paper-wiki 0.2.0__tar.gz

paper_wiki-0.2.0/PKG-INFO

@@ -0,0 +1,101 @@
+Metadata-Version: 2.4
+Name: paper-wiki
+Version: 0.2.0
+Summary: Deep paper reading wiki: search, summarize, and analyze arXiv papers into a three-layer markdown wiki
+Author: mblank
+License-Expression: MIT
+Project-URL: Homepage, https://github.com/mblank5/paper-wiki
+Project-URL: Repository, https://github.com/mblank5/paper-wiki
+Keywords: arxiv,paper,wiki,summarization,cli
+Classifier: Development Status :: 3 - Alpha
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+Requires-Dist: click>=8.0
+Requires-Dist: openai>=1.0
+Requires-Dist: python-dotenv>=1.0
+Requires-Dist: rich>=13.0
+Requires-Dist: requests>=2.28
+
+# paper-wiki
+
+Deep paper reading wiki: search, summarize, and analyze arXiv papers into a three-layer markdown wiki.
+
+## Features
+
+- Search arXiv papers via [deepxiv](https://data.rag.ac.cn) CLI
+- Download paper PDFs
+- Generate per-section Chinese summaries using LLM
+- Produce a full paper analysis with 5 dimensions:
+  - ELI5 (小学生也能懂)
+  - Core innovations (核心创新点)
+  - Test datasets & metrics (测试集与指标)
+  - Core weaknesses (核心缺点)
+  - Future directions (后期演进方向)
+
+## Three-Layer Wiki
+
+```
+wiki/{arxiv_id}/
+├── README.md           # Layer 1: full summary + 5 analysis aspects
+├── sections/           # Layer 2: per-section Chinese summaries
+│   ├── 01_introduction.md
+│   ├── 02_method.md
+│   └── ...
+└── paper.pdf           # Layer 3: original PDF
+```
+
+## Install
+
+```bash
+pip install paper-wiki
+```
+
+### Prerequisites
+
+- [deepxiv](https://pypi.org/project/deepxiv/) CLI installed and configured (`pip install deepxiv`)
+- OpenAI-compatible API endpoint
+
+## Quick Start
+
+```bash
+# Configure API
+export OPENAI_API_KEY=sk-xxx
+export OPENAI_BASE_URL=https://api.openai.com/v1
+export OPENAI_MODEL=gpt-4o
+
+# Process a paper by arXiv ID
+paper-wiki process 2409.05591
+
+# Or by URL
+paper-wiki process https://arxiv.org/pdf/2604.27393
+
+# Search and interactively select
+paper-wiki search "RAG long context"
+
+# Search and auto-select first result
+paper-wiki search "transformer memory" --first
+
+# List processed papers
+paper-wiki list
+```
+
+## Configuration
+
+Create a `.env` file in your working directory:
+
+```env
+OPENAI_API_KEY=sk-xxx
+OPENAI_BASE_URL=https://api.openai.com/v1
+OPENAI_MODEL=gpt-4o
+```
+
+## License
+
+MIT

paper_wiki-0.2.0/README.md

@@ -0,0 +1,76 @@
+# paper-wiki
+
+Deep paper reading wiki: search, summarize, and analyze arXiv papers into a three-layer markdown wiki.
+
+## Features
+
+- Search arXiv papers via [deepxiv](https://data.rag.ac.cn) CLI
+- Download paper PDFs
+- Generate per-section Chinese summaries using LLM
+- Produce a full paper analysis with 5 dimensions:
+  - ELI5 (小学生也能懂)
+  - Core innovations (核心创新点)
+  - Test datasets & metrics (测试集与指标)
+  - Core weaknesses (核心缺点)
+  - Future directions (后期演进方向)
+
+## Three-Layer Wiki
+
+```
+wiki/{arxiv_id}/
+├── README.md           # Layer 1: full summary + 5 analysis aspects
+├── sections/           # Layer 2: per-section Chinese summaries
+│   ├── 01_introduction.md
+│   ├── 02_method.md
+│   └── ...
+└── paper.pdf           # Layer 3: original PDF
+```
+
+## Install
+
+```bash
+pip install paper-wiki
+```
+
+### Prerequisites
+
+- [deepxiv](https://pypi.org/project/deepxiv/) CLI installed and configured (`pip install deepxiv`)
+- OpenAI-compatible API endpoint
+
+## Quick Start
+
+```bash
+# Configure API
+export OPENAI_API_KEY=sk-xxx
+export OPENAI_BASE_URL=https://api.openai.com/v1
+export OPENAI_MODEL=gpt-4o
+
+# Process a paper by arXiv ID
+paper-wiki process 2409.05591
+
+# Or by URL
+paper-wiki process https://arxiv.org/pdf/2604.27393
+
+# Search and interactively select
+paper-wiki search "RAG long context"
+
+# Search and auto-select first result
+paper-wiki search "transformer memory" --first
+
+# List processed papers
+paper-wiki list
+```
+
+## Configuration
+
+Create a `.env` file in your working directory:
+
+```env
+OPENAI_API_KEY=sk-xxx
+OPENAI_BASE_URL=https://api.openai.com/v1
+OPENAI_MODEL=gpt-4o
+```
+
+## License
+
+MIT

