priorwork-mcp 0.1.0__tar.gz

priorwork_mcp-0.1.0/.gitignore

@@ -0,0 +1,58 @@
+# Ignore NeurIPS papers directory
+*_papers/
+*_paper_hashes.json
+paper_sections_index*
+
+# Data directories (PDFs, indexes, text store)
+data/
+tests/test_data/
+
+# Ignore environment files
+.env
+.env.*
+
+# Python virtual environment
+.venv/
+venv/
+ENV/
+
+# Python cache files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# Package build artifacts
+*.egg-info/
+build/
+dist/
+
+# Index files (large)
+*.zip
+*.index
+test_index*/
+paper_index_gte_temp/
+
+# Log files
+*.log
+
+# macOS system files
+.DS_Store
+.DS_Store?
+._*
+.Spotlight-V100
+.Trashes
+ehthumbs.db
+Thumbs.db # Exclude PDFs from git tracking
+tests/test_data/citation_test_pdfs/*.pdf
+
+# Temporary and generated files
+.coverage
+coverage.xml
+test_confidence_display.py
+
+# Test generation script (one-time use, requires full corpus)
+generate_citation_test.py
+
+# IDE settings
+.vscode/
+usage.db

priorwork_mcp-0.1.0/PKG-INFO

@@ -0,0 +1,77 @@
+Metadata-Version: 2.4
+Name: priorwork-mcp
+Version: 0.1.0
+Summary: MCP server for priorwork.fyi — academic ML paper search for Claude Code and Cursor
+Project-URL: Homepage, https://priorwork.fyi
+Project-URL: Repository, https://github.com/stefanmagureanu/Steference-review
+License: MIT
+Keywords: academic-search,machine-learning,mcp,model-context-protocol,papers,semantic-search
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.9
+Requires-Dist: httpx>=0.24.0
+Requires-Dist: mcp>=1.2.0
+Description-Content-Type: text/markdown
+
+# priorwork-mcp
+
+MCP server for [priorwork.fyi](https://priorwork.fyi) — semantic search over 114,000+ papers
+from major ML, NLP, and Computer Vision conferences (NeurIPS, CVPR, ICML, ICLR, ACL, EMNLP,
+ICCV, AISTATS, and more), directly from Claude Code, Cursor, or any MCP client.
+
+## Tools
+
+| Tool | Description |
+|---|---|
+| `search_papers` | Semantic search from a natural-language query — returns ranked papers with authors, venue, year, similarity score, download URL, and abstract snippet |
+| `search_by_pdf_url` | Give it a public PDF link (e.g. an arXiv PDF) and get semantically similar papers from the index |
+| `list_indexes` | List the available paper indexes and their sizes |
+
+## Setup
+
+### Claude Code
+
+```bash
+claude mcp add priorwork -- uvx priorwork-mcp
+```
+
+### Cursor / other MCP clients
+
+Add to your MCP configuration:
+
+```json
+{
+  "mcpServers": {
+    "priorwork": {
+      "command": "uvx",
+      "args": ["priorwork-mcp"]
+    }
+  }
+}
+```
+
+## Configuration
+
+Configuration is via environment variables (pass with `-e KEY=value` in `claude mcp add`,
+or an `"env"` block in JSON configs):
+
+| Variable | Required | Description |
+|---|---|---|
+| `PRIORWORK_API_KEY` | No | API key from priorwork.fyi (enables authenticated features) |
+| `PRIORWORK_API_URL` | No | Override the API base URL (default `https://priorwork.fyi`) |
+
+## Example
+
+> "Find prior work on spatiotemporal forecasting with graph neural networks"
+
+```
+1. Taming Local Effects in Graph-based Spatiotemporal Forecasting — Andrea Cini, et al. (2023 · NeurIPS)  [score: 0.802]
+2. FourierGNN: Rethinking Multivariate Time Series Forecasting from a Pure Graph Perspective — ...
+```
+
+## License
+
+MIT

