lollms-client 0.17.0__py3-none-any.whl → 0.17.2__py3-none-any.whl

examples/text_2_image.py

@@ -14,8 +14,13 @@ LOLLMS_CLIENT_ID = "my_lollms_client_id" # Replace with your actual client ID or
 
 # Initialize LollmsClient, enabling the TTI 'lollms' binding
 # The service_key here is used as client_id by the TTI binding for lollms
+# lc = LollmsClient(
+#     tti_binding_name="lollms"
+# )
+
+# make sure you have a OPENAI_API_KEY environment variable
 lc = LollmsClient(
-    tti_binding_name="lollms"
+    tti_binding_name="dalle"
 )
 
 if not lc.tti:
@@ -79,8 +84,8 @@ def test_generate_image():
     ASCIIColors.cyan("\n--- Testing Generate Image ---")
     prompt = "A futuristic cityscape at sunset, neon lights, flying vehicles"
     negative_prompt = "blurry, low quality, ugly, text, watermark"
-    width = 512
-    height = 512
+    width = 1024
+    height = 1024
     home_dir = Path.home()
     documents_dir = home_dir / "Documents"
     output_filename = documents_dir/"generated_lollms_image.jpg"

examples/text_2_image_diffusers.py

@@ -0,0 +1,274 @@
+# lollms_client/examples/test_tti_bindings.py
+from lollms_client import LollmsClient
+from lollms_client.lollms_types import MSG_TYPE # If using callbacks
+from ascii_colors import ASCIIColors, trace_exception
+from PIL import Image
+from pathlib import Path
+import io
+import os
+import platform # For opening image
+import subprocess # For opening image
+import shutil # For cleanup of diffusers env
+import json # For pretty printing dicts
+from typing import Optional
+try:
+    from huggingface_hub import snapshot_download
+    HUGGINGFACE_HUB_AVAILABLE = True
+except ImportError:
+    HUGGINGFACE_HUB_AVAILABLE = False
+    snapshot_download = None
+    ASCIIColors.warning("huggingface_hub library not found. Diffusers model download test will be skipped.")
+    ASCIIColors.warning("Please install with: pip install huggingface-hub")
+
+LOLLMS_CLIENT_ID = "my_lollms_test_client_id" 
+
+# --- Diffusers Test Specific Configuration ---
+# Using a standard Stable Diffusion model
+DIFFUSERS_MODEL_ID = "runwayml/stable-diffusion-v1-5" # Standard SD 1.5 model
+DIFFUSERS_LOCAL_MODEL_NAME = "sd-v1-5-test-model" # Folder name for the downloaded model
+TEMP_DIFFUSERS_ENV_DIR_NAME = "temp_diffusers_lollms_env_for_test"
+BASE_TEST_PATH = Path(__file__).parent 
+
+def setup_diffusers_environment(base_path: Path):
+    ASCIIColors.cyan(f"\n--- Setting up Diffusers Test Environment for {DIFFUSERS_MODEL_ID} at {base_path} ---")
+    if not HUGGINGFACE_HUB_AVAILABLE:
+        raise ImportError("huggingface_hub is not available. Cannot set up Diffusers environment.")
+
+    temp_env_root = base_path / TEMP_DIFFUSERS_ENV_DIR_NAME
+    if temp_env_root.exists(): 
+        shutil.rmtree(temp_env_root)
+    temp_env_root.mkdir(parents=True, exist_ok=True)
+
+    lollms_paths = {
+        "personal_models_path": temp_env_root / "personal_models",
+        "models_zoo_path": temp_env_root / "models_zoo", 
+        "shared_cache_path": temp_env_root / "shared_cache", 
+        "tti_bindings_path": base_path.parent / "tti_bindings" 
+    }
+    for p_key, p_val in lollms_paths.items():
+        if p_key != "tti_bindings_path": 
+            Path(p_val).mkdir(parents=True, exist_ok=True)
+    
+    diffusers_models_dir = lollms_paths["personal_models_path"] / "diffusers_models"
+    diffusers_models_dir.mkdir(parents=True, exist_ok=True)
+    
+    model_target_dir = diffusers_models_dir / DIFFUSERS_LOCAL_MODEL_NAME
+    
+    ASCIIColors.info(f"Attempting to download {DIFFUSERS_MODEL_ID} to {model_target_dir}...")
+    try:
+        # SD 1.5 often has fp16 revision, which is smaller and faster if GPU is used
+        # For CPU test, main revision is fine. Safetensors are preferred.
+        snapshot_download(
+            repo_id=DIFFUSERS_MODEL_ID,
