academia-mcp 1.8.1__tar.gz → 1.9.1__tar.gz

academia_mcp-1.9.1/PKG-INFO

@@ -0,0 +1,170 @@
+Metadata-Version: 2.4
+Name: academia-mcp
+Version: 1.9.1
+Summary: MCP server that provides different tools to search for scientific publications
+Author-email: Ilya Gusev <phoenixilya@gmail.com>
+Project-URL: Homepage, https://github.com/IlyaGusev/academia_mcp
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.12
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: mcp>=1.10.1
+Requires-Dist: xmltodict>=0.14.0
+Requires-Dist: types-xmltodict>=0.14.0
+Requires-Dist: requests>=2.32.0
+Requires-Dist: types-requests>=2.32.0
+Requires-Dist: pypdf>=5.1.0
+Requires-Dist: beautifulsoup4>=4.12.0
+Requires-Dist: types-beautifulsoup4>=4.12.0
+Requires-Dist: markdownify==0.14.1
+Requires-Dist: acl-anthology==0.5.2
+Requires-Dist: markdown==3.7.0
+Requires-Dist: types-markdown==3.7.0.20250322
+Requires-Dist: huggingface-hub>=0.32.4
+Requires-Dist: fire>=0.7.0
+Requires-Dist: openai>=1.97.1
+Requires-Dist: jinja2>=3.1.6
+Requires-Dist: datasets>=4.0.0
+Requires-Dist: pymupdf>=1.26.4
+Requires-Dist: pillow>=11.3.0
+Requires-Dist: pydantic-settings>=2.6.0
+Dynamic: license-file
+
+# Academia MCP
+
+[![PyPI](https://img.shields.io/pypi/v/academia-mcp?label=PyPI%20package)](https://pypi.org/project/academia-mcp/)
+[![CI](https://github.com/IlyaGusev/academia_mcp/actions/workflows/python.yml/badge.svg)](https://github.com/IlyaGusev/academia_mcp/actions/workflows/python.yml)
+[![License](https://img.shields.io/github/license/IlyaGusev/academia_mcp)](LICENSE)
+[![smithery badge](https://smithery.ai/badge/@IlyaGusev/academia_mcp)](https://smithery.ai/server/@IlyaGusev/academia_mcp)
+[![Verified on MseeP](https://mseep.ai/badge.svg)](https://mseep.ai/app/e818878b-c3a6-4b3d-a5b4-e54dcd1f1fed)
+
+MCP server with tools to search, fetch, analyze, and report on scientific papers and datasets.
+
+### Features
+- ArXiv search and download
+- ACL Anthology search
+- Hugging Face datasets search
+- Semantic Scholar citations and references
+- Web search via Exa, Brave, or Tavily
+- Web page crawler, LaTeX compilation, PDF reading
+- Optional LLM-powered tools for document QA and research proposal workflows
+
+### Requirements
+- Python 3.12+
+
+### Install
+- Using pip (end users):
+```bash
+pip3 install academia-mcp
+```
+
+- For development (uv + Makefile):
+```bash
+uv venv .venv
+make install
+```
+
+### Quickstart
+- Run over HTTP (default transport):
+```bash
+uv run -m academia_mcp --transport streamable-http
+```
+
+- Run over stdio (for local MCP clients like Claude Desktop):
+```bash
+python -m academia_mcp --transport stdio
+```
+
+Notes:
+- Transports: `stdio`, `sse`, `streamable-http`.
+- `host`/`port` are used for HTTP transports; ignored for `stdio`. Default port is `5056` (or `PORT`).
+
+### Claude Desktop config
+```json
+{
+  "mcpServers": {
+    "academia": {
+      "command": "python3",
+      "args": [
+        "-m",
+        "academia_mcp",
+        "--transport",
+        "stdio"
+      ]
+    }
+  }
+}
+```
+
+### Available tools (one-liners)
+- `arxiv_search`: Query arXiv with field-specific queries and filters.
+- `arxiv_download`: Fetch a paper by ID and convert to structured text (HTML/PDF modes).
