BioContextAI Knowledgebase MCP

Instructions

Implementation Reference

• src/biocontext_kb/core/pride/_search_pride_proteins.py:9-95 (handler)

  Main handler function for the bc_search_pride_proteins tool (registered as search_pride_proteins). Queries PRIDE API for proteins in a project, supports keyword search, sorting, pagination. Includes input schema via Annotated Fields.

  @core_mcp.tool()
  def search_pride_proteins(
      project_accession: Annotated[
          str,
          Field(description="PRIDE project accession to search proteins in"),
      ],
      keyword: Annotated[
          Optional[str],
          Field(description="Search keyword for protein names or accessions"),
      ] = None,
      page_size: Annotated[
          int,
          Field(description="Number of results to return (max 100)"),
      ] = 20,
      sort_field: Annotated[
          str,
          Field(description="Sort field: accession, proteinName, or gene"),
      ] = "accession",
      sort_direction: Annotated[
          str,
          Field(description="Sort direction: ASC or DESC"),
      ] = "ASC",
  ) -> dict:
      """Search for proteins identified in a specific PRIDE mass spectrometry project. Useful for finding specific proteins in proteomics datasets.
  
      Returns:
          dict: Proteins list with accessions, names, genes, sequences, modifications, associated projects or error message.
      """
      base_url = "https://www.ebi.ac.uk/pride/ws/archive/v3/pride-ap/search/proteins"
  
      # Build query parameters
      params: dict[str, str | int] = {"projectAccession": project_accession}
  
      if page_size > 100:
          page_size = 100
      params["pageSize"] = page_size
      params["page"] = 0
  
      # Add keyword search
      if keyword:
          params["keyword"] = keyword
  
      # Validate and set sort parameters
      valid_sort_fields = ["accession", "proteinName", "gene"]
      if sort_field not in valid_sort_fields:
          sort_field = "accession"
      params["sortField"] = sort_field
  
      valid_sort_directions = ["ASC", "DESC"]
      if sort_direction.upper() not in valid_sort_directions:
          sort_direction = "ASC"
      params["sortDirection"] = sort_direction.upper()
  
      try:
          response = requests.get(base_url, params=params)
          response.raise_for_status()
  
          search_results = response.json()
  
          if not search_results:
              return {"results": [], "count": 0, "message": f"No proteins found in PRIDE project {project_accession}"}
  
          # Process results to include key information
          processed_results = []
          for protein in search_results:
              processed_protein = {
                  "protein_accession": protein.get("proteinAccession"),
                  "protein_name": protein.get("proteinName"),
                  "gene": protein.get("gene"),
                  "project_count": protein.get("projectCount", 0),
              }
              processed_results.append(processed_protein)
  
          return {
              "results": processed_results,
              "count": len(processed_results),
              "project_accession": project_accession,
              "search_criteria": {"keyword": keyword, "sort_field": sort_field, "sort_direction": sort_direction},
          }
  
      except requests.exceptions.HTTPError as e:
          if e.response.status_code == 404:
              return {"error": f"PRIDE project {project_accession} not found or has no protein data"}
          return {"error": f"HTTP error: {e}"}
      except Exception as e:
          return {"error": f"Exception occurred: {e!s}"}

• src/biocontext_kb/core/__init__.py:16-16 (registration)

  Registers pride tools by importing from .pride which includes search_pride_proteins decorated with @core_mcp.tool().

  from .pride import *

• src/biocontext_kb/core/pride/__init__.py:3-3 (registration)

  Explicit import of the search_pride_proteins function in pride module for higher-level import.

  from ._search_pride_proteins import search_pride_proteins

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

  Defines the FastMCP instance 'core_mcp' with prefix 'BC' used to register all tools including search_pride_proteins via @core_mcp.tool() decorator. Tool names likely prefixed as 'bc_search_pride_proteins'.

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