+            local_dir=str(model_target_dir),
+            local_dir_use_symlinks=False,
+            cache_dir=str(lollms_paths["shared_cache_path"] / "huggingface_hub_cache"),
+            # revision="fp16", # Optional: if you want the fp16 variant specifically
+            allow_patterns=["*.json", "*.txt", "*.safetensors"], # Prefer safetensors
+            # ignore_patterns=["*.bin", "*.ckpt"], # Ignore older formats if safetensors exist
+        )
+        ASCIIColors.green(f"Model {DIFFUSERS_MODEL_ID} downloaded successfully.")
+        
+    except Exception as e:
+        trace_exception(e)
+        ASCIIColors.error(f"Failed to download model {DIFFUSERS_MODEL_ID} or assertion failed: {e}")
+        if model_target_dir.exists():
+            ASCIIColors.info(f"Contents of {model_target_dir} ({model_target_dir.resolve()}):")
+            for item_path in model_target_dir.rglob('*'):
+                if item_path.is_file(): ASCIIColors.info(f"  - {item_path.relative_to(model_target_dir)}")
+                elif item_path.is_dir(): ASCIIColors.info(f"  DIR: {item_path.relative_to(model_target_dir)}")
+        raise
+
+    binding_instance_config = {
+        "model_id_or_path": DIFFUSERS_LOCAL_MODEL_NAME, 
+        "pipeline_class_name": None, # Let AutoPipelineForText2Image handle SD 1.5
+        "device": "cpu", # Keep CPU for stable CI/testing
+        "torch_dtype_str": "float32", # float32 for CPU
+        "num_inference_steps": 10, # Fewer steps for faster test
+        "default_width": 512, # Standard for SD 1.5
+        "default_height": 512,
+        "safety_checker_on": False, # Commonly disabled for local testing
+        "lollms_paths": lollms_paths,
+        "hf_variant": None # If using main revision, no variant needed. Use "fp16" if you downloaded that revision.
+    }
+    return binding_instance_config, lollms_paths
+
+def cleanup_diffusers_environment(base_path: Path):
+    ASCIIColors.cyan("\n--- Cleaning up Diffusers Test Environment ---")
+    temp_env_root = base_path / TEMP_DIFFUSERS_ENV_DIR_NAME
+    if temp_env_root.exists():
+        try:
+            shutil.rmtree(temp_env_root)
+            ASCIIColors.info(f"Cleaned up Diffusers temp environment: {temp_env_root}")
+        except Exception as e:
+            ASCIIColors.warning(f"Could not fully clean up {temp_env_root}: {e}")
+            trace_exception(e)
+
+def test_list_tti_services(lc: LollmsClient):
+    ASCIIColors.cyan("\n--- Testing List TTI Services ---")
+    try:
+        services = lc.tti.list_services()
+        if services:
+            ASCIIColors.green(f"Available TTI Services for binding '{lc.tti.binding_name}':")
+            for i, service in enumerate(services):
+                print(f"  {i+1}. Name: {service.get('name')}, Caption: {service.get('caption')}, Help: {service.get('help')}")
+            assert len(services) > 0, "Expected at least one service to be listed."
+        else:
+            ASCIIColors.yellow("No TTI services listed or an empty list was returned.")
+    except Exception as e:
+        ASCIIColors.error(f"Error listing TTI services: {e}")
+        trace_exception(e)
+        raise
+
+def test_get_tti_settings(lc: LollmsClient):
+    ASCIIColors.cyan("\n--- Testing Get Active TTI Settings ---")
+    try:
+        settings = lc.tti.get_settings() # This should be a list of dicts
+        if settings: 
+            ASCIIColors.green(f"Current Active TTI Settings/Template for binding '{lc.tti.binding_name}':")
+            for setting_item in settings:
+                # Ensure setting_item is a dictionary before trying to access .get()
+                if isinstance(setting_item, dict):
+                    print(f"  - Name: {setting_item.get('name')}, Type: {setting_item.get('type')}, Value: {setting_item.get('value')}") 
+                else:
