fixed dataworkflow

scievo/workflows/data_workflow.py

@@ -194,12 +194,74 @@ class DataWorkflow(BaseModel):
         if result_state.output_summary:
             return result_state.output_summary
 
-        # Fallback: try to read saved analysis.md file
+        # Fallback 1: Try to extract from generate_summary node in intermediate_state
+        for item in reversed(result_state.intermediate_state):
+            if item.get("node_name") == "generate_summary":
+                output = item.get("output", "")
+                if output and output != "No summary generated":
+                    logger.info("Extracted summary from intermediate_state")
+                    return output
+
+        # Fallback 2: Try to extract from last assistant message in history
+        # (the summary might be in the last message)
+        if result_state.history:
+            for msg in reversed(result_state.history):
+                if hasattr(msg, "role") and msg.role == "assistant":
+                    if hasattr(msg, "content") and msg.content:
+                        content = msg.content
+                        # Check if this looks like a summary (not an error message)
+                        if content and not content.startswith("Failed to generate"):
+                            logger.info("Extracted summary from last assistant message")
+                            return content
+
+        # Fallback 3: Try to read saved analysis.md file
         analysis_file = self.workspace_path / "analysis.md"
         if analysis_file.exists():
+            logger.info("Extracted summary from analysis.md file")
             return analysis_file.read_text()
 
-        raise RuntimeError("Data analysis completed but no summary was generated.")
+        # Fallback 4: Try data_analysis.md
+        data_analysis_file = self.workspace_path / "data_analysis.md"
+        if data_analysis_file.exists():
+            logger.info("Extracted summary from data_analysis.md file")
+            return data_analysis_file.read_text()
+
+        # Fallback 5: Generate a basic summary from available information
+        logger.warning("No summary found, generating basic summary from available data")
+        summary_parts = []
+
+        if result_state.paper_search_summary:
+            summary_parts.append(f"## Paper Search Results\n{result_state.paper_search_summary}")
+
+        if result_state.papers:
+            summary_parts.append(
+                f"\n## Papers Found\nFound {len(result_state.papers)} relevant papers."
+            )
+
+        if result_state.datasets:
+            summary_parts.append(
+                f"\n## Datasets Found\nFound {len(result_state.datasets)} relevant datasets."
+            )
+
+        if result_state.metrics:
+            summary_parts.append(
+                f"\n## Metrics\nExtracted {len(result_state.metrics)} evaluation metrics."
+            )
+
+        if result_state.intermediate_state:
+            # Include some intermediate outputs
+            summary_parts.append("\n## Analysis Process")
+            for item in result_state.intermediate_state[-5:]:  # Last 5 items
+                node_name = item.get("node_name", "unknown")
+                output = item.get("output", "")
+                if output and len(output) < 500:  # Only include short outputs
+                    summary_parts.append(f"\n### {node_name}\n{output[:300]}...")
+
+        if summary_parts:
+            return "\n".join(summary_parts)
+
+        # Last resort: return a minimal summary
+        return "Data analysis completed. No detailed summary available. Please check the intermediate states for more information."
 
     def _finalize(self, success: bool):
         """Finalize the workflow."""