priorwork_mcp-0.1.0/README.md

@@ -0,0 +1,59 @@
+# priorwork-mcp
+
+MCP server for [priorwork.fyi](https://priorwork.fyi) — semantic search over 114,000+ papers
+from major ML, NLP, and Computer Vision conferences (NeurIPS, CVPR, ICML, ICLR, ACL, EMNLP,
+ICCV, AISTATS, and more), directly from Claude Code, Cursor, or any MCP client.
+
+## Tools
+
+| Tool | Description |
+|---|---|
+| `search_papers` | Semantic search from a natural-language query — returns ranked papers with authors, venue, year, similarity score, download URL, and abstract snippet |
+| `search_by_pdf_url` | Give it a public PDF link (e.g. an arXiv PDF) and get semantically similar papers from the index |
+| `list_indexes` | List the available paper indexes and their sizes |
+
+## Setup
+
+### Claude Code
+
+```bash
+claude mcp add priorwork -- uvx priorwork-mcp
+```
+
+### Cursor / other MCP clients
+
+Add to your MCP configuration:
+
+```json
+{
+  "mcpServers": {
+    "priorwork": {
+      "command": "uvx",
+      "args": ["priorwork-mcp"]
+    }
+  }
+}
+```
+
+## Configuration
+
+Configuration is via environment variables (pass with `-e KEY=value` in `claude mcp add`,
+or an `"env"` block in JSON configs):
+
+| Variable | Required | Description |
+|---|---|---|
+| `PRIORWORK_API_KEY` | No | API key from priorwork.fyi (enables authenticated features) |
+| `PRIORWORK_API_URL` | No | Override the API base URL (default `https://priorwork.fyi`) |
+
+## Example
+
+> "Find prior work on spatiotemporal forecasting with graph neural networks"
+
+```
+1. Taming Local Effects in Graph-based Spatiotemporal Forecasting — Andrea Cini, et al. (2023 · NeurIPS)  [score: 0.802]
+2. FourierGNN: Rethinking Multivariate Time Series Forecasting from a Pure Graph Perspective — ...
+```
+
+## License
+
+MIT