+- `anthology_search`: Search ACL Anthology with fielded queries and optional date filtering.
+- `hf_datasets_search`: Find Hugging Face datasets with filters and sorting.
+- `s2_get_citations`: List papers citing a given arXiv paper (Semantic Scholar Graph).
+- `s2_get_references`: List papers referenced by a given arXiv paper.
+- `visit_webpage`: Fetch and normalize a web page.
+- `web_search`: Unified search wrapper; available when at least one of Exa/Brave/Tavily keys is set.
+- `exa_web_search`, `brave_web_search`, `tavily_web_search`: Provider-specific search.
+- `get_latex_templates_list`, `get_latex_template`: Enumerate and fetch built-in LaTeX templates.
+- `compile_latex`: Compile LaTeX to PDF in `WORKSPACE_DIR`.
+- `read_pdf`: Extract text per page from a PDF.
+- `download_pdf_paper`, `review_pdf_paper`: Download and optionally review PDFs (requires LLM + workspace).
+- `document_qa`: Answer questions over provided document chunks (requires LLM).
+- `extract_bitflip_info`, `generate_research_proposals`, `score_research_proposals`: Research proposal helpers (requires LLM).
+
+Availability notes:
+- Set `WORKSPACE_DIR` to enable `compile_latex`, `read_pdf`, `download_pdf_paper`, and `review_pdf_paper`.
+- Set `OPENROUTER_API_KEY` to enable LLM tools (`document_qa`, `review_pdf_paper`, and bitflip tools).
+- Set one or more of `EXA_API_KEY`, `BRAVE_API_KEY`, `TAVILY_API_KEY` to enable `web_search` and provider tools.
+
+### Environment variables
+Set as needed depending on which tools you use:
+
+- `OPENROUTER_API_KEY`: required for LLM-related tools.
+- `BASE_URL`: override OpenRouter base URL.
+- `DOCUMENT_QA_MODEL_NAME`: override default model for `document_qa`.
+- `BITFLIP_MODEL_NAME`: override default model for bitflip tools.
+- `TAVILY_API_KEY`: enables Tavily in `web_search`.
+- `EXA_API_KEY`: enables Exa in `web_search` and `visit_webpage`.
+- `BRAVE_API_KEY`: enables Brave in `web_search`.
+- `WORKSPACE_DIR`: directory for generated files (PDFs, temp artifacts).
+- `PORT`: HTTP port (default `5056`).
+
+You can put these in a `.env` file in the project root.
+
+### Docker
+Build the image:
+```bash
+docker build -t academia_mcp .
+```
+
+Run the server (HTTP):
+```bash
+docker run --rm -p 5056:5056 \
+  -e PORT=5056 \
+  -e OPENROUTER_API_KEY=your_key_here \
+  -e WORKSPACE_DIR=/workspace \
+  -v "$PWD/workdir:/workspace" \
+  academia_mcp
+```
+
+Or use existing image: `phoenix120/academia_mcp`
+
+### Examples
+- [Comprehensive report screencast (YouTube)](https://www.youtube.com/watch?v=4bweqQcN6w8)
+- [Single paper screencast (YouTube)](https://www.youtube.com/watch?v=IAAPMptJ5k8)
+
+### Makefile targets
+- `make install`: install the package in editable mode with uv
+- `make validate`: run black, flake8, and mypy (strict)
+- `make test`: run the test suite with pytest
+- `make publish`: build and publish using uv
+
+### LaTeX/PDF requirements
+Only needed for LaTeX/PDF tools. Ensure a LaTeX distribution is installed and `pdflatex` is on PATH, as well as `latexmk`. On Debian/Ubuntu:
+```bash
+sudo apt install texlive-latex-base texlive-fonts-recommended texlive-latex-extra texlive-science latexmk
+```

