spiderforce4ai 2.4.3__py3-none-any.whl → 2.4.6__py3-none-any.whl

spiderforce4ai/__init__.py

@@ -586,7 +586,7 @@ class SpiderForce4AI:
                 for result in results:
                     if result.status == "success":
                         try:
-                            result.extraction_result = await agent.process_content(result.url, result.markdown)
+                            result.extraction_result = agent.process_content(result.url, result.markdown)
                             progress.update(llm_task, advance=1)
                         except Exception as e:
                             console.print(f"[red]Error in post-extraction processing for {result.url}: {str(e)}[/red]")

spiderforce4ai/post_extraction_agent.py

@@ -164,12 +164,9 @@ class PostExtractionAgent:
                 self.config.output_file.rename(backup_path)
             self.config.output_file.touch()
 
-    async def _process_single_content(self, url: str, content: str) -> Optional[Dict]:
+    def _process_single_content(self, url: str, content: str) -> Optional[Dict]:
         """Process a single piece of content through the LLM."""
         try:
-            # Apply rate limiting
-            await self.rate_limiter.acquire()
-            
             # Replace placeholder in messages with actual content
             messages = [
                 {**msg, 'content': msg['content'].replace('{here_markdown_content}', content)}
@@ -183,7 +180,8 @@ class PostExtractionAgent:
             
             for attempt in range(max_retries):
                 try:
-                    response = await completion(
+                    # Call completion synchronously
+                    response = completion(
                         model=self.config.model,
                         messages=messages,
                         max_tokens=self.config.max_tokens,
@@ -192,19 +190,31 @@ class PostExtractionAgent:
                         api_base=self.config.base_url
                     )
                     
-                    # Parse response
-                    extracted_data = json.loads(response.choices[0].message.content)
-                    self.buffer.remove_request(url)  # Remove from buffer if successful
-                    return extracted_data
+                    # Log raw response for debugging
+                    raw_content = response.choices[0].message.content
+                    logger.debug(f"Raw LLM response for {url}: {raw_content}")
                     
-                except json.JSONDecodeError as e:
-                    last_error = f"Invalid JSON response from LLM: {e}"
-                    if attempt < max_retries - 1:
-                        await asyncio.sleep(retry_delay * (attempt + 1))
+                    try:
+                        extracted_data = json.loads(raw_content)
+                        self.buffer.remove_request(url)  # Remove from buffer if successful
+                        return extracted_data
+                    except json.JSONDecodeError as e:
+                        error_msg = (
+                            f"Invalid JSON response from LLM for {url}:\n"
+                            f"Error: {str(e)}\n"
+                            f"Raw content: {raw_content[:500]}..."  # First 500 chars of response
+                        )
+                        logger.error(error_msg)
+                        last_error = error_msg
+                        if attempt < max_retries - 1:
+                            time.sleep(retry_delay * (attempt + 1))
+                            
                 except Exception as e:
-                    last_error = str(e)
+                    error_msg = f"LLM processing error for {url}: {str(e)}"
+                    logger.error(error_msg)
+                    last_error = error_msg
                     if attempt < max_retries - 1:
-                        await asyncio.sleep(retry_delay * (attempt + 1))
+                        time.sleep(retry_delay * (attempt + 1))
             
             # If we get here, all retries failed
             raise Exception(last_error)
@@ -214,6 +224,20 @@ class PostExtractionAgent:
             self.buffer.add_failed_request(url, content, str(e))
             return None
 
+    def _save_result_sync(self, url: str, result: Dict) -> None:
+        """Save individual or combined results synchronously."""
+        try:
+            if self.config.combine_output and self.config.output_file:
+                self.results[url] = result
+                with open(self.config.output_file, 'w') as f:
+                    json.dump(self.results, f, indent=2)
+            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:
+                    json.dump(result, f, indent=2)
+        except Exception as e:
+            logger.error(f"Error saving results for {url}: {str(e)}")
+
     async def _save_result(self, url: str, result: Dict) -> None:
         """Save individual or combined results."""
         try:
@@ -228,25 +252,44 @@ class PostExtractionAgent:
         except Exception as e:
             logger.error(f"Error saving results for {url}: {str(e)}")
 