+                    ASCIIColors.warning(f"Found non-dict item in settings list: {setting_item}")
+            assert isinstance(settings, list) and len(settings) > 0, "Expected settings to be a non-empty list of dicts."
+        elif isinstance(settings, dict) and not settings: 
+             ASCIIColors.yellow("No active TTI service or settings configured on the server (empty dict).")
+        else:
+            ASCIIColors.yellow("Could not retrieve TTI settings or format unexpected.")
+            print(f"Received: {settings}")
+    except Exception as e:
+        ASCIIColors.error(f"Error getting TTI settings: {e}")
+        trace_exception(e)
+        raise
+
+def test_set_tti_settings(lc: LollmsClient):
+    ASCIIColors.cyan("\n--- Testing Set Active TTI Settings ---")
+    if lc.tti.binding_name == "diffusers":
+        ASCIIColors.info("Attempting to change 'num_inference_steps' for Diffusers.")
+        try:
+            original_settings = lc.tti.get_settings()
+            original_steps = None
+            for s_item in original_settings:
+                if isinstance(s_item, dict) and s_item.get('name') == 'num_inference_steps':
+                    original_steps = s_item['value']
+                    break
+            assert original_steps is not None, "Could not find 'num_inference_steps' in Diffusers settings."
+            
+            new_steps = int(original_steps) + 1
+            settings_to_set = [{"name": "num_inference_steps", "value": new_steps}] 
+            
+            success = lc.tti.set_settings(settings_to_set) 
+            if success:
+                ASCIIColors.green(f"Successfully sent request to set 'num_inference_steps' to {new_steps}.")
+                current_settings_after_set = lc.tti.get_settings()
+                current_config_steps = None
+                for s_item_after in current_settings_after_set:
+                     if isinstance(s_item_after, dict) and s_item_after.get('name') == 'num_inference_steps':
+                        current_config_steps = s_item_after['value']
+                        break
+                assert current_config_steps == new_steps, f"Verification failed: settings show {current_config_steps}, expected {new_steps}"
+                ASCIIColors.green("Setting change verified in binding's settings.")
+            else:
+                ASCIIColors.red("Failed to set TTI settings (binding indicated failure or no change).")
+                assert False, "set_settings returned False"
+        except Exception as e:
+            ASCIIColors.error(f"Error setting Diffusers TTI settings: {e}")
+            trace_exception(e)
+            raise
+    else: 
+        ASCIIColors.yellow(f"Skipping actual setting change for '{lc.tti.binding_name}' in this detailed test.")
+
+def test_generate_image(lc: LollmsClient, output_filename: Path, prompt: str, negative_prompt: Optional[str], width: int, height: int):
+    ASCIIColors.cyan(f"\n--- Testing Generate Image ({lc.tti.binding_name}) ---")
+    ASCIIColors.info(f"Output to: {output_filename}")
+    ASCIIColors.info(f"Prompt: {prompt}")
+    if negative_prompt: ASCIIColors.info(f"Negative Prompt: {negative_prompt}")
+    ASCIIColors.info(f"Dimensions: {width}x{height}")
+    
+    try:
+        image_bytes = lc.tti.generate_image(
+            prompt=prompt,
+            negative_prompt=negative_prompt,
+            width=width,
+            height=height
+        )
+
+        if image_bytes:
+            ASCIIColors.green(f"Image generated successfully ({len(image_bytes)} bytes).")
+            try:
+                image = Image.open(io.BytesIO(image_bytes))
+                image.save(output_filename)
+                ASCIIColors.green(f"Image saved as {output_filename}")
+                if os.name == 'nt': 
+                    os.startfile(str(output_filename))
+                elif os.name == 'posix': 