paper_wiki-0.2.0/paper_wiki/__init__.py

@@ -0,0 +1 @@
+"""paper-wiki: Deep paper reading wiki generator."""

paper_wiki-0.2.0/paper_wiki/cli.py

@@ -0,0 +1,143 @@
+"""CLI entry point for paper-wiki."""
+
+import re
+
+import click
+from rich.console import Console
+from rich.table import Table
+
+from . import deepxiv_client as dx
+from .wiki import process_paper, list_papers
+
+console = Console()
+
+_ARXIV_ID_RE = re.compile(r"(?:^|/)(\d{4}\.\d{4,5})(?:$|\.pdf|v\d+)")
+_ARXIV_URL_RE = re.compile(r"arxiv\.org/(?:abs|pdf)/(\d{4}\.\d{4,5})")
+
+
+def _extract_arxiv_id(text: str) -> str | None:
+    """Extract arXiv ID from raw ID, URL, or return None."""
+    text = text.strip()
+    # Pure ID like 2604.27393
+    if re.fullmatch(r"\d{4}\.\d{4,5}", text):
+        return text
+    # URL like https://arxiv.org/abs/2604.27393 or .../pdf/2604.27393.pdf
+    m = _ARXIV_URL_RE.search(text)
+    if m:
+        return m.group(1)
+    return None
+
+
+@click.group()
+@click.option("--output-dir", default="wiki", envvar="PAPER_WIKI_DIR", help="Wiki output directory")
+@click.pass_context
+def cli(ctx, output_dir):
+    """paper-wiki: Deep paper reading wiki generator."""
+    ctx.ensure_object(dict)
+    ctx.obj["output_dir"] = output_dir
+
+
+@cli.command()
+@click.argument("arxiv_id")
+@click.option("--model", default=None, help="Override LLM model name")
+@click.pass_context
+def process(ctx, arxiv_id, model):
+    """Process an arXiv paper by ID or URL into a wiki page."""
+    output_dir = ctx.obj["output_dir"]
+    arxiv_id = _extract_arxiv_id(arxiv_id) or arxiv_id
+    console.print(f"[bold blue]Processing paper:[/] {arxiv_id}")
+    try:
+        paper_dir = process_paper(arxiv_id, output_dir=output_dir, model=model)
+        console.print(f"[bold green]Done![/] Wiki page: {paper_dir}/README.md")
+    except Exception as e:
+        console.print(f"[bold red]Error:[/] {e}")
+        raise SystemExit(1)
+
+
+@cli.command()
+@click.argument("query")
+@click.option("--limit", "-l", default=5, help="Number of search results")
+@click.option("--first", is_flag=True, help="Auto-select the first result")
+@click.option("--model", default=None, help="Override LLM model name")
+@click.pass_context
+def search(ctx, query, limit, first, model):
+    """Search for papers and process into wiki.
+
+    If QUERY looks like an arXiv ID (e.g. 2604.27393) or arXiv URL,
+    skip search and process directly.
+    """
+    output_dir = ctx.obj["output_dir"]
+
+    # Detect direct arXiv ID / URL
+    direct_id = _extract_arxiv_id(query)
+    if direct_id:
+        console.print(f"[bold blue]Detected arXiv ID:[/] {direct_id}, processing directly...")
+        paper_dir = process_paper(direct_id, output_dir=output_dir, model=model)
+        console.print(f"[bold green]Done![/] Wiki page: {paper_dir}/README.md")
+        return
+
+    console.print(f"[bold blue]Searching:[/] {query}")
+
+    try:
+        results = dx.search(query, limit=limit)
+    except Exception as e:
+        console.print(f"[bold red]Search error:[/] {e}")
+        raise SystemExit(1)
+
+    if not results:
+        console.print("[yellow]No results found.[/]")
+        return
+
+    # Display results
+    table = Table(title="Search Results")
+    table.add_column("#", style="bold")
+    table.add_column("arXiv ID")
+    table.add_column("Title")
+    table.add_column("Date")
+
+    for i, r in enumerate(results, 1):
+        arxiv_id = r.get("arxiv_id", r.get("id", ""))
+        title = r.get("title", "")[:60]
+        date = r.get("publish_at", "")[:10]
+        table.add_row(str(i), arxiv_id, title, date)
+
+    console.print(table)
+
+    if first:
+        selected = results[0]
+    else:
+        choice = click.prompt("Select a paper (#) or press Enter to cancel", default="")
+        if not choice:
+            return
+        try:
+            selected = results[int(choice) - 1]
+        except (ValueError, IndexError):
+            console.print("[red]Invalid selection.[/]")
+            return
+
+    arxiv_id = selected.get("arxiv_id", selected.get("id", ""))
+    console.print(f"\n[bold blue]Processing:[/] {selected.get('title', arxiv_id)}")
+    paper_dir = process_paper(arxiv_id, output_dir=output_dir, model=model)
+    console.print(f"[bold green]Done![/] Wiki page: {paper_dir}/README.md")
+
+
+@cli.command(name="list")
+@click.pass_context
+def list_cmd(ctx):
+    """List all processed papers in the wiki."""
+    output_dir = ctx.obj["output_dir"]
+    papers = list_papers(output_dir)
+
+    if not papers:
+        console.print(f"[yellow]No papers found in {output_dir}/[/]")
+        return
+
+    table = Table(title="Processed Papers")
+    table.add_column("arXiv ID", style="bold")
+    table.add_column("Title")
+    table.add_column("Path")
+
+    for p in papers:
+        table.add_row(p["id"], p["title"][:60], p["path"])
+
+    console.print(table)