academia_mcp-1.9.1/README.md

@@ -0,0 +1,136 @@
+# Academia MCP
+
+[![PyPI](https://img.shields.io/pypi/v/academia-mcp?label=PyPI%20package)](https://pypi.org/project/academia-mcp/)
+[![CI](https://github.com/IlyaGusev/academia_mcp/actions/workflows/python.yml/badge.svg)](https://github.com/IlyaGusev/academia_mcp/actions/workflows/python.yml)
+[![License](https://img.shields.io/github/license/IlyaGusev/academia_mcp)](LICENSE)
+[![smithery badge](https://smithery.ai/badge/@IlyaGusev/academia_mcp)](https://smithery.ai/server/@IlyaGusev/academia_mcp)
+[![Verified on MseeP](https://mseep.ai/badge.svg)](https://mseep.ai/app/e818878b-c3a6-4b3d-a5b4-e54dcd1f1fed)
+
+MCP server with tools to search, fetch, analyze, and report on scientific papers and datasets.
+
+### Features
+- ArXiv search and download
+- ACL Anthology search
+- Hugging Face datasets search
+- Semantic Scholar citations and references
+- Web search via Exa, Brave, or Tavily
+- Web page crawler, LaTeX compilation, PDF reading
+- Optional LLM-powered tools for document QA and research proposal workflows
+
+### Requirements
+- Python 3.12+
+
+### Install
+- Using pip (end users):
+```bash
+pip3 install academia-mcp
+```
+
+- For development (uv + Makefile):
+```bash
+uv venv .venv
+make install
+```
+
+### Quickstart
+- Run over HTTP (default transport):
+```bash
+uv run -m academia_mcp --transport streamable-http
+```
+
+- Run over stdio (for local MCP clients like Claude Desktop):
+```bash
+python -m academia_mcp --transport stdio
+```
+
+Notes:
+- Transports: `stdio`, `sse`, `streamable-http`.
+- `host`/`port` are used for HTTP transports; ignored for `stdio`. Default port is `5056` (or `PORT`).
+
+### Claude Desktop config
+```json
+{
+  "mcpServers": {
+    "academia": {
+      "command": "python3",
+      "args": [
+        "-m",
+        "academia_mcp",
+        "--transport",
+        "stdio"
+      ]
+    }
+  }
+}
+```
+
+### Available tools (one-liners)
+- `arxiv_search`: Query arXiv with field-specific queries and filters.
+- `arxiv_download`: Fetch a paper by ID and convert to structured text (HTML/PDF modes).
+- `anthology_search`: Search ACL Anthology with fielded queries and optional date filtering.
+- `hf_datasets_search`: Find Hugging Face datasets with filters and sorting.
+- `s2_get_citations`: List papers citing a given arXiv paper (Semantic Scholar Graph).
+- `s2_get_references`: List papers referenced by a given arXiv paper.
+- `visit_webpage`: Fetch and normalize a web page.
+- `web_search`: Unified search wrapper; available when at least one of Exa/Brave/Tavily keys is set.
+- `exa_web_search`, `brave_web_search`, `tavily_web_search`: Provider-specific search.
+- `get_latex_templates_list`, `get_latex_template`: Enumerate and fetch built-in LaTeX templates.
+- `compile_latex`: Compile LaTeX to PDF in `WORKSPACE_DIR`.
+- `read_pdf`: Extract text per page from a PDF.
+- `download_pdf_paper`, `review_pdf_paper`: Download and optionally review PDFs (requires LLM + workspace).
+- `document_qa`: Answer questions over provided document chunks (requires LLM).
+- `extract_bitflip_info`, `generate_research_proposals`, `score_research_proposals`: Research proposal helpers (requires LLM).
+
+Availability notes:
+- Set `WORKSPACE_DIR` to enable `compile_latex`, `read_pdf`, `download_pdf_paper`, and `review_pdf_paper`.
+- Set `OPENROUTER_API_KEY` to enable LLM tools (`document_qa`, `review_pdf_paper`, and bitflip tools).
+- Set one or more of `EXA_API_KEY`, `BRAVE_API_KEY`, `TAVILY_API_KEY` to enable `web_search` and provider tools.
+
+### Environment variables
+Set as needed depending on which tools you use:
+
+- `OPENROUTER_API_KEY`: required for LLM-related tools.
+- `BASE_URL`: override OpenRouter base URL.
+- `DOCUMENT_QA_MODEL_NAME`: override default model for `document_qa`.
+- `BITFLIP_MODEL_NAME`: override default model for bitflip tools.
+- `TAVILY_API_KEY`: enables Tavily in `web_search`.
+- `EXA_API_KEY`: enables Exa in `web_search` and `visit_webpage`.
+- `BRAVE_API_KEY`: enables Brave in `web_search`.
+- `WORKSPACE_DIR`: directory for generated files (PDFs, temp artifacts).
+- `PORT`: HTTP port (default `5056`).
+
+You can put these in a `.env` file in the project root.
+
+### Docker
+Build the image:
+```bash
+docker build -t academia_mcp .
+```
+
+Run the server (HTTP):
+```bash
+docker run --rm -p 5056:5056 \
+  -e PORT=5056 \
+  -e OPENROUTER_API_KEY=your_key_here \
+  -e WORKSPACE_DIR=/workspace \
+  -v "$PWD/workdir:/workspace" \
+  academia_mcp
+```
+
+Or use existing image: `phoenix120/academia_mcp`
+
+### Examples
+- [Comprehensive report screencast (YouTube)](https://www.youtube.com/watch?v=4bweqQcN6w8)
+- [Single paper screencast (YouTube)](https://www.youtube.com/watch?v=IAAPMptJ5k8)
+
+### Makefile targets
+- `make install`: install the package in editable mode with uv
+- `make validate`: run black, flake8, and mypy (strict)
+- `make test`: run the test suite with pytest
+- `make publish`: build and publish using uv
+
+### LaTeX/PDF requirements
+Only needed for LaTeX/PDF tools. Ensure a LaTeX distribution is installed and `pdflatex` is on PATH, as well as `latexmk`. On Debian/Ubuntu:
+```bash
+sudo apt install texlive-latex-base texlive-fonts-recommended texlive-latex-extra texlive-science latexmk
+```

