Get GOV.UK Page Section

Instructions

Implementation Reference

• govuk_mcp/server.py:469-478 (registration)

  The tool 'govuk_get_section' is registered via the @mcp.tool decorator with name='govuk_get_section' and annotations for title, readOnlyHint, destructiveHint, idempotentHint, openWorldHint.

  @mcp.tool(
      name="govuk_get_section",
      annotations={
          "title": "Get GOV.UK Page Section",
          "readOnlyHint": True,
          "destructiveHint": False,
          "idempotentHint": True,
          "openWorldHint": True,
      },
  )

• govuk_mcp/server.py:480-502 (handler)

  The async function govuk_get_section handles the tool logic: fetches content from GOV.UK Content API using the base_path, then calls parsers.extract_section(payload, anchor) to extract the HTML for the named section. Raises ValueError if the anchor is not found.

  async def govuk_get_section(
      base_path: Annotated[str, Field(description="GOV.UK base_path, e.g. '/universal-credit'", min_length=1, max_length=500)],
      anchor: Annotated[str, Field(description="Section anchor ID from govuk_get_content sections list", min_length=1, max_length=200)],
      ctx: Context,
  ) -> dict:
      """Get the HTML content of one named section of a GOV.UK page.
  
      Use govuk_get_content first to get the list of available section anchors,
      then call this with the anchor of the section you want to read.
      """
      path = base_path.lstrip("/")
      client = _client(ctx)
      resp = await client.get(f"{CONTENT_BASE}/{path}")
      resp.raise_for_status()
      payload = resp.json()
      try:
          html = parsers.extract_section(payload, anchor)
      except KeyError:
          raise ValueError(
              f"Section '{anchor}' not found in '{base_path}'. "
              "Use govuk_get_content to list available anchors."
          )
      return {"base_path": base_path, "anchor": anchor, "content": html}

• govuk_mcp/parsers.py:85-114 (helper)

  The extract_section function is the core parsing helper. It handles both guide-format pages (using details.parts by slug) and standard pages (slicing HTML between <h2 id='anchor'> tags). Raises KeyError if anchor not found.

  def extract_section(payload: dict, anchor: str) -> str:
      """Return the content for a named section.
  
      For guide-format pages: returns the body of the part whose slug matches
      anchor. For standard pages: returns the HTML slice from `<h2 id="anchor">`
      to the next `<h2 id="...">`. Raise KeyError if no such anchor exists.
      """
      details = payload.get("details", {})
  
      # Guide format: look in parts by slug
      parts = details.get("parts") or []
      if parts:
          for part in parts:
              if part.get("slug") == anchor:
                  return part.get("body", "")
          raise KeyError(f"No part with slug {anchor!r}")
  
      # Standard format: h2 slice in body
      body = details.get("body", "")
      pattern = re.compile(
          rf'<h2\b[^>]*\bid="{re.escape(anchor)}"[^>]*>',
          re.IGNORECASE,
      )
      m = pattern.search(body)
      if not m:
          raise KeyError(f"No <h2 id={anchor!r}> in body")
      rest = body[m.end():]
      next_m = _HEADING_WITH_ID.search(rest)
      end = m.end() + next_m.start() if next_m else len(body)
      return body[m.start():end]

• govuk_mcp/server.py:481-484 (schema)

  Input parameters defined via Pydantic Field annotations: base_path (str, 1-500 chars) and anchor (str, 1-200 chars). Return type is dict (informal, not a Pydantic model).

      base_path: Annotated[str, Field(description="GOV.UK base_path, e.g. '/universal-credit'", min_length=1, max_length=500)],
      anchor: Annotated[str, Field(description="Section anchor ID from govuk_get_content sections list", min_length=1, max_length=200)],
      ctx: Context,
  ) -> dict: