BioContextAI Knowledgebase MCP

Instructions

Implementation Reference

• src/biocontext_kb/core/uniprot/_get_uniprot_protein_info.py:9-87 (handler)

  The handler function `get_uniprot_protein_info` for the tool, decorated with `@core_mcp.tool()`, which queries the UniProt REST API to retrieve protein information based on ID, name, gene symbol, species, and includes references optionally.

  @core_mcp.tool()
  def get_uniprot_protein_info(
      protein_id: Annotated[
          Optional[str],
          Field(description="Protein accession number (e.g., 'P04637')"),
      ] = None,
      protein_name: Annotated[
          Optional[str],
          Field(description="Protein name to search for (e.g., 'P53')"),
      ] = None,
      gene_symbol: Annotated[
          Optional[str],
          Field(description="Gene symbol to search for (e.g., 'TP53')"),
      ] = None,
      species: Annotated[
          Optional[str],
          Field(description="Taxonomy ID (e.g., '10090') or species name"),
      ] = None,
      include_references: Annotated[
          bool,
          Field(description="Include references and cross-references in response"),
      ] = False,
  ) -> dict:
      """Retrieve protein information from UniProt database. Provide at least one of protein_id, protein_name, or gene_symbol.
  
      Returns:
          dict: Protein information with accession, proteinDescription, genes, organism, sequence, functions, keywords, references or error message.
      """
      base_url = "https://rest.uniprot.org/uniprotkb/search"
  
      # Ensure at least one search parameter was provided
      if not protein_id and not protein_name and not gene_symbol:
          return {"error": "At least one of protein_id or protein_name or gene_symbol must be provided."}
  
      query_parts = []
  
      if protein_id:
          query_parts.append(f"accession:{protein_id}")
  
      elif protein_name:
          query_parts.append(f"protein_name:{protein_name}")
  
      elif gene_symbol:
          query_parts.append(f"gene:{gene_symbol}")
  
      if species:
          species = str(species).strip()
  
          # Try to determine if it's a taxonomy ID (numeric) or a name
          if species.isdigit():
              query_parts.append(f"organism_id:{species}")
          else:
              query_parts.append(f'taxonomy_name:"{species}"')
  
      query = " AND ".join(query_parts)
  
      params: dict[str, str | int] = {
          "query": query,
          "format": "json",
      }
  
      try:
          response = requests.get(base_url, params=params)
          response.raise_for_status()
  
          result = response.json()
          if not result.get("results"):
              return {"error": "No results found for the given query."}
  
          first_result = result["results"][0]
  
          # Remove references and cross-references by default to reduce response size
          if not include_references:
              first_result.pop("references", None)
              first_result.pop("uniProtKBCrossReferences", None)
  
          return first_result
      except Exception as e:
          return {"error": f"Exception occurred: {e!s}"}

• src/biocontext_kb/app.py:35-39 (registration)

  Registration of the `core_mcp` server (containing the uniprot tool) into the main `mcp_app` with namespace prefix `slugify('BC') = 'bc'`, resulting in the tool name `bc_get_uniprot_protein_info`.

  for mcp in [core_mcp, *(await get_openapi_mcps())]:
      await mcp_app.import_server(
          mcp,
          slugify(mcp.name),
      )

• src/biocontext_kb/core/_server.py:3-6 (registration)

  Definition of `core_mcp` FastMCP server instance named 'BC', into which tools like `get_uniprot_protein_info` are registered via `@tool()` decorators, later namespaced as 'bc_'.

  core_mcp = FastMCP(  # type: ignore
      "BC",
      instructions="Provides access to biomedical knowledge bases.",
  )

• src/biocontext_kb/core/uniprot/_get_uniprot_protein_info.py:10-31 (schema)

  Pydantic schema definition for the tool inputs using Annotated and Field for protein_id, protein_name, gene_symbol, species, and include_references.

  def get_uniprot_protein_info(
      protein_id: Annotated[
          Optional[str],
          Field(description="Protein accession number (e.g., 'P04637')"),
      ] = None,
      protein_name: Annotated[
          Optional[str],
          Field(description="Protein name to search for (e.g., 'P53')"),
      ] = None,
      gene_symbol: Annotated[
          Optional[str],
          Field(description="Gene symbol to search for (e.g., 'TP53')"),
      ] = None,
      species: Annotated[
          Optional[str],
          Field(description="Taxonomy ID (e.g., '10090') or species name"),
      ] = None,
      include_references: Annotated[
          bool,
          Field(description="Include references and cross-references in response"),
      ] = False,
  ) -> dict:

• src/biocontext_kb/core/uniprot/__init__.py:1-7 (helper)

  Exports the `get_uniprot_protein_info` function, allowing it to be imported into the core module for registration.

  from ._get_uniprot_id_by_protein_symbol import get_uniprot_id_by_protein_symbol
  from ._get_uniprot_protein_info import get_uniprot_protein_info
  
  __all__ = [
      "get_uniprot_id_by_protein_symbol",
      "get_uniprot_protein_info",
  ]