{academia_mcp-1.8.1 → academia_mcp-1.9.1}/academia_mcp/files.py

@@ -1,7 +1,8 @@
-import os
 from typing import Optional
 from pathlib import Path
 
+from academia_mcp.settings import settings
+
 DIR_PATH = Path(__file__).parent
 ROOT_PATH = DIR_PATH.parent
 DEFAULT_WORKSPACE_DIR_PATH: Path = DIR_PATH / "workdir"
@@ -14,7 +15,7 @@ class WorkspaceDirectory:
     @classmethod
     def get_dir(cls) -> Path:
         if cls.workspace_dir is None:
-            return Path(os.getenv("WORKSPACE_DIR", DEFAULT_WORKSPACE_DIR_PATH))
+            return Path(settings.WORKSPACE_DIR)
         return cls.workspace_dir
 
     @classmethod

{academia_mcp-1.8.1 → academia_mcp-1.9.1}/academia_mcp/llm.py

@@ -1,10 +1,11 @@
-import os
 from typing import List, Dict, Any
 
 from pydantic import BaseModel
 from openai import AsyncOpenAI
 from openai.types.chat.chat_completion_message import ChatCompletionMessage
 
+from academia_mcp.settings import settings
+
 
 class ChatMessage(BaseModel):  # type: ignore
     role: str
@@ -15,9 +16,9 @@ ChatMessages = List[ChatMessage]
 
 
 async def llm_acall(model_name: str, messages: ChatMessages, **kwargs: Any) -> str:
-    key = os.getenv("OPENROUTER_API_KEY", "")
+    key = settings.OPENROUTER_API_KEY
     assert key, "Please set OPENROUTER_API_KEY in the environment variables"
-    base_url = os.getenv("BASE_URL", "https://openrouter.ai/api/v1")
+    base_url = settings.BASE_URL
 
