1.13.3
+    # via gradio
+groovy==0.1.2
+    # via gradio
+h11==0.16.0
+    # via
+    #   httpcore
+    #   uvicorn
+hf-xet==1.2.0 ; platform_machine == 'aarch64' or platform_machine == 'amd64' or platform_machine == 'arm64' or platform_machine == 'x86_64'
+    # via huggingface-hub
+httpcore==1.0.9
+    # via httpx
+httpx==0.28.1
+    # via
+    #   gradio
+    #   gradio-client
+    #   safehttpx
+huggingface-hub==0.36.0
+    # via
+    #   gradio
+    #   gradio-client
+    #   tokenizers
+    #   transformers
+idna==3.11
+    # via
+    #   anyio
+    #   httpx
+    #   requests
+janome==0.5.0
+    # via fast-bunkai
+jinja2==3.1.6
+    # via
+    #   gradio
+    #   torch
+joblib==1.5.2
+    # via nltk
+markdown-it-py==4.0.0
+    # via rich
+markupsafe==3.0.3
+    # via
+    #   gradio
+    #   jinja2
+mdurl==0.1.2
+    # via markdown-it-py
+mpmath==1.3.0
+    # via sympy
+networkx==3.5
+    # via torch
+nltk==3.9.2
+    # via open-provence-demo
+numpy==2.3.4
+    # via
+    #   gradio
+    #   pandas
+    #   transformers
+nvidia-cublas-cu12==12.8.4.1 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via
+    #   nvidia-cudnn-cu12
+    #   nvidia-cusolver-cu12
+    #   torch
+nvidia-cuda-cupti-cu12==12.8.90 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via torch
+nvidia-cuda-nvrtc-cu12==12.8.93 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via torch
+nvidia-cuda-runtime-cu12==12.8.90 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via torch
+nvidia-cudnn-cu12==9.10.2.21 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via torch
+nvidia-cufft-cu12==11.3.3.83 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via torch
+nvidia-cufile-cu12==1.13.1.3 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via torch
+nvidia-curand-cu12==10.3.9.90 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via torch
+nvidia-cusolver-cu12==11.7.3.90 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via torch
+nvidia-cusparse-cu12==12.5.8.93 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via
+    #   nvidia-cusolver-cu12
+    #   torch
+nvidia-cusparselt-cu12==0.7.1 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via torch
+nvidia-nccl-cu12==2.27.5 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via torch
+nvidia-nvjitlink-cu12==12.8.93 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via
+    #   nvidia-cufft-cu12
+    #   nvidia-cusolver-cu12
+    #   nvidia-cusparse-cu12
+    #   torch
+nvidia-nvshmem-cu12==3.3.20 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via torch
+nvidia-nvtx-cu12==12.8.90 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via torch
+orjson==3.11.4
+    # via gradio
+packaging==25.0
+    # via
+    #   gradio
+    #   gradio-client
+    #   huggingface-hub
+    #   transformers
+pandas==2.3.3
+    # via gradio
+pillow==11.3.0
+    # via gradio
+pydantic==2.11.10
+    # via
+    #   fastapi
+    #   gradio
+pydantic-core==2.33.2
+    # via pydantic
+pydub==0.25.1
+    # via gradio
+pygments==2.19.2
+    # via rich
+python-dateutil==2.9.0.post0
+    # via pandas
+python-multipart==0.0.20
+    # via gradio
+pytz==2025.2
+    # via pandas
+pyyaml==6.0.3
+    # via
+    #   gradio
+    #   huggingface-hub
+    #   transformers
+regex==2025.10.23
+    # via
+    #   nltk
+    #   transformers
+requests==2.32.5
+    # via
+    #   huggingface-hub
+    #   transformers
+rich==14.2.0
+    # via typer
+ruff==0.14.2
+    # via gradio
+safehttpx==0.1.7
+    # via gradio
+safetensors==0.6.2
+    # via transformers
+semantic-version==2.10.0
+    # via gradio
+setuptools==80.9.0
+    # via torch
+shellingham==1.5.4
+    # via typer
+six==1.17.0
+    # via python-dateutil
+sniffio==1.3.1
+    # via anyio
+starlette==0.48.0
+    # via
+    #   fastapi
+    #   gradio
+sympy==1.14.0
+    # via torch
+tokenizers==0.22.1
+    # via transformers
+tomlkit==0.13.3
+    # via gradio
+torch==2.9.0
+    # via open-provence-demo
+tqdm==4.67.1
+    # via
+    #   huggingface-hub
+    #   nltk
+    #   transformers
+transformers==4.57.1
+    # via open-provence-demo
+triton==3.5.0 ; platform_machine == 'x86_64' and sys_platform == 'linux'
+    # via torch
+typer==0.20.0
+    # via gradio
+typing-extensions==4.15.0
+    # via
+    #   anyio
+    #   fastapi
+    #   gradio
+    #   gradio-client
+    #   huggingface-hub
+    #   pydantic
+    #   pydantic-core
+    #   starlette
+    #   torch
+    #   typer
+    #   typing-inspection
+typing-inspection==0.4.2
+    # via pydantic
+tzdata==2025.2
+    # via pandas
+urllib3==2.5.0
+    # via requests
+uvicorn==0.38.0
+    # via gradio
+websockets==15.0.1
+    # via gradio-client