Add BioC full text tool to MCP server

[turbomam]

Problem

The get_full_text_from_bioc function is available as a CLI command but not available in the MCP server (it's commented out in main.py).

Current State

CLI: ✅ Available

uvx --from artl-mcp artl-cli get-full-text-from-bioc --pmid "12345678"

MCP: ❌ Disabled

# main.py line ~264
# mcp.tool(get_full_text_from_bioc)

Impact

• Inconsistency between CLI and MCP functionality
• MCP users cannot access BioC XML format full text
• Documentation incorrectly lists BioC as available source

BioC Format Benefits

BioC (BioCreative) is a structured XML format that:

• Preserves document structure (sections, paragraphs)
• Includes annotations and metadata
• Provides cleaner text than raw PDF extraction
• Is available from NCBI for many PMC articles

Proposed Solution

1. Uncomment mcp.tool(get_full_text_from_bioc) in main.py
2. Test the tool via MCP
3. Update documentation to show it's available in both MCP and CLI
4. Add integration test for MCP usage

Related

• Issue Enable disabled MCP tools (DOI metadata, abstracts, PDF extraction, etc.) #212 - General issue about disabled MCP tools
• src/artl_mcp/tools.py:~1200 - Implementation
• src/artl_mcp/main.py:~264 - Commented registration

References

• NCBI BioC API: https://www.ncbi.nlm.nih.gov/research/bionlp/RESTful/pmcoa.cgi/BioC_xml/{pmid}/ascii