onecite 0.0.6

Universal citation management and academic reference toolkit

OneCite

The Universal Citation & Academic Reference Toolkit

Effortlessly convert messy, unstructured references into perfectly formatted, standardized citations.

OneCite is a powerful command-line tool and Python library designed to automate the tedious process of citation management. Feed it anything—DOIs, paper titles,arXiv IDs, or even a mix—and get clean, accurate bibliographic entries in return.

🚀 OneCite for Web is coming.

Dropping soon at hezhiang.com/onecite.

✨ Features • 🚀 Quick Start • 📖 Advanced Usage • 🤖 AI Integration • ⚙️ Configuration • 🤝 Contributing

---

✨ Features

OneCite is packed with features to streamline your entire academic workflow, from initial search to final formatting.

• 🔍 Smart Recognition: Utilizes fuzzy matching against CrossRef and Google Scholar APIs to find the correct reference even from incomplete or slightly inaccurate information.
• 📚 Universal Format Support: Accepts .txt and .bib inputs and can output to BibTeX, APA, and MLA formats, adapting to any project's requirements.
• 🎯 High-Accuracy Refinement: A 4-stage processing pipeline cleans, queries, validates, and formats your entries to ensure the highest quality output.
• 🤖 Intelligent Auto-Completion: Automatically discovers and fills in missing bibliographic data like journal, volume, pages, and author lists.
• 🎛️ Interactive Mode: When multiple potential matches are found, an interactive prompt lets you choose the correct entry, giving you full control over ambiguous references.
• ⚙️ Customizable Templates: A flexible YAML-based template system allows for complete control over the output fields and their priority.
• 🎓 Broad Paper Type Support: Natively understands and processes journal articles, conference papers (NIPS, CVPR, ICML, etc.), and arXiv preprints with ease.
• 📄 Seamless arXiv & URL Integration: Automatically fetches metadata for arXiv IDs and can extract identifiers directly from arxiv.org or doi.org URLs.

🚀 Quick Start

Get up and running with OneCite in under a minute.

Installation

# Recommended: Install from PyPI
pip install onecite

# Or, install from source for the latest version
git clone https://github.com/HzaCode/OneCite.git
cd OneCite
pip install -e .

Basic Usage

1. Create an input file (references.txt):

   10.1038/nature14539
   
   Attention is all you need
   Vaswani et al.
   NIPS 2017
   

2. Run the command:

   onecite process references.txt -o results.bib --quiet
   

3. Get perfectly formatted output (results.bib):

   @article{LeCun2015Deep,
     doi = "10.1038/nature14539",
     title = "Deep learning",
     author = "LeCun, Yann and Bengio, Yoshua and Hinton, Geoffrey",
     journal = "Nature",
     year = 2015,
     volume = 521,
     number = 7553,
     pages = "436-444",
     publisher = "Springer Science and Business Media LLC",
     url = "https://doi.org/10.1038/nature14539",
   }
   
   @inproceedings{Vaswani2017Attention,
     arxiv = "1706.03762",
     title = "Attention Is All You Need",
     author = "Vaswani, Ashish and Shazeer, Noam and Parmar, Niki and Uszkoreit, Jakob and Jones, Llion and Gomez, Aidan N. and Kaiser, Lukasz and Polosukhin, Illia",
     booktitle = "Advances in Neural Information Processing Systems",
     year = 2017,
     url = "https://arxiv.org/abs/1706.03762",
   }
   

📖 Advanced Usage

🎨 Multiple Output Formats (APA, MLA)

# Generate APA formatted citations
onecite process refs.txt --output-format apa
# → LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436-444.
# → Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., ... & Polosukhin, I. (2017). Attention is all you need. In Advances in Neural Information Processing Systems.

# Generate MLA formatted citations
onecite process refs.txt --output-format mla
# → LeCun, Yann, Yoshua Bengio, and Geoffrey Hinton. "Deep Learning." Nature 521.7553 (2015): 436-444.
# → Vaswani, Ashish, et al. "Attention Is All You Need." Advances in Neural Information Processing Systems. 2017.

🤖 Interactive Disambiguation

For ambiguous entries, use the --interactive flag to ensure accuracy.

Command:

onecite process ambiguous.txt --interactive

Example Interaction:

1. Deep learning
   Authors: LeCun, Yann; Bengio, Yoshua; Hinton, Geoffrey
   Journal: Nature
   Year: 2015
   Match Score: 92.5
   DOI: 10.1038/nature14539

2. Deep belief networks
   Authors: Hinton, Geoffrey E.
   Journal: Scholarpedia
   Year: 2009
   Match Score: 78.3
   DOI: 10.4249/scholarpedia.5947

Please select (1-2, 0=skip): 1
✅ Selected: Deep learning

🐍 Use as a Python Library

Integrate OneCite's processing power directly into your Python scripts.

from onecite import process_references

# Define a callback for non-interactive selection (e.g., always choose the best match)
def auto_select_callback(candidates):
    return 0

result = process_references(
    input_content="Deep learning review\nLeCun, Bengio, Hinton\nNature 2015",
    input_type="txt",
    output_format="bibtex",
    interactive_callback=auto_select_callback
)

print(result['output_content'])

📑 Supported Input Types

OneCite is designed to be flexible and understands various common academic identifiers.