+                    try:
+                        opener = "open" if platform.system() == "Darwin" else "xdg-open"
+                        subprocess.run([opener, str(output_filename)], check=False, timeout=5)
+                    except Exception:
+                        ASCIIColors.yellow(f"Could not auto-open image. Please find it at {output_filename}")
+            except Exception as e:
+                ASCIIColors.error(f"Error processing or saving image: {e}")
+                trace_exception(e)
+                raw_output_filename = output_filename.with_suffix(".raw_data")
+                with open(raw_output_filename, "wb") as f_raw: f_raw.write(image_bytes)
+                ASCIIColors.yellow(f"Raw image data saved as {raw_output_filename} for inspection.")
+                raise
+        else:
+            ASCIIColors.red("Image generation returned empty bytes.")
+            assert False, "Image generation returned empty bytes"
+    except Exception as e:
+        ASCIIColors.error(f"Error during image generation: {e}")
+        trace_exception(e)
+        raise
+
+if __name__ == "__main__":
+    # --- DALL-E Test ---
+    ASCIIColors.magenta("\n\n========== DALL-E Binding Test ==========")
+    if not os.environ.get("OPENAI_API_KEY"):
+        ASCIIColors.warning("OPENAI_API_KEY environment variable not set. Skipping DALL-E tests.")
+    else:
+        try:
+            lc_dalle = LollmsClient(tti_binding_name="dalle", service_key=LOLLMS_CLIENT_ID)
+            if not lc_dalle.tti: ASCIIColors.error("DALL-E TTI binding could not be initialized.")
+            else:
+                test_list_tti_services(lc_dalle)
+                test_get_tti_settings(lc_dalle)
+                test_set_tti_settings(lc_dalle) 
+                test_generate_image(lc_dalle, BASE_TEST_PATH / "generated_dalle_image.png",
+                                    "A vibrant oil painting of a mythical creature in an enchanted forest",
+                                    "photorealistic, modern, ugly, deformed", 1024, 1024)
+        except Exception as e:
+            ASCIIColors.error(f"DALL-E test block failed: {e}"); trace_exception(e)
+
+    # --- Diffusers Test ---
+    ASCIIColors.magenta("\n\n========== Diffusers Binding Test ==========")
+    if not HUGGINGFACE_HUB_AVAILABLE:
+        ASCIIColors.warning("Skipping Diffusers tests as huggingface_hub is not available.")
+    else:
+        diffusers_binding_config = None
+        try:
+            diffusers_binding_config, _ = setup_diffusers_environment(BASE_TEST_PATH)
+            lc_diffusers = LollmsClient(tti_binding_name="diffusers", binding_config=diffusers_binding_config, service_key=LOLLMS_CLIENT_ID)
+            if not lc_diffusers.tti: raise RuntimeError("Diffusers TTI binding failed to initialize")
+            
+            test_list_tti_services(lc_diffusers)
+            test_get_tti_settings(lc_diffusers)
+            test_set_tti_settings(lc_diffusers)
+            
+            gen_width = lc_diffusers.tti.config.get("default_width", 512) 
+            gen_height = lc_diffusers.tti.config.get("default_height", 512)
+            diffusers_prompt = "A majestic griffin soaring through a cloudy sky, detailed feathers, fantasy art" 
+            diffusers_negative_prompt = "ugly, blurry, low quality, watermark, text, simple background"
+            
+            test_generate_image(lc_diffusers, BASE_TEST_PATH / "generated_diffusers_image.png",
+                                diffusers_prompt, diffusers_negative_prompt, 
+                                gen_width, gen_height)
+        except Exception as e:
+            ASCIIColors.error(f"Diffusers test block failed: {e}"); trace_exception(e)
+        finally:
+            cleanup_diffusers_environment(BASE_TEST_PATH)
+
+    ASCIIColors.magenta("\n\n========== All Tests Finished ==========")