     client = AsyncOpenAI(base_url=base_url, api_key=key)
     response: ChatCompletionMessage = (

academia_mcp-1.9.1/academia_mcp/server.py

@@ -0,0 +1,133 @@
+import socket
+import logging
+from logging.config import dictConfig
+from typing import Optional, Literal
+
+import fire  # type: ignore
+from mcp.server.fastmcp import FastMCP
+from uvicorn.config import LOGGING_CONFIG as UVICORN_LOGGING_CONFIG
+
+from academia_mcp.settings import settings
+from academia_mcp.tools.arxiv_search import arxiv_search
+from academia_mcp.tools.arxiv_download import arxiv_download
+from academia_mcp.tools.s2 import (
+    s2_get_citations,
+    s2_get_references,
+    s2_corpus_id_from_arxiv_id,
+    s2_get_info,
+)
+from academia_mcp.tools.hf_datasets_search import hf_datasets_search
+from academia_mcp.tools.anthology_search import anthology_search
+from academia_mcp.tools.document_qa import document_qa
+from academia_mcp.tools.latex import (
+    compile_latex,
+    get_latex_template,
+    get_latex_templates_list,
+    read_pdf,
+)
+from academia_mcp.tools.web_search import (
+    web_search,
+    tavily_web_search,
+    exa_web_search,
+    brave_web_search,
+)
+from academia_mcp.tools.visit_webpage import visit_webpage
+from academia_mcp.tools.bitflip import (
+    extract_bitflip_info,
+    generate_research_proposals,
+    score_research_proposals,
+)
+from academia_mcp.tools.review import review_pdf_paper, download_pdf_paper
+
+
+def configure_uvicorn_style_logging(level: int = logging.INFO) -> None:
+    config = {**UVICORN_LOGGING_CONFIG}
+    config["disable_existing_loggers"] = False
+    config["root"] = {"handlers": ["default"], "level": logging.getLevelName(level)}
+    dictConfig(config)
+
+
+def find_free_port() -> int:
+    for port in range(5000, 6001):
+        try:
+            with socket.socket() as s:
+                s.bind(("", port))
+                return port
+        except Exception:
+            continue
+    raise RuntimeError("No free port in range 5000-6000 found")
+
+
+def run(
+    host: str = "0.0.0.0",
+    port: Optional[int] = None,
+    mount_path: str = "/",
+    streamable_http_path: str = "/mcp",
+    transport: Literal["stdio", "sse", "streamable-http"] = "streamable-http",
+    disable_web_search_tools: bool = False,
+    disable_llm_tools: bool = False,
+) -> None:
+    configure_uvicorn_style_logging()
+    server = FastMCP(
+        "Academia MCP",
+        stateless_http=True,
+        streamable_http_path=streamable_http_path,
+        mount_path=mount_path,
+    )
+    logger = logging.getLogger(__name__)
+
+    server.add_tool(arxiv_search)
+    server.add_tool(arxiv_download)
+    server.add_tool(s2_get_citations)
+    server.add_tool(s2_get_references)
+    server.add_tool(s2_corpus_id_from_arxiv_id)
+    server.add_tool(s2_get_info)
+    server.add_tool(hf_datasets_search)
+    server.add_tool(anthology_search)
+    server.add_tool(get_latex_template)
+    server.add_tool(get_latex_templates_list)
+    server.add_tool(visit_webpage)
+
+    if settings.WORKSPACE_DIR:
+        server.add_tool(compile_latex)
+        server.add_tool(download_pdf_paper)
+        server.add_tool(read_pdf)
+    else:
+        logger.warning(
+            "WORKSPACE_DIR is not set, compile_latex/download_pdf_paper/read_pdf will not be available!"
+        )
+
+    if not disable_web_search_tools:
+        if settings.TAVILY_API_KEY:
+            server.add_tool(tavily_web_search)
+        if settings.EXA_API_KEY:
+            server.add_tool(exa_web_search)
+        if settings.BRAVE_API_KEY:
+            server.add_tool(brave_web_search)
+        if settings.EXA_API_KEY or settings.BRAVE_API_KEY or settings.TAVILY_API_KEY:
+            server.add_tool(web_search)
+        else:
+            logger.warning("No web search tools keys are set, web_search will not be available!")
+
+    if not disable_llm_tools and settings.OPENROUTER_API_KEY:
+        server.add_tool(extract_bitflip_info)
+        server.add_tool(generate_research_proposals)
+        server.add_tool(score_research_proposals)
+        server.add_tool(document_qa)
+        if settings.WORKSPACE_DIR:
+            server.add_tool(review_pdf_paper)
+    else:
+        logger.warning("No OpenRouter API key is set, LLM-related tools will not be available!")
+
+    if port is None:
+        if settings.PORT is not None:
+            port = int(settings.PORT)
+        else:
+            port = find_free_port()
+    server.settings.port = port
+    server.settings.host = host
+    server.run(transport=transport)
+
+
+if __name__ == "__main__":
+    fire.Fire(run)