priorwork_mcp-0.1.0/priorwork_mcp.py

@@ -0,0 +1,138 @@
+"""MCP server for priorwork.fyi — academic ML paper search.
+
+Set PRIORWORK_API_KEY to your API key from priorwork.fyi.
+Set PRIORWORK_API_URL to override the base URL (optional).
+"""
+
+import os
+import httpx
+from mcp.server.fastmcp import FastMCP
+
+API_URL = os.getenv("PRIORWORK_API_URL", "https://priorwork.fyi").rstrip("/")
+API_KEY = os.getenv("PRIORWORK_API_KEY", "")
+
+mcp = FastMCP("priorwork")
+
+
+def _headers() -> dict:
+    return {"X-API-Key": API_KEY} if API_KEY else {}
+
+
+def _format_results(results: list) -> str:
+    if not results:
+        return "No results found."
+    lines = []
+    for i, r in enumerate(results, 1):
+        title = r.get("title", "Unknown title")
+        authors = r.get("authors", "")
+        year = r.get("year", "")
+        venue = r.get("conference") or r.get("venue", "")
+        score = r.get("similarity", r.get("score", 0))
+        url = r.get("download_url") or r.get("url", "")
+        abstract = r.get("abstract", "")
+
+        header = f"{i}. {title}"
+        if authors:
+            header += f" — {authors}"
+        meta = " · ".join(filter(None, [str(year) if year else "", venue]))
+        if meta:
+            header += f" ({meta})"
+        header += f"  [score: {score:.3f}]"
+        lines.append(header)
+        if url:
+            lines.append(f"   {url}")
+        if abstract:
+            lines.append(f"   {abstract[:250].strip()}...")
+        lines.append("")
+    return "\n".join(lines).strip()
+
+
+def _flatten_results(data: dict) -> list:
+    # PDF search responses nest results per uploaded file: {"files": [{"results": [...]}]}
+    if isinstance(data.get("files"), list):
+        flat: list = []
+        for f in data["files"]:
+            if isinstance(f, dict):
+                flat.extend(f.get("results") or [])
+        return flat
+    results = data.get("results", [])
+    if isinstance(results, dict):
+        flat = []
+        for v in results.values():
+            flat.extend(v if isinstance(v, list) else [])
+        return flat
+    return results
+
+
+@mcp.tool()
+def search_papers(query: str, max_results: int = 10) -> str:
+    """Search for academic ML papers semantically related to a text query.
+
+    Returns a ranked list of papers with titles, authors, year, venue,
+    similarity score, download URL, and abstract snippet.
+
+    Args:
+        query: Natural language description of the research topic or question.
+        max_results: Number of results to return (1-10, default 10).
+    """
+    max_results = max(1, min(10, max_results))
+    with httpx.Client(timeout=30) as client:
+        resp = client.post(
+            f"{API_URL}/api/query",
+            json={"query": query, "max_results": max_results},
+            headers=_headers(),
+        )
+    resp.raise_for_status()
+    return _format_results(_flatten_results(resp.json()))
+
+
+@mcp.tool()
+def search_by_pdf_url(url: str, max_results: int = 10) -> str:
+    """Search for academic ML papers related to a paper at a given PDF URL.
+
+    Downloads the PDF from the URL (e.g. an arXiv PDF link), extracts its
+    content, and returns semantically similar papers from the index.
+
+    Args:
+        url: Public URL to a PDF file (http or https).
+        max_results: Number of results to return (1-10, default 10).
+    """
+    max_results = max(1, min(10, max_results))
+    with httpx.Client(timeout=60) as client:
+        resp = client.post(
+            f"{API_URL}/api/search/pdf-url",
+            json={"url": url, "max_results": max_results},
+            headers=_headers(),
+        )
+    resp.raise_for_status()
+    return _format_results(_flatten_results(resp.json()))
+
+
+@mcp.tool()
+def list_indexes() -> str:
+    """List the paper indexes available on priorwork.fyi and their paper counts."""
+    with httpx.Client(timeout=10) as client:
+        resp = client.get(f"{API_URL}/api/indexes", headers=_headers())
+    resp.raise_for_status()
+    data = resp.json()
+    indexes = data.get("indexes", [])
+    if not indexes:
+        return "No indexes available."
+    lines = []
+    for idx in indexes:
+        name = idx.get("name", "unknown")
+        count = idx.get("paper_count", "?")
+        provider = idx.get("provider", "")
+        line = f"  - {name}: {count:,} papers" if isinstance(count, int) else f"  - {name}: {count} papers"
+        if provider:
+            line += f" ({provider})"
+        lines.append(line)
+    return "Available indexes:\n" + "\n".join(lines)
+
+
+def main():
+    mcp.run()
+
+
+if __name__ == "__main__":
+    main()

priorwork_mcp-0.1.0/pyproject.toml

@@ -0,0 +1,36 @@
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[project]
+name = "priorwork-mcp"
+version = "0.1.0"
+description = "MCP server for priorwork.fyi — academic ML paper search for Claude Code and Cursor"
+readme = "README.md"
+requires-python = ">=3.9"
+license = {text = "MIT"}
+keywords = ["mcp", "model-context-protocol", "academic-search", "machine-learning", "papers", "semantic-search"]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Intended Audience :: Science/Research",
+    "License :: OSI Approved :: MIT License",
+    "Programming Language :: Python :: 3",
+    "Topic :: Scientific/Engineering :: Artificial Intelligence",
+]
+dependencies = [
+    "mcp>=1.2.0",
+    "httpx>=0.24.0",
+]
+
+[project.scripts]
+priorwork-mcp = "priorwork_mcp:main"
+
+[project.urls]
+Homepage = "https://priorwork.fyi"
+Repository = "https://github.com/stefanmagureanu/Steference-review"
+
+[tool.hatch.build.targets.wheel]
+packages = ["priorwork_mcp.py"]
+
+[tool.hatch.build.targets.sdist]
+include = ["priorwork_mcp.py", "README.md", "pyproject.toml"]