lollms_client/__init__.py

@@ -6,7 +6,7 @@ from lollms_client.lollms_discussion import LollmsDiscussion, LollmsMessage
 from lollms_client.lollms_utilities import PromptReshaper # Keep general utilities
 from lollms_client.lollms_functions import FunctionCalling_Library
 
-__version__ = "0.17.0"
+__version__ = "0.17.2"
 
 # Optionally, you could define __all__ if you want to be explicit about exports
 __all__ = [

lollms_client/llm_bindings/llamacpp/__init__.py

@@ -24,7 +24,7 @@ pm.ensure_packages(["requests", "pillow"]) # pillow for dummy image in test
 if not pm.is_installed("llama-cpp-binaries"):
     def install_llama_cpp():
         system = platform.system()
-        python_version_simple = f"py{sys.version_info.major}{sys.version_info.minor}" # e.g. py310 for 3.10
+        python_version_simple = f"py{sys.version_info.major}" # e.g. py310 for 3.10
 
         # Determine CUDA suffix based on common recent versions. Adjust if needed.
         # For simplicity, we'll target a common recent CUDA version.
@@ -35,13 +35,13 @@ if not pm.is_installed("llama-cpp-binaries"):
 
 
         if system == "Windows":
-            # llama_cpp_binaries-0.12.0+cu124-py3-none-win_amd64.whl
-            url = f"https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.12.0/llama_cpp_binaries-0.12.0{cuda_suffix}-{python_version_simple}-none-win_amd64.whl"
-            fallback_url = "https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.12.0/llama_cpp_binaries-0.12.0+cu124-py3-none-win_amd64.whl" # Generic py3
+            # llama_cpp_binaries-0.14.0+cu124-py3-none-win_amd64.whl
+            url = f"https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.12.0/llama_cpp_binaries-0.14.0{cuda_suffix}-{python_version_simple}-none-win_amd64.whl"
+            fallback_url = "https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.14.0/llama_cpp_binaries-0.14.0+cu124-py3-none-win_amd64.whl" # Generic py3
         elif system == "Linux":
-            # llama_cpp_binaries-0.12.0+cu124-py3-none-linux_x86_64.whl
-            url = f"https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.12.0/llama_cpp_binaries-0.12.0{cuda_suffix}-{python_version_simple}-none-linux_x86_64.whl"
-            fallback_url = "https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.12.0/llama_cpp_binaries-0.12.0+cu124-py3-none-linux_x86_64.whl" # Generic py3
+            # llama_cpp_binaries-0.14.0+cu124-py3-none-linux_x86_64.whl
+            url = f"https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.14.0/llama_cpp_binaries-0.14.0{cuda_suffix}-{python_version_simple}-none-linux_x86_64.whl"
+            fallback_url = "https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.14.0/llama_cpp_binaries-0.14.0+cu124-py3-none-linux_x86_64.whl" # Generic py3
         else:
             ASCIIColors.warning(f"Unsupported OS for prebuilt llama-cpp-binaries: {system}. Please install manually.")
             return
@@ -915,4 +915,4 @@ if __name__ == '__main__':
             else:
                 ASCIIColors.green("All servers shut down correctly.")
 