paper_wiki-0.2.0/paper_wiki/deepxiv_client.py

@@ -0,0 +1,61 @@
+"""Wrapper around the deepxiv CLI for paper search and retrieval."""
+
+import json
+import subprocess
+
+import requests
+
+
+class DeepxivError(Exception):
+    """Error from deepxiv CLI."""
+
+
+def _run(cmd: list[str]) -> str:
+    result = subprocess.run(cmd, capture_output=True, text=True)
+    if result.returncode != 0:
+        msg = result.stderr.strip() or f"deepxiv exited with code {result.returncode}"
+        raise DeepxivError(f"{' '.join(cmd)}\n{msg}")
+    return result.stdout
+
+
+def search(query: str, limit: int = 10) -> list[dict]:
+    """Search papers on arXiv. Returns list of result dicts."""
+    out = _run(["deepxiv", "search", query, "-f", "json", "-l", str(limit)])
+    data = json.loads(out)
+    if isinstance(data, list):
+        return data
+    return data.get("result", data.get("results", []))
+
+
+def get_metadata(arxiv_id: str) -> dict:
+    """Get paper metadata (title, authors, sections list, etc.)."""
+    out = _run(["deepxiv", "paper", arxiv_id, "--head", "-f", "json"])
+    return json.loads(out)
+
+
+def get_brief(arxiv_id: str) -> dict:
+    """Get brief info (title, TLDR, keywords, citations, github_url)."""
+    out = _run(["deepxiv", "paper", arxiv_id, "--brief", "-f", "json"])
+    return json.loads(out)
+
+
+def get_section(arxiv_id: str, section_name: str) -> str:
+    """Get full markdown content of a specific section."""
+    return _run(["deepxiv", "paper", arxiv_id, "--section", section_name])
+
+
+def get_sections_list(arxiv_id: str) -> list[dict]:
+    """Extract sections list from paper metadata."""
+    meta = get_metadata(arxiv_id)
+    return meta.get("sections", [])
+
+
+def download_pdf(arxiv_id: str, dest_path: str) -> str:
+    """Download paper PDF to dest_path. Returns dest_path."""
+    meta = get_metadata(arxiv_id)
+    url = meta.get("src_url") or f"https://arxiv.org/pdf/{arxiv_id}"
+    resp = requests.get(url, timeout=120)
+    resp.raise_for_status()
+    with open(dest_path, "wb") as f:
+        f.write(resp.content)
+    return dest_path

