analyze_full_text

Instructions

Implementation Reference

• dracor_mcp_fastmcp.py:655-747 (handler)

  The complete implementation of the 'analyze_full_text' tool. Includes the @mcp.tool decorator for registration and the full handler function that analyzes the play's TEI XML structure (acts, scenes, speeches, stage directions), metadata, characters, plain text stats, and returns a comprehensive dictionary with the analysis.

  @mcp.tool("analyze_full_text")
  def analyze_full_text(corpus_name: str, play_name: str) -> Dict:
      """Analyze the full text of a play, including dialogue and stage directions."""
      try:
          # Get the TEI XML as primary source
          tei_result = get_tei_text(corpus_name, play_name)
          if "error" in tei_result:
              # Fall back to the plain text if TEI fails
              full_text = get_full_text(corpus_name, play_name)
              if "error" in full_text:
                  return {"error": full_text["error"]}
              has_tei = False
              text_content = full_text["text"]
          else:
              has_tei = True
              tei_text = tei_result["tei_text"]
              
              # Simple XML parsing to extract basic structure
              # In a production environment, use a proper XML parser library
              import re
              
              # Extract title
              title_match = re.search(r'<title[^>]*>([^<]+)</title>', tei_text)
              title = title_match.group(1) if title_match else "Unknown"
              
              # Extract author(s)
              author_matches = re.findall(r'<author[^>]*>([^<]+)</author>', tei_text)
              authors = author_matches if author_matches else ["Unknown"]
              
              # Extract acts
              acts = re.findall(r'<div type="act"[^>]*>(.*?)</div>', tei_text, re.DOTALL)
              act_count = len(acts)
              
              # Extract scenes
              scenes = re.findall(r'<div type="scene"[^>]*>(.*?)</div>', tei_text, re.DOTALL)
              scene_count = len(scenes)
              
              # Extract speeches
              speeches = re.findall(r'<sp[^>]*>(.*?)</sp>', tei_text, re.DOTALL)
              speech_count = len(speeches)
              
              # Extract stage directions
              stage_directions = re.findall(r'<stage[^>]*>(.*?)</stage>', tei_text, re.DOTALL)
              stage_direction_count = len(stage_directions)
              
              # Also get the plain text for easier processing
              full_text = get_full_text(corpus_name, play_name)
              text_content = full_text.get("text", "")
              
          # Get play metadata
          play_info = get_play(corpus_name, play_name)
          if "error" in play_info:
              return {"error": play_info["error"]}
              
          # Get character list
          characters = get_characters(corpus_name, play_name)
          if "error" in characters:
              return {"error": characters["error"]}
          
          result = {
              "play": play_info.get("play", {}),
              "characters": characters.get("characters", []),
              "text": text_content,
          }
          
          # Add TEI-specific analysis if available
          if has_tei:
              result["tei_analysis"] = {
                  "title": title,
                  "authors": authors,
                  "structure": {
                      "acts": act_count,
                      "scenes": scene_count,
                      "speeches": speech_count,
                      "stage_directions": stage_direction_count
                  },
                  "text_sample": {
                      "first_speech": speeches[0] if speeches else "",
                      "first_stage_direction": stage_directions[0] if stage_directions else ""
                  }
              }
          
          # Add basic text analysis in either case
          result["analysis"] = {
              "text_length": len(text_content),
              "character_count": len(characters.get("characters", [])),
              "dialogue_to_direction_ratio": text_content.count("\n\nDIALOGUE:") / 
                                            (text_content.count("\n\nSTAGE DIRECTIONS:") or 1)
          }
          
          return result
      except Exception as e:
          return {"error": str(e)}

• dracor_mcp_fastmcp.py:655-655 (registration)

  The MCP tool registration decorator specifically naming this tool 'analyze_full_text'.

  @mcp.tool("analyze_full_text")