-    ASCIIColors.yellow("\nLlamaCppServerBinding test finished.")
+    ASCIIColors.yellow("\nLlamaCppServerBinding test finished.")

lollms_client/lollms_core.py

@@ -61,11 +61,12 @@ class LollmsClient():
                  ctx_size: Optional[int] = 8192,
                  n_predict: Optional[int] = 4096,
                  stream: bool = False,
-                 temperature: float = 0.1,
-                 top_k: int = 50,
-                 top_p: float = 0.95,
-                 repeat_penalty: float = 0.8,
-                 repeat_last_n: int = 40,
+                 temperature: float = 0.7, # Ollama default is 0.8, common default 0.7
+                 top_k: int = 40,          # Ollama default is 40
+                 top_p: float = 0.9,       # Ollama default is 0.9
+                 repeat_penalty: float = 1.1, # Ollama default is 1.1
+                 repeat_last_n: int = 64,  # Ollama default is 64
+
                  seed: Optional[int] = None,
                  n_threads: int = 8,
                  streaming_callback: Optional[Callable[[str, MSG_TYPE], None]] = None,

lollms_client/lollms_llm_binding.py

@@ -157,29 +157,25 @@ class LollmsLLMBinding(ABC):
 
 
     def split_discussion(self, lollms_prompt_string: str, system_keyword="!@>system:", user_keyword="!@>user:", ai_keyword="!@>assistant:") -> list:
-
         """
-        Abra-cadabra! Splits a combined LoLLMs prompt string into a list of
-        OpenAI chat messages.
-
-        Each segment starting with system_keyword/user_keyword/ai_keyword
-        becomes a {"role": ..., "content": ...} dict.
+        Splits a LoLLMs prompt into a list of OpenAI-style messages.
+        If the very first chunk has no prefix, it's assigned to "system".
         """
-        # Build a regex that looks ahead for any of the three markers
+        # Regex to split on any of the three prefixes (lookahead)
         pattern = r"(?={}|{}|{})".format(
             re.escape(system_keyword),
             re.escape(user_keyword),
             re.escape(ai_keyword)
         )
-        # Split the big string into little bits at each keyword
         parts = re.split(pattern, lollms_prompt_string)
         messages = []
 
         for part in parts:
             part = part.strip()
             if not part:
-                continue  # Skip empty rabbitholes
+                continue
 
+            # Determine role and strip prefix if present
             if part.startswith(system_keyword):
                 role = "system"
                 content = part[len(system_keyword):].strip()
@@ -190,16 +186,22 @@ class LollmsLLMBinding(ABC):
                 role = "assistant"
                 content = part[len(ai_keyword):].strip()
             else:
-                # Unknown segment—maybe a ghost?
-                continue
+                # No prefix: if it's the first valid chunk, treat as system
+                if not messages:
+                    role = "system"
+                    content = part
+                else:
+                    # otherwise skip unrecognized segments
+                    continue
 
             messages.append({"role": role, "content": content})
-        if messages[-1]["content"]=="":
-            del messages[-1]
+            if messages[-1]["content"]=="":
+                del messages[-1]
         return messages
 
 
 
+
 class LollmsLLMBindingManager:
     """Manages binding discovery and instantiation"""
 

lollms_client/stt_bindings/whisper/__init__.py

@@ -30,7 +30,7 @@ try:
     torch_index_url = None
     if preferred_torch_device_for_install == "cuda":
         # Specify a common CUDA version index. Pip should resolve the correct torch version.
-        # As of late 2023/early 2024, cu118 or cu121 are common. Let's use cu121.
+        # As of late 2023/early 2024, cu118 or cu121 are common. Let's use cu126.
         # Users with different CUDA setups might need to pre-install torch manually.
         torch_index_url = "https://download.pytorch.org/whl/cu126"
         ASCIIColors.info(f"Attempting to ensure PyTorch with CUDA support (target index: {torch_index_url})")