spiderforce4ai 2.5.6__py3-none-any.whl → 2.5.8__py3-none-any.whl

spiderforce4ai/post_extraction_agent.py

@@ -167,6 +167,8 @@ class PostExtractionAgent:
                 backup_path = self.config.output_file.with_suffix(f".bak_{int(time.time())}")
                 self.config.output_file.rename(backup_path)
             self.config.output_file.touch()
+            # Initialize empty results dictionary
+            self.results = {}
 
     def _process_single_content(self, url: str, content: str) -> Optional[Dict]:
         """Process a single piece of content through the LLM."""
@@ -259,8 +261,18 @@ class PostExtractionAgent:
         try:
             if self.config.combine_output and self.config.output_file:
                 self.results[url] = result
+                # Save combined results
                 with open(self.config.output_file, 'w') as f:
                     json.dump(self.results, f, indent=2)
+                
+                # Cleanup backup files after successful save
+                for backup_file in self.config.output_file.parent.glob(f"{self.config.output_file.stem}.bak_*"):
+                    try:
+                        backup_file.unlink()
+                        logger.info(f"Cleaned up backup file: {backup_file}")
+                    except Exception as e:
+                        logger.warning(f"Failed to remove backup file {backup_file}: {e}")
+                        
             elif not self.config.combine_output and self.config.output_file:
                 individual_file = self.config.output_file.parent / f"{url.replace('/', '_')}.json"
                 with open(individual_file, 'w') as f:
@@ -299,13 +311,6 @@ class PostExtractionAgent:
                         # Add URL to result before transformation
                         result['url'] = url
                         
-                        # Check for required fields and set to None if missing
-                        required_fields = ['ContactInformation', 'CallToAction', 'KeyPoints', 'Title', 'Description', 'CanonicalUrl']
-                        for field in required_fields:
-                            if field not in result:
-                                logger.warning(f"Missing field '{field}' in LLM response for {url}, setting to None")
-                                result[field] = None
-                        
                         logger.info(f"Executing custom transformer function for {url}")
                         transformed_result = self.config.custom_transform_function(result)
                         logger.info(f"Successfully applied custom transformation for {url}")

{spiderforce4ai-2.5.6.dist-info → spiderforce4ai-2.5.8.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: spiderforce4ai
-Version: 2.5.6
+Version: 2.5.8
 Summary: Python wrapper for SpiderForce4AI HTML-to-Markdown conversion service with LLM post-processing
 Home-page: https://petertam.pro
 Author: Piotr Tamulewicz

spiderforce4ai-2.5.8.dist-info/RECORD

@@ -0,0 +1,7 @@
+spiderforce4ai/__init__.py,sha256=BlrRIrtpDUHjmDedqgXP1KbAAsAH9vwFPncUR5VGGyM,44804
+spiderforce4ai/post_extraction_agent.py,sha256=nyc6V21E3laTetFG2aEBMmFiXv5goumo8eHoupp3mKA,16430
+spiderforce4ai-2.5.8.dist-info/METADATA,sha256=eitdlF0yEzoFabYliZWvNSx7hGQMBDVvKwgCh21lU18,9012
+spiderforce4ai-2.5.8.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+spiderforce4ai-2.5.8.dist-info/entry_points.txt,sha256=ibARQxOlDiL1ho12zbDZt4Uq5RKSIk_qk159ZlZ46hc,59
+spiderforce4ai-2.5.8.dist-info/top_level.txt,sha256=Kth7A21Js7DCp0j5XBBi-FE45SCLouZkeNZU__Yr9Yk,15
+spiderforce4ai-2.5.8.dist-info/RECORD,,

spiderforce4ai-2.5.6.dist-info/RECORD

@@ -1,7 +0,0 @@
-spiderforce4ai/__init__.py,sha256=BlrRIrtpDUHjmDedqgXP1KbAAsAH9vwFPncUR5VGGyM,44804
-spiderforce4ai/post_extraction_agent.py,sha256=OP8_gVWzWdQ3B6d0vd54IbcopHSVpAkWrUmnrDVT5_k,16327
-spiderforce4ai-2.5.6.dist-info/METADATA,sha256=TjgWAC0iHPrSaedmZL-1PZWCoLEMZ7NJUhd74tfWj6A,9012
-spiderforce4ai-2.5.6.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
-spiderforce4ai-2.5.6.dist-info/entry_points.txt,sha256=ibARQxOlDiL1ho12zbDZt4Uq5RKSIk_qk159ZlZ46hc,59
-spiderforce4ai-2.5.6.dist-info/top_level.txt,sha256=Kth7A21Js7DCp0j5XBBi-FE45SCLouZkeNZU__Yr9Yk,15
-spiderforce4ai-2.5.6.dist-info/RECORD,,