-    async def process_content(self, url: str, content: str) -> Optional[Dict]:
+    def process_content(self, url: str, content: str) -> Optional[Dict]:
         """Process content with retry mechanism."""
+        logger.info(f"Starting content processing for {url}")
+        
         for attempt in range(self.config.max_retries):
-            result = await self._process_single_content(url, content)
+            logger.info(f"Processing attempt {attempt + 1}/{self.config.max_retries} for {url}")
+            
+            result = self._process_single_content(url, content)
             if result:
+                logger.info(f"Successfully processed content for {url}")
+                
                 # Apply custom transformation if provided
                 if self.config.custom_transform_function:
                     try:
                         result = self.config.custom_transform_function(result)
+                        logger.info(f"Applied custom transformation for {url}")
                     except Exception as e:
-                        logger.error(f"Error in custom transform for {url}: {str(e)}")
+                        error_msg = f"Error in custom transform for {url}: {str(e)}"
+                        logger.error(error_msg)
+                        console.print(f"[red]{error_msg}[/red]")
+                
+                # Save result synchronously
+                try:
+                    self._save_result_sync(url, result)
+                    logger.info(f"Saved results for {url}")
+                except Exception as e:
+                    error_msg = f"Error saving results for {url}: {str(e)}"
+                    logger.error(error_msg)
+                    console.print(f"[red]{error_msg}[/red]")
                 
-                await self._save_result(url, result)
                 return result
             
             # Wait before retry
             if attempt < self.config.max_retries - 1:
-                await asyncio.sleep(self.config.retry_delay)
+                logger.info(f"Attempt {attempt + 1} failed for {url}, waiting {self.config.retry_delay}s before retry")
+                time.sleep(self.config.retry_delay)
         
+        logger.error(f"All processing attempts failed for {url}")
         return None
 
     async def process_bulk_content(self, content_map: Dict[str, str]) -> Dict[str, Optional[Dict]]:

{spiderforce4ai-2.4.3.dist-info → spiderforce4ai-2.4.6.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: spiderforce4ai
-Version: 2.4.3
+Version: 2.4.6
 Summary: Python wrapper for SpiderForce4AI HTML-to-Markdown conversion service with LLM post-processing
 Home-page: https://petertam.pro
 Author: Piotr Tamulewicz

spiderforce4ai-2.4.6.dist-info/RECORD

@@ -0,0 +1,7 @@
+spiderforce4ai/__init__.py,sha256=PPpJLowJhgoRijsF2ebmdkFbIriI_yIFlCi1wL6hSP8,42267
+spiderforce4ai/post_extraction_agent.py,sha256=Ty9-Ai3_RVT86RrPUxKEzf4oUh-Wr7bk2aM87Je_WvE,13580
+spiderforce4ai-2.4.6.dist-info/METADATA,sha256=7FaEgAHdD-8a0XmuDMkpAUjAQ7ZmFTD89IqQM17nllI,9012
+spiderforce4ai-2.4.6.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+spiderforce4ai-2.4.6.dist-info/entry_points.txt,sha256=ibARQxOlDiL1ho12zbDZt4Uq5RKSIk_qk159ZlZ46hc,59
+spiderforce4ai-2.4.6.dist-info/top_level.txt,sha256=Kth7A21Js7DCp0j5XBBi-FE45SCLouZkeNZU__Yr9Yk,15
+spiderforce4ai-2.4.6.dist-info/RECORD,,

spiderforce4ai-2.4.3.dist-info/RECORD

@@ -1,7 +0,0 @@
-spiderforce4ai/__init__.py,sha256=iwCLSvooHtFAo-rU52-nsFgyn99Dflpt_OpSrIW-PqA,42273
-spiderforce4ai/post_extraction_agent.py,sha256=m00-y0SCoutUnxsMwHxPaW-qRm4o5alQWjggDStUSrg,11249
-spiderforce4ai-2.4.3.dist-info/METADATA,sha256=-i_vH6DDs4xVFVdDfaFG_Xka0pqXCSQdCrKgym5r5b0,9012
-spiderforce4ai-2.4.3.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
-spiderforce4ai-2.4.3.dist-info/entry_points.txt,sha256=ibARQxOlDiL1ho12zbDZt4Uq5RKSIk_qk159ZlZ46hc,59
-spiderforce4ai-2.4.3.dist-info/top_level.txt,sha256=Kth7A21Js7DCp0j5XBBi-FE45SCLouZkeNZU__Yr9Yk,15
-spiderforce4ai-2.4.3.dist-info/RECORD,,