paper_wiki-0.2.0/paper_wiki/llm.py

@@ -0,0 +1,47 @@
+"""OpenAI-compatible LLM client for paper analysis."""
+
+import os
+
+from dotenv import load_dotenv
+from openai import OpenAI
+
+load_dotenv()
+
+
+def _default_client() -> OpenAI:
+    return OpenAI(
+        api_key=os.getenv("OPENAI_API_KEY", ""),
+        base_url=os.getenv("OPENAI_BASE_URL", "https://api.openai.com/v1"),
+    )
+
+
+_client: OpenAI | None = None
+
+
+def get_client() -> OpenAI:
+    global _client
+    if _client is None:
+        _client = _default_client()
+    return _client
+
+
+def chat(
+    system_prompt: str,
+    user_content: str,
+    model: str | None = None,
+    temperature: float = 0.3,
+    max_tokens: int = 4096,
+) -> str:
+    """Single-turn chat completion. Returns assistant message content."""
+    client = get_client()
+    model = model or os.getenv("OPENAI_MODEL", "gpt-4o")
+    resp = client.chat.completions.create(
+        model=model,
+        temperature=temperature,
+        max_tokens=max_tokens,
+        messages=[
+            {"role": "system", "content": system_prompt},
+            {"role": "user", "content": user_content},
+        ],
+    )
+    return resp.choices[0].message.content or ""

paper_wiki-0.2.0/paper_wiki/prompts.py

@@ -0,0 +1,101 @@
+"""LLM prompt templates for paper analysis."""
+
+SYSTEM_ROLE = "你是一位资深AI/CS领域的研究论文分析专家，擅长用中文进行学术分析和通俗讲解。"
+
+SECTION_SUMMARY_PROMPT = SYSTEM_ROLE + """
+
+请对以下论文章节内容进行深度总结。要求：
+1. 用中文撰写，300-500字
+2. 准确概括该章节的核心内容、方法论和关键发现
+3. 保留重要的技术细节和数学概念（如有）
+4. 语言专业但清晰易懂
+
+章节名称：{section_name}
+
+章节内容：
+{content}
+"""
+
+FULL_SUMMARY_PROMPT = SYSTEM_ROLE + """
+
+以下是一篇论文各章节的中文总结。请基于这些章节总结，撰写一篇完整的论文概述。要求：
+1. 用中文撰写，800-1200字
+2. 涵盖论文的研究背景、核心方法、实验结果和结论
+3. 逻辑连贯，层次分明
+4. 面向具有一定学术背景的读者
+
+论文标题：{title}
+
+各章节总结：
+{sections}
+"""
+
+ELI5_PROMPT = SYSTEM_ROLE + """
+
+请用小学生（10岁左右）能听懂的方式，解释以下论文的核心内容。要求：
+1. 用中文撰写，200-400字
+2. 用生活中的比喻来解释复杂概念
+3. 不要使用专业术语，如果必须使用则要给出通俗解释
+4. 让孩子能理解这篇论文"解决了一个什么问题"以及"怎么解决的"
+
+论文标题：{title}
+
+论文概述：
+{summary}
+"""
+
+INNOVATIONS_PROMPT = SYSTEM_ROLE + """
+
+请分析以下论文的核心创新点。要求：
+1. 用中文列出3-5个核心创新点
+2. 每个创新点用简洁的一句话概括，然后附上简短说明
+3. 重点区分"真正的新贡献"和"已有方法的改进"
+4. 按重要性排序
+
+论文标题：{title}
+
+论文概述：
+{summary}
+"""
+
+TEST_METRICS_PROMPT = SYSTEM_ROLE + """
+
+请提取以下论文中使用的测试集和评估指标。要求：
+1. 列出论文使用的主要测试/评估数据集
+2. 列出主要评估指标及论文报告的具体数值
+3. 如有对比实验，列出基线方法和本文方法的结果对比
+4. 用中文撰写，但数据集名称和指标名称保留英文原文
+
+论文标题：{title}
+
+论文概述：
+{summary}
+"""
+
+WEAKNESSES_PROMPT = SYSTEM_ROLE + """
+
+请分析以下论文的核心缺点和局限性。要求：
+1. 用中文列出3-5个主要缺点
+2. 从以下维度分析：方法论的局限性、实验设计的不足、适用范围的限制、潜在的偏差
+3. 客观公正，既不过分苛刻也不回避问题
+4. 每个缺点给出具体论据支持
+
+论文标题：{title}
+
+论文概述：
+{summary}
+"""
+
+FUTURE_DIRECTIONS_PROMPT = SYSTEM_ROLE + """
+
+请基于以下论文的内容，分析其后期可能的研究演进方向。要求：
+1. 用中文列出3-5个未来研究方向
+2. 包括论文作者自己提到的方向和你基于领域知识的合理推测
+3. 每个方向说明：要解决什么问题、可能的思路、预期影响
+4. 结合当前AI领域的最新发展趋势
+
+论文标题：{title}
+
+论文概述：
+{summary}
+"""