academia_mcp-1.9.1/academia_mcp/settings.py

@@ -0,0 +1,33 @@
+from pathlib import Path
+from typing import Optional
+
+from pydantic_settings import BaseSettings, SettingsConfigDict
+
+
+class Settings(BaseSettings):
+    BASE_URL: str = "https://openrouter.ai/api/v1"
+
+    OPENROUTER_API_KEY: str = ""
+    TAVILY_API_KEY: Optional[str] = None
+    EXA_API_KEY: Optional[str] = None
+    BRAVE_API_KEY: Optional[str] = None
+
+    REVIEW_MODEL_NAME: str = "gpt-5"
+    BITFLIP_MODEL_NAME: str = "deepseek/deepseek-chat-v3-0324"
+    BITFLIP_MAX_COMPLETION_TOKENS: int = 16384
+    DOCUMENT_QA_MODEL_NAME: str = "deepseek/deepseek-chat-v3-0324"
+    DOCUMENT_QA_QUESTION_MAX_LENGTH: int = 10000
+    DOCUMENT_QA_DOCUMENT_MAX_LENGTH: int = 200000
+
+    PORT: int = 5056
+    WORKSPACE_DIR: Optional[Path] = None
+
+    model_config = SettingsConfigDict(
+        env_file=".env",
+        env_file_encoding="utf-8",
+        env_prefix="",
+        extra="ignore",
+    )
+
+
+settings = Settings()

{academia_mcp-1.8.1 → academia_mcp-1.9.1}/academia_mcp/tools/__init__.py

@@ -2,7 +2,7 @@ from .arxiv_search import arxiv_search
 from .anthology_search import anthology_search
 from .arxiv_download import arxiv_download
 from .hf_datasets_search import hf_datasets_search
-from .s2_citations import s2_get_references, s2_get_citations
+from .s2 import s2_get_references, s2_get_citations, s2_corpus_id_from_arxiv_id, s2_get_info
 from .document_qa import document_qa
 from .latex import (
     compile_latex,
@@ -13,7 +13,7 @@ from .latex import (
 from .web_search import web_search, tavily_web_search, exa_web_search, brave_web_search
 from .visit_webpage import visit_webpage
 from .bitflip import extract_bitflip_info, generate_research_proposals, score_research_proposals
-from .review import review_pdf_paper, download_pdf_paper
+from .review import review_pdf_paper, download_pdf_paper, review_pdf_paper_by_url
 
 __all__ = [
     "arxiv_search",
@@ -21,6 +21,8 @@ __all__ = [
     "anthology_search",
     "s2_get_references",
     "s2_get_citations",
+    "s2_corpus_id_from_arxiv_id",
+    "s2_get_info",
     "hf_datasets_search",
     "document_qa",
     "compile_latex",
@@ -35,6 +37,7 @@ __all__ = [
     "generate_research_proposals",
     "score_research_proposals",
     "review_pdf_paper",
+    "review_pdf_paper_by_url",
     "download_pdf_paper",
     "read_pdf",
 ]

