"""
Riksarkivet Search MCP Server.
This module sets up the FastMCP server and registers search tools
for searching transcribed historical documents.
"""
from fastmcp import FastMCP
from .search_tool import register_search_tool
search_mcp = FastMCP(
name="ra-search-mcp",
instructions="""
🏛️ Riksarkivet (RA) Search MCP Server
This server provides search access to historical documents from the Swedish National Archives.
AVAILABLE TOOLS:
🔍 search_transcribed - Search AI-transcribed text in digitised documents
- Searches full-text transcriptions of historical documents
- Returns documents and pages containing the keyword with text snippets
- Best for: Finding specific phrases, names, or terms in document content
- Offset parameter required to encourage comprehensive discovery
- Provides direct links to images and ALTO XML
- Supports advanced Solr search syntax (see SEARCH SYNTAX below)
📋 search_metadata - Search document metadata (titles, names, places)
- Searches metadata fields: titles, personal names, place names, descriptions
- Can search both digitised and non-digitised materials (2M+ records)
- Best for: Finding documents by person names, places, or archival descriptions
- Set only_digitised=False to include non-digitised materials
- Supports same advanced Solr search syntax
SEARCH STRATEGY FOR MAXIMUM DISCOVERY:
1. Start with search_transcribed(keyword, offset=0) for initial hits (use syntax guide below when searching) and check some metadata with search_metadata
2. Continue pagination with increasing offsets (50, 100, 150...) if nothing interesting pop up but prioritze transribed materials
3. EXPLORE RELATED TERMS: Search for similar/related words to gather comprehensive context
- Historical variants and spellings (e.g., "trolldom" + "häxa" + "trollkona")
- Synonyms and related concepts (e.g., "satan" + "djävul" for devil-related terms)
- Different word forms (e.g., "trolleri" + "trollkonst" for witchcraft variants)
- Period-appropriate terminology and archaic spellings
4. Note reference codes and page numbers from results for detailed browsing with browse tools, use them to go into interesting matches
SEARCH SYNTAX (Solr Query Syntax):
Basic Search:
- "Stockholm" - Exact term search
- "Stock*" - Wildcard (multiple characters)
- "St?ckholm" - Wildcard (single character)
Fuzzy & Proximity:
- "Stockholm~" - Fuzzy search (edit distance 2)
- "Stockholm~1" - Fuzzy with custom edit distance
- '\"Stockholm trolldom\"~10' - Proximity (within 10 words)
Boolean Operators:
- "(Stockholm AND trolldom)" - Both terms required
- "(Stockholm OR Göteborg)" - Either term (or both)
- "(Stockholm NOT trolldom)" - First without second
- "+Stockholm -trolldom" - Require/exclude terms
Boosting & Grouping:
- "Stockholm^4 troll*" - Boost term relevance (4x)
- '(\"Stockholm dom*\"^4 Reg*)' - Boost phrase with wildcards
- "((Stockholm OR Göteborg) AND troll*)" - Complex grouping
TYPICAL WORKFLOW:
1. Comprehensive search: search_transcribed(term, 0), then search_transcribed(term, 50), etc.
2. Search related terms in parallel to build complete context
3. Use advanced syntax for precise queries (Boolean, wildcards, fuzzy, proximity)
4. Review hit summaries to identify most relevant documents across all searches
5. Use browse tools (separate server) for detailed examination of specific pages
SESSION MEMORY & AVOIDING REDUNDANT CALLS:
- These tools track what they have shown you in this session. Re-calling the same search returns compact stubs for already-seen documents, saving tokens.
- PREFER referencing data already in your conversation context over re-calling a tool. If you already have search results with reference codes, page numbers, or snippets, use them directly.
- Only call a search tool again when you need genuinely NEW information: a different keyword, different offset, or different parameters.
- Pass dedup=False to any tool if you truly need full results repeated.
All tools return rich, formatted text optimized for LLM understanding.
""",
)
# Register search tool
register_search_tool(search_mcp)
