lollms-client 0.29.2__tar.gz → 0.29.3__tar.gz

{lollms_client-0.29.2/lollms_client.egg-info → lollms_client-0.29.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lollms_client
-Version: 0.29.2
+Version: 0.29.3
 Summary: A client library for LoLLMs generate endpoint
 Author-email: ParisNeo <parisneoai@gmail.com>
 License: Apache Software License

{lollms_client-0.29.2 → lollms_client-0.29.3}/lollms_client/__init__.py

@@ -8,7 +8,7 @@ from lollms_client.lollms_utilities import PromptReshaper # Keep general utiliti
 from lollms_client.lollms_mcp_binding import LollmsMCPBinding, LollmsMCPBindingManager
 from lollms_client.lollms_llm_binding import LollmsLLMBindingManager
 
-__version__ = "0.29.2" # Updated version
+__version__ = "0.29.3" # Updated version
 
 # Optionally, you could define __all__ if you want to be explicit about exports
 __all__ = [

{lollms_client-0.29.2 → lollms_client-0.29.3}/lollms_client/lollms_core.py

@@ -147,9 +147,6 @@ class LollmsClient():
             available = self.binding_manager.get_available_bindings()
             raise ValueError(f"Failed to create LLM binding: {binding_name}. Available: {available}")
 
-        # Determine the effective host address (use LLM binding's if initial was None)
-        effective_host_address = self.host_address
-
         # --- Modality Binding Setup ---
         self.tts_binding_manager = LollmsTTSBindingManager(tts_bindings_dir)
         self.tti_binding_manager = LollmsTTIBindingManager(tti_bindings_dir)
@@ -2961,7 +2958,6 @@ Provide the final aggregated answer in {output_format} format, directly addressi
         callback("Deep analysis complete.", MSG_TYPE.MSG_TYPE_STEP_END)
         return final_output
 
-
     def summarize(
         self,
         text_to_summarize: str,
@@ -2990,6 +2986,7 @@ Provide the final aggregated answer in {output_format} format, directly addressi
                                           is not lost at the boundaries. Defaults to 250.
             streaming_callback (Optional[Callable], optional): A callback function to receive real-time updates
                                                              on the process (e.g., which chunk is being processed).
+                                                             It receives a message, a message type, and optional metadata.
                                                              Defaults to None.
             **kwargs: Additional keyword arguments to be passed to the generation method (e.g., temperature, top_p).
 
@@ -3004,12 +3001,17 @@ Provide the final aggregated answer in {output_format} format, directly addressi
         
         if len(tokens) <= chunk_size_tokens:
             if streaming_callback:
-                streaming_callback("Text is short enough for a single summary.", MSG_TYPE.MSG_TYPE_STEP)
+                streaming_callback("Text is short enough for a single summary.", MSG_TYPE.MSG_TYPE_STEP, {"progress": 0})
             
             prompt_objective = contextual_prompt or "Provide a comprehensive summary of the following text."
             final_prompt = f"{prompt_objective}\n\n--- Text to Summarize ---\n{text_to_summarize}"
             
-            return self.generate_text(final_prompt, **kwargs)
+            summary = self.generate_text(final_prompt, **kwargs)
+            
+            if streaming_callback:
+                streaming_callback("Summary generated.", MSG_TYPE.MSG_TYPE_STEP, {"progress": 100})
+            
+            return summary
 
         # --- Stage 1: Chunking and Independent Summarization ---
         chunks = []
@@ -3021,13 +3023,21 @@ Provide the final aggregated answer in {output_format} format, directly addressi
 
         chunk_summaries = []
         
+        # Total steps include each chunk plus the final synthesis step
+        total_steps = len(chunks) + 1
+        
         # Define the prompt for summarizing each chunk
         summarization_objective = contextual_prompt or "Summarize the key points of the following text excerpt."
         chunk_summary_prompt_template = f"{summarization_objective}\n\n--- Text Excerpt ---\n{{chunk_text}}"
 
         for i, chunk in enumerate(chunks):
+            progress_before = (i / total_steps) * 100
             if streaming_callback:
-                streaming_callback(f"Summarizing chunk {i + 1} of {len(chunks)}...", MSG_TYPE.MSG_TYPE_STEP_START, {"id": f"chunk_{i+1}"})
+                streaming_callback(
+                    f"Summarizing chunk {i + 1} of {len(chunks)}...", 
+                    MSG_TYPE.MSG_TYPE_STEP_START, 
+                    {"id": f"chunk_{i+1}", "progress": progress_before}
+                )
 
             prompt = chunk_summary_prompt_template.format(chunk_text=chunk)
             