paper_wiki-0.2.0/paper_wiki/wiki.py

@@ -0,0 +1,167 @@
+"""Wiki generation orchestration: fetch paper, summarize, assemble."""
+
+import re
+import os
+from pathlib import Path
+
+from rich.progress import Progress, SpinnerColumn, TextColumn
+
+from . import deepxiv_client as dx
+from . import llm
+from . import prompts
+
+
+def _slug(name: str) -> str:
+    """Convert section name to a file-system friendly slug."""
+    # Remove leading numbers like "1. " or "Appendix A "
+    s = re.sub(r"^(\d+\.?\s*|Appendix\s+[A-Z]\s*)", "", name)
+    s = s.strip().lower()
+    s = re.sub(r"[^a-z0-9]+", "_", s)
+    return s.strip("_") or "section"
+
+
+def process_paper(arxiv_id: str, output_dir: str = "wiki", model: str | None = None) -> Path:
+    """Full pipeline: fetch paper, summarize sections, generate wiki."""
+    paper_dir = Path(output_dir) / arxiv_id
+    sections_dir = paper_dir / "sections"
+    sections_dir.mkdir(parents=True, exist_ok=True)
+
+    with Progress(
+        SpinnerColumn(),
+        TextColumn("[progress.description]{task.description}"),
+        transient=True,
+    ) as progress:
+        # 1. Fetch metadata and brief
+        t = progress.add_task("Fetching paper metadata...", total=None)
+        metadata = dx.get_metadata(arxiv_id)
+        brief = dx.get_brief(arxiv_id)
+        title = metadata.get("title", arxiv_id)
+        raw_authors = metadata.get("authors", [])
+        if isinstance(raw_authors, str):
+            authors = raw_authors
+        elif raw_authors and isinstance(raw_authors[0], dict):
+            authors = ", ".join(a["name"] for a in raw_authors)
+        else:
+            authors = ", ".join(str(a) for a in raw_authors)
+        pub_date = metadata.get("publish_at", "")[:10]
+        sections_list = metadata.get("sections", [])
+        progress.update(t, completed=True)
+
+        # 2. Download PDF
+        progress.update(t, description="Downloading PDF...")
+        dx.download_pdf(arxiv_id, str(paper_dir / "paper.pdf"))
+        progress.update(t, completed=True)
+
+        # 3. Summarize each section
+        section_summaries: list[tuple[str, str, str]] = []  # (name, filename, summary)
+        for i, sec in enumerate(sections_list, 1):
+            sec_name = sec["name"]
+            progress.update(t, description=f"Summarizing section: {sec_name}...")
+            filename = f"{i:02d}_{_slug(sec_name)}.md"
+
+            content = dx.get_section(arxiv_id, sec_name)
+            summary = llm.chat(
+                prompts.SECTION_SUMMARY_PROMPT.format(section_name=sec_name, content=content),
+                "",
+                model=model,
+            )
+
+            sec_path = sections_dir / filename