• DOI: 10.1038/nature14539
• Conference Papers: Attention is all you need, Vaswani et al., NIPS 2017
• arXiv ID: 1706.03762
• URLs: https://arxiv.org/abs/1706.03762

🤖 AI Assistant Integration (MCP)

OneCite provides complete Model Context Protocol (MCP) support, enabling AI assistants to directly use all of OneCite's functionality for literature search, processing, and formatting.

✨ Available Functions

• cite - Generate single academic citations
  • Supports DOI, paper titles, arXiv IDs, and other input types
  • Supports APA, MLA, BibTeX, and other output formats
• batch_cite - Batch citation generation
  • Process multiple literature sources at once
  • Automatically handle different input types
• search - Academic literature search
  • Search for relevant literature based on keywords
  • Return structured literature information

🚀 Quick Start

1. Install OneCite (if not already installed):

   pip install onecite
   

2. Test MCP server:

   onecite-mcp
   

3. Configure AI assistant: Add to settings.json in MCP-supported editors:

   {
     "mcpServers": {
       "onecite": {
         "command": "onecite-mcp",
         "args": [],
         "env": {}
       }
     }
   }
   

4. Restart your editor, and the AI assistant will have access to OneCite's complete functionality!

📊 Test Status

✅ Server Startup - MCP server starts and responds normally
✅ Citation Function - DOI parsing and formatting work correctly
✅ Batch Processing - Multi-source batch processing works normally
✅ Search Function - Literature search functionality works correctly
✅ Command Line Tool - onecite-mcp command is available

💡 Usage Examples

After configuration, you can directly tell your AI assistant:

• "Generate an APA format citation for this DOI: 10.1038/nature14539"
• "Batch process these references and generate BibTeX format"
• "Search for the latest papers on machine learning"

The AI assistant will automatically call OneCite's corresponding functions and return results.

⚙️ Configuration

📋 Command Line Options

Option	Description	Default
--input-type	Input format (txt, bib)	txt
--output-format	Output format (bibtex, apa, mla)	bibtex
--template	Specify a custom template YAML to use	journal_article_full
--interactive	Enable interactive mode for disambiguation	False
--quiet	Suppress verbose logging	False
--output, -o	Path to the output file	stdout

🎨 Custom Templates

Define custom output formats using a simple YAML template.

Example my_template.yaml:

name: my_template
entry_type: "@article"
fields:
  - name: author
    required: true
  - name: title  
    required: true
  - name: journal
    required: true
  - name: year
    required: true
  - name: doi
    required: false
    source_priority: [crossref_api]

Usage:` ``bash onecite process refs.txt --template my_template.yaml```

🔄 Core Processing Pipeline

OneCite ensures high accuracy and quality through a sophisticated four-stage processing pipeline. The diagram below shows the complete workflow from raw input to final formatted output.

💡 MCP Integration: Through Model Context Protocol, AI assistants can directly invoke this complete processing pipeline without requiring users to manually operate the command line.

graph TD
    A["Input Content"] --> B["Stage 1: Parsing Module"]
    
    B --> B1{"Input Type?"}
    B1 -->|TXT| B2["Parse Text<br/>- Split entries by double newlines<br/>- Extract DOIs and URLs<br/>- Generate query strings"]
    B1 -->|BIB| B3["Parse BibTeX<br/>- Parse existing entries<br/>- Extract metadata"]
    B2 --> C["Raw Entry List<br/>RawEntry"]
    B3 --> C
    
    C --> D["Stage 2: Identification Module"]
    D --> D1{"DOI exists?"}
    
    D1 -->|Yes| D2["Validate DOI format<br/>Regex matching"]
    D2 --> D3["Verify DOI via CrossRef API"]
    D3 --> D4{"DOI exists and valid?"}
    
    D4 -->|Yes| D5["Get metadata from CrossRef<br/>Status: identified"]
    D4 -->|No| D6["DOI format valid but not found<br/>Continue fuzzy search"]
    
    D1 -->|No| D7["Check arXiv ID in URL"]
    D7 --> D8{"Found arXiv ID?"}
    D8 -->|Yes| D9["Extract arXiv ID<br/>Continue processing"]
    D8 -->|No| D10["Check known paper database<br/>Built-in paper matching"]
    