@@ -3035,8 +3045,14 @@ Provide the final aggregated answer in {output_format} format, directly addressi
                 # Generate summary for the current chunk
                 chunk_summary = self.generate_text(prompt, **kwargs)
                 chunk_summaries.append(chunk_summary)
+                
+                progress_after = ((i + 1) / total_steps) * 100
                 if streaming_callback:
-                    streaming_callback(f"Chunk {i + 1} summarized.", MSG_TYPE.MSG_TYPE_STEP_END, {"id": f"chunk_{i+1}", "summary_snippet": chunk_summary[:100]})
+                    streaming_callback(
+                        f"Chunk {i + 1} summarized. Progress: {progress_after:.0f}%", 
+                        MSG_TYPE.MSG_TYPE_STEP_END, 
+                        {"id": f"chunk_{i+1}", "summary_snippet": chunk_summary[:100], "progress": progress_after}
+                    )
             except Exception as e:
                 trace_exception(e)
                 if streaming_callback:
@@ -3045,8 +3061,13 @@ Provide the final aggregated answer in {output_format} format, directly addressi
                 chunk_summaries.append(f"[Error summarizing chunk {i+1}]")
 
         # --- Stage 2: Final Synthesis of All Chunk Summaries ---
+        progress_before_synthesis = (len(chunks) / total_steps) * 100
         if streaming_callback:
-            streaming_callback("Synthesizing all chunk summaries into a final version...", MSG_TYPE.MSG_TYPE_STEP_START, {"id": "final_synthesis"})
+            streaming_callback(
+                "Synthesizing all chunk summaries into a final version...", 
+                MSG_TYPE.MSG_TYPE_STEP_START, 
+                {"id": "final_synthesis", "progress": progress_before_synthesis}
+            )
 
         combined_summaries = "\n\n---\n\n".join(chunk_summaries)
         
@@ -3064,7 +3085,11 @@ Provide the final aggregated answer in {output_format} format, directly addressi
         final_summary = self.generate_text(final_synthesis_prompt, **kwargs)
         
         if streaming_callback:
-            streaming_callback("Final summary synthesized.", MSG_TYPE.MSG_TYPE_STEP_END, {"id": "final_synthesis"})
+            streaming_callback(
+                "Final summary synthesized.", 
+                MSG_TYPE.MSG_TYPE_STEP_END, 
+                {"id": "final_synthesis", "progress": 100}
+            )
 
         return final_summary.strip()
 

{lollms_client-0.29.2 → lollms_client-0.29.3}/lollms_client/lollms_discussion.py

@@ -1180,6 +1180,7 @@ class LollmsDiscussion:
                 "- Key decisions or conclusions reached.\n"
                 "- Important entities, projects, or topics mentioned that are likely to recur.\n"
                 "Format the output as a concise list of bullet points. Be brief and factual. "
+                "Do not repeat information that is already in the User Data Zone or the Memory"
                 "If no new, significant long-term information is present, output the single word: 'NOTHING'."
             )
             
@@ -1272,8 +1273,8 @@ class LollmsDiscussion:
                         "content": str,
                         "tokens": int,
                         "breakdown": {
-                            "system_prompt": str,
-                            "memory": str,
+                            "system_prompt": {"content": str, "tokens": int},
+                            "memory": {"content": str, "tokens": int},
                             ...
                         }
                     },
@@ -1291,40 +1292,66 @@ class LollmsDiscussion:
             "current_tokens": 0,
             "zones": {}
         }
+        tokenizer = self.lollmsClient.count_tokens
 
         # --- 1. Assemble and Tokenize the Entire System Context Block ---
         system_prompt_text = (self._system_prompt or "").strip()
-        data_zone_text = self.get_full_data_zone() # This already formats all zones correctly
-        
-        pruning_summary_text = ""
-        if self.pruning_summary and self.pruning_point_id:
-            pruning_summary_text = f"--- Conversation Summary ---\n{self.pruning_summary.strip()}"
+        data_zone_text = self.get_full_data_zone()
+        pruning_summary_content = (self.pruning_summary or "").strip()
+
+        pruning_summary_block = ""
+        if pruning_summary_content and self.pruning_point_id:
+            pruning_summary_block = f"--- Conversation Summary ---\n{pruning_summary_content}"
 