+            sec_path.write_text(f"# {sec_name}\n\n{summary}\n", encoding="utf-8")
+            section_summaries.append((sec_name, filename, summary))
+
+        progress.update(t, completed=True)
+
+        # 4. Full paper summary from section summaries
+        progress.update(t, description="Generating full paper summary...")
+        all_sections_text = "\n\n---\n\n".join(
+            f"### {name}\n{summary}" for name, _, summary in section_summaries
+        )
+        full_summary = llm.chat(
+            prompts.FULL_SUMMARY_PROMPT.format(title=title, sections=all_sections_text),
+            "",
+            model=model,
+        )
+        progress.update(t, completed=True)
+
+        # 5. Five analysis aspects
+        aspects = {}
+        aspect_defs = [
+            ("eli5", "小学生也能懂", prompts.ELI5_PROMPT),
+            ("innovations", "核心创新点", prompts.INNOVATIONS_PROMPT),
+            ("test_metrics", "测试集与指标", prompts.TEST_METRICS_PROMPT),
+            ("weaknesses", "核心缺点", prompts.WEAKNESSES_PROMPT),
+            ("future", "后期演进方向", prompts.FUTURE_DIRECTIONS_PROMPT),
+        ]
+        for key, label, prompt_tmpl in aspect_defs:
+            progress.update(t, description=f"Analyzing: {label}...")
+            aspects[key] = llm.chat(
+                prompt_tmpl.format(title=title, summary=full_summary),
+                "",
+                model=model,
+            )
+        progress.update(t, completed=True)
+
+    # 6. Assemble README.md
+    section_links = "\n".join(
+        f"- [{name}](sections/{fname})" for name, fname, _ in section_summaries
+    )
+
+    readme = f"""# {title}
+
+> Authors: {authors} | Published: {pub_date} | arXiv: [{arxiv_id}](https://arxiv.org/abs/{arxiv_id})
+
+## 论文概述
+
+{full_summary}
+
+## 小学生也能懂
+
+{aspects["eli5"]}
+
+## 核心创新点
+
+{aspects["innovations"]}
+
+## 测试集与指标
+
+{aspects["test_metrics"]}
+
+## 核心缺点
+
+{aspects["weaknesses"]}
+
+## 后期演进方向
+
+{aspects["future"]}
+
+---
+
+## 章节详解
+
+{section_links}
+
+## 原文
+
+- [PDF](paper.pdf)
+"""
+
+    readme_path = paper_dir / "README.md"
+    readme_path.write_text(readme, encoding="utf-8")
+    return paper_dir
+
+
+def list_papers(output_dir: str = "wiki") -> list[dict]:
+    """List all processed papers in the wiki directory."""
+    wiki = Path(output_dir)
+    if not wiki.exists():
+        return []
+    papers = []
+    for d in sorted(wiki.iterdir()):
+        readme = d / "README.md"
+        if d.is_dir() and readme.exists():
+            # Read first line for title
+            first_line = readme.read_text(encoding="utf-8").split("\n")[0]
+            title = first_line.lstrip("# ").strip()
+            papers.append({"id": d.name, "title": title, "path": str(d)})
+    return papers