{academia_mcp-1.8.1 → academia_mcp-1.9.1}/academia_mcp/tools/bitflip.py

@@ -2,7 +2,6 @@
 # https://web.stanford.edu/class/cs197c/slides/02-literature-search.pdf
 
 import json
-import os
 import random
 from typing import List, Optional, Any, Dict
 
@@ -12,6 +11,7 @@ from datasets import load_dataset  # type: ignore
 from academia_mcp.tools.arxiv_download import arxiv_download
 from academia_mcp.utils import extract_json, encode_prompt
 from academia_mcp.llm import llm_acall, ChatMessage
+from academia_mcp.settings import settings
 
 
 class ProposalDataset:
@@ -201,7 +201,7 @@ async def extract_bitflip_info(arxiv_id: str) -> str:
     Args:
         arxiv_id: The arXiv ID of the paper to extract the Bit-Flip information from.
     """
-    model_name = os.getenv("BITFLIP_MODEL_NAME", "deepseek/deepseek-chat-v3-0324")
+    model_name = settings.BITFLIP_MODEL_NAME
     paper = arxiv_download(arxiv_id)
     abstract = json.loads(paper)["abstract"]
     prompt = encode_prompt(EXTRACT_PROMPT, abstract=abstract)
@@ -240,8 +240,8 @@ async def generate_research_proposals(
     ]
     Use `json.loads` to deserialize the result if you want to get specific items.
     """
-    model_name = os.getenv("BITFLIP_MODEL_NAME", "deepseek/deepseek-chat-v3-0324")
-    max_completion_tokens = int(os.getenv("BITFLIP_MAX_COMPLETION_TOKENS", 16384))
+    model_name = settings.BITFLIP_MODEL_NAME
+    max_completion_tokens = int(settings.BITFLIP_MAX_COMPLETION_TOKENS)
     examples = ProposalDataset.get_dataset()[:]
     examples = random.choices(examples, k=2)
 
@@ -293,7 +293,7 @@ async def score_research_proposals(proposals: str | List[str | Dict[str, Any] |
     Args:
         proposals: A list of JSON strings with research proposals.
     """
-    model_name = os.getenv("BITFLIP_MODEL_NAME", "deepseek/deepseek-chat-v3-0324")
+    model_name = settings.BITFLIP_MODEL_NAME
     if isinstance(proposals, str):
         proposals = json.loads(proposals)
         assert isinstance(proposals, list), "Proposals should be a list of JSON strings"

{academia_mcp-1.8.1 → academia_mcp-1.9.1}/academia_mcp/tools/document_qa.py

@@ -1,14 +1,12 @@
-import os
 import json
 from typing import List, Any, Dict
-from dotenv import load_dotenv
 
 from pydantic import BaseModel
 
 from academia_mcp.llm import llm_acall
 from academia_mcp.utils import truncate_content
+from academia_mcp.settings import settings
 
-load_dotenv()
 
 PROMPT = """You are a helpful assistant that answers questions about documents accurately and concisely.
 Please answer the following questions based solely on the provided document.
@@ -65,10 +63,10 @@ async def document_qa(
         document = json.dumps(document)
     assert document and document.strip(), "Please provide non-empty 'document'"
 
-    question = truncate_content(question, 10000)
-    document = truncate_content(document, 200000)
+    question = truncate_content(question, settings.DOCUMENT_QA_QUESTION_MAX_LENGTH)
+    document = truncate_content(document, settings.DOCUMENT_QA_DOCUMENT_MAX_LENGTH)
 
-    model_name = os.getenv("DOCUMENT_QA_MODEL_NAME", "deepseek/deepseek-chat-v3-0324")
+    model_name = settings.DOCUMENT_QA_MODEL_NAME
     prompt = PROMPT.format(question=question, document=document)
     content = await llm_acall(
         model_name=model_name, messages=[ChatMessage(role="user", content=prompt)]