-        # Combine all parts that go into the system block, separated by newlines
         full_system_content_parts = [
-            part for part in [system_prompt_text, data_zone_text, pruning_summary_text] if part
+            part for part in [system_prompt_text, data_zone_text, pruning_summary_block] if part
         ]
         full_system_content = "\n\n".join(full_system_content_parts).strip()
 
         if full_system_content:
-            # Create the final system block as it would be exported
             system_block = f"!@>system:\n{full_system_content}\n"
-            system_tokens = self.lollmsClient.count_tokens(system_block)
+            system_tokens = tokenizer(system_block)
             
-            # Create the breakdown for user visibility
             breakdown = {}
             if system_prompt_text:
-                breakdown["system_prompt"] = system_prompt_text
-            if self.memory and self.memory.strip():
-                breakdown["memory"] = self.memory.strip()
-            if self.user_data_zone and self.user_data_zone.strip():
-                breakdown["user_data_zone"] = self.user_data_zone.strip()
-            if self.discussion_data_zone and self.discussion_data_zone.strip():
-                breakdown["discussion_data_zone"] = self.discussion_data_zone.strip()
-            if self.personality_data_zone and self.personality_data_zone.strip():
-                breakdown["personality_data_zone"] = self.personality_data_zone.strip()
-            if self.pruning_summary and self.pruning_summary.strip():
-                breakdown["pruning_summary"] = self.pruning_summary.strip()
+                breakdown["system_prompt"] = {
+                    "content": system_prompt_text,
+                    "tokens": tokenizer(system_prompt_text)
+                }
+
+            memory_text = (self.memory or "").strip()
+            if memory_text:
+                breakdown["memory"] = {
+                    "content": memory_text,
+                    "tokens": tokenizer(memory_text)
+                }
+
+            user_data_text = (self.user_data_zone or "").strip()
+            if user_data_text:
+                breakdown["user_data_zone"] = {
+                    "content": user_data_text,
+                    "tokens": tokenizer(user_data_text)
+                }
+            
+            discussion_data_text = (self.discussion_data_zone or "").strip()
+            if discussion_data_text:
+                breakdown["discussion_data_zone"] = {
+                    "content": discussion_data_text,
+                    "tokens": tokenizer(discussion_data_text)
+                }
+
+            personality_data_text = (self.personality_data_zone or "").strip()
+            if personality_data_text:
+                breakdown["personality_data_zone"] = {
+                    "content": personality_data_text,
+                    "tokens": tokenizer(personality_data_text)
+                }
+
+            if pruning_summary_content:
+                breakdown["pruning_summary"] = {
+                    "content": pruning_summary_content,
+                    "tokens": tokenizer(pruning_summary_content)
+                }
 
             result["zones"]["system_context"] = {
                 "content": full_system_content,
@@ -1340,7 +1367,6 @@ class LollmsDiscussion:
             branch = self.get_branch(branch_tip_id)
             messages_to_render = branch
             
-            # Adjust for pruning to get the active set of messages
             if self.pruning_summary and self.pruning_point_id:
                 pruning_index = -1
                 for i, msg in enumerate(branch):
@@ -1363,7 +1389,7 @@ class LollmsDiscussion:
             message_count = len(messages_to_render)
 
         if messages_text:
-            tokens = self.lollmsClient.count_tokens(messages_text)
+            tokens = tokenizer(messages_text)
             result["zones"]["message_history"] = {
                 "content": messages_text,
                 "tokens": tokens,
@@ -1371,11 +1397,10 @@ class LollmsDiscussion:
             }
 
         # --- 3. Finalize the Total Count ---
-        # This remains the most accurate way to get the final count, as it uses the
-        # exact same export logic as the chat method.
         result["current_tokens"] = self.count_discussion_tokens("lollms_text", branch_tip_id)
 
         return result
+    
     def switch_to_branch(self, branch_id):
         self.active_branch_id = branch_id
 
@@ -1412,4 +1437,4 @@ class LollmsDiscussion:
         new_metadata = (self.metadata or {}).copy()
         new_metadata[itemname] = item_value
         self.metadata = new_metadata
-        self.commit()
+        self.commit()

{lollms_client-0.29.2 → lollms_client-0.29.3/lollms_client.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lollms_client
-Version: 0.29.2
+Version: 0.29.3
 Summary: A client library for LoLLMs generate endpoint
 Author-email: ParisNeo <parisneoai@gmail.com>
 License: Apache Software License