paper_wiki-0.2.0/paper_wiki.egg-info/PKG-INFO

@@ -0,0 +1,101 @@
+Metadata-Version: 2.4
+Name: paper-wiki
+Version: 0.2.0
+Summary: Deep paper reading wiki: search, summarize, and analyze arXiv papers into a three-layer markdown wiki
+Author: mblank
+License-Expression: MIT
+Project-URL: Homepage, https://github.com/mblank5/paper-wiki
+Project-URL: Repository, https://github.com/mblank5/paper-wiki
+Keywords: arxiv,paper,wiki,summarization,cli
+Classifier: Development Status :: 3 - Alpha
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+Requires-Dist: click>=8.0
+Requires-Dist: openai>=1.0
+Requires-Dist: python-dotenv>=1.0
+Requires-Dist: rich>=13.0
+Requires-Dist: requests>=2.28
+
+# paper-wiki
+
+Deep paper reading wiki: search, summarize, and analyze arXiv papers into a three-layer markdown wiki.
+
+## Features
+
+- Search arXiv papers via [deepxiv](https://data.rag.ac.cn) CLI
+- Download paper PDFs
+- Generate per-section Chinese summaries using LLM
+- Produce a full paper analysis with 5 dimensions:
+  - ELI5 (小学生也能懂)
+  - Core innovations (核心创新点)
+  - Test datasets & metrics (测试集与指标)
+  - Core weaknesses (核心缺点)
+  - Future directions (后期演进方向)
+
+## Three-Layer Wiki
+
+```
+wiki/{arxiv_id}/
+├── README.md           # Layer 1: full summary + 5 analysis aspects
+├── sections/           # Layer 2: per-section Chinese summaries
+│   ├── 01_introduction.md
+│   ├── 02_method.md
+│   └── ...
+└── paper.pdf           # Layer 3: original PDF
+```
+
+## Install
+
+```bash
+pip install paper-wiki
+```
+
+### Prerequisites
+
+- [deepxiv](https://pypi.org/project/deepxiv/) CLI installed and configured (`pip install deepxiv`)
+- OpenAI-compatible API endpoint
+
+## Quick Start
+
+```bash
+# Configure API
+export OPENAI_API_KEY=sk-xxx
+export OPENAI_BASE_URL=https://api.openai.com/v1
+export OPENAI_MODEL=gpt-4o
+
+# Process a paper by arXiv ID
+paper-wiki process 2409.05591
+
+# Or by URL
+paper-wiki process https://arxiv.org/pdf/2604.27393
+
+# Search and interactively select
+paper-wiki search "RAG long context"
+
+# Search and auto-select first result
+paper-wiki search "transformer memory" --first
+
+# List processed papers
+paper-wiki list
+```
+
+## Configuration
+
+Create a `.env` file in your working directory:
+
+```env
+OPENAI_API_KEY=sk-xxx
+OPENAI_BASE_URL=https://api.openai.com/v1
+OPENAI_MODEL=gpt-4o
+```
+
+## License
+
+MIT

paper_wiki-0.2.0/paper_wiki.egg-info/SOURCES.txt

@@ -0,0 +1,14 @@
+README.md
+pyproject.toml
+paper_wiki/__init__.py
+paper_wiki/cli.py
+paper_wiki/deepxiv_client.py
+paper_wiki/llm.py
+paper_wiki/prompts.py
+paper_wiki/wiki.py
+paper_wiki.egg-info/PKG-INFO
+paper_wiki.egg-info/SOURCES.txt
+paper_wiki.egg-info/dependency_links.txt
+paper_wiki.egg-info/entry_points.txt
+paper_wiki.egg-info/requires.txt
+paper_wiki.egg-info/top_level.txt

paper_wiki-0.2.0/paper_wiki.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

paper_wiki-0.2.0/paper_wiki.egg-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+paper-wiki = paper_wiki.cli:cli

paper_wiki-0.2.0/paper_wiki.egg-info/requires.txt

@@ -0,0 +1,5 @@
+click>=8.0
+openai>=1.0
+python-dotenv>=1.0
+rich>=13.0
+requests>=2.28

paper_wiki-0.2.0/paper_wiki.egg-info/top_level.txt

@@ -0,0 +1 @@
+paper_wiki

paper_wiki-0.2.0/pyproject.toml

@@ -0,0 +1,42 @@
+[project]
+name = "paper-wiki"
+version = "0.2.0"
+description = "Deep paper reading wiki: search, summarize, and analyze arXiv papers into a three-layer markdown wiki"
+readme = "README.md"
+license = "MIT"
+requires-python = ">=3.10"
+authors = [
+    { name = "mblank" },
+]
+keywords = ["arxiv", "paper", "wiki", "summarization", "cli"]
+classifiers = [
+    "Development Status :: 3 - Alpha",
+    "Environment :: Console",
+    "Intended Audience :: Science/Research",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Topic :: Scientific/Engineering :: Artificial Intelligence",
+]
+dependencies = [
+    "click>=8.0",
+    "openai>=1.0",
+    "python-dotenv>=1.0",
+    "rich>=13.0",
+    "requests>=2.28",
+]
+
+[project.urls]
+Homepage = "https://github.com/mblank5/paper-wiki"
+Repository = "https://github.com/mblank5/paper-wiki"
+
+[project.scripts]
+paper-wiki = "paper_wiki.cli:cli"
+
+[build-system]
+requires = ["setuptools>=68.0"]
+build-backend = "setuptools.build_meta"
+
+[tool.setuptools.packages.find]
+include = ["paper_wiki*"]

paper_wiki-0.2.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+