    D6 --> D11["Multi-source fuzzy search"]
    D9 --> D11
    D10 --> D11
    
    D11 --> D11A["CrossRef Search"]
    D11 --> D11B["Google Scholar Search"]
    D11A --> D12["Score candidate results"]
    D11B --> D12
    
    D12 --> D13{"Match score?"}
    D13 -->|">80 points"| D14["Auto-select best match<br/>Status: identified"]
    D13 -->|"70-80 points"| D15["Interactive selection<br/>User chooses from candidates"]
    D13 -->|"<70 points"| D16["Mark as identification failed<br/>Status: identification_failed"]
    
    D15 --> D17["User selection result<br/>Status: identified"]
    D5 --> E["Identified Entry List<br/>IdentifiedEntry"]
    D14 --> E
    D16 --> E
    D17 --> E
    
    E --> F["Stage 3: Enrichment Module"]
    F --> F1{"Entry status?"}
    F1 -->|identified| F2["Enrich metadata"]
    F1 -->|failed| F3["Skip enrichment<br/>Status: enrichment_failed"]
    
    F2 --> F4{"Data source type?"}
    F4 -->|DOI| F5["Get complete metadata from CrossRef"]
    F4 -->|"arXiv ID"| F6["Get metadata from arXiv API"]
    F4 -->|"Search result"| F7["Convert search metadata format"]
    
    F5 --> F8["Generate BibTeX key<br/>FirstAuthorYearTitle format"]
    F6 --> F8
    F7 --> F8
    
    F8 --> F9["Complete missing fields by template<br/>Use priority rules"]
    F9 --> F10["Determine entry type<br/>@article or @inproceedings"]
    F10 --> F11["Status: completed"]
    
    F3 --> G["Completed Entry List<br/>CompletedEntry"]
    F11 --> G
    
    G --> H["Stage 4: Formatting Module"]
    H --> H1{"Output format?"}
    
    H1 -->|BibTeX| H2["Format as BibTeX<br/>- Generate @entry format<br/>- Include all fields"]
    H1 -->|APA| H3["Format as APA style<br/>- Author-date format<br/>- Standard punctuation"]
    H1 -->|MLA| H4["Format as MLA style<br/>- Author-page format<br/>- Specific citation rules"]
    
    H2 --> I["Final Output<br/>Formatted citation string list"]
    H3 --> I
    H4 --> I
    
    I --> J["Processing Report<br/>- Total entries<br/>- Success count<br/>- Failed entry list"]
    
    %% Error handling
    D3 -.->|"API error"| D11
    F5 -.->|"API error"| F3
    F6 -.->|"API error"| F3
    H2 -.->|"Format error"| H5["Add to failed entries"]
    H3 -.->|"Format error"| H5
    H4 -.->|"Format error"| H5
    H5 --> J
    
    %% Template system
    T["Template System"] --> F9
    T --> T1["journal_article_full.yaml<br/>Complete journal article template"]
    T --> T2["conference_paper.yaml<br/>Conference paper template"]
    T1 --> T3["Field requirements<br/>- Required fields<br/>- Optional fields<br/>- Data source priority"]
    T2 --> T3
    
    %% External data sources
    DS["External Data Sources"] --> D3
    DS --> F5
    DS --> F6
    DS --> DS1["CrossRef API<br/>- DOI validation<br/>- Metadata retrieval<br/>- Consistency check"]
    DS --> DS2["arXiv API<br/>- Paper metadata<br/>- PDF information"]
    DS --> DS3["Google Scholar<br/>- Fuzzy search<br/>- Citation data<br/>- Timeout handling"]
    
    %% Style definitions
    classDef stageBox fill:#e1f5fe,stroke:#01579b,stroke-width:2px
    classDef decisionBox fill:#fff3e0,stroke:#e65100,stroke-width:2px
    classDef processBox fill:#f3e5f5,stroke:#4a148c,stroke-width:2px
    classDef outputBox fill:#e8f5e8,stroke:#2e7d32,stroke-width:2px
    classDef systemBox fill:#fafafa,stroke:#424242,stroke-width:2px
    
    class B,D,F,H stageBox
    class B1,D1,D4,D8,D13,F1,F4,H1 decisionBox
    class B2,B3,D2,D3,D5,D6,D7,D9,D10,D11,D11A,D11B,D12,D14,D15,D16,D17,F2,F3,F5,F6,F7,F8,F9,F10,F11,H2,H3,H4,H5 processBox
    class C,E,G,I,J outputBox
    class T,T1,T2,T3,DS,DS1,DS2,DS3 systemBox

🤝 Contributing

Contributions are welcome! Please see CONTRIBUTING.md for development guidelines and instructions on how to submit pull requests.

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.

---

OneCite - Simple, Accurate, and Powerful Citation Management ✨

⭐ Star on GitHub • 🚀 Try the Web App • 📖 Read the Docs • 🐛 Report an Issue • 💬 Start a Discussion