Buckets:

download
raw
15.2 kB
<<<<<<< SEARCH
if choice == "1":
# --- SECTION 1: LLM BINDING ---
available_llms = []
if llm_bindings_dir.exists():
available_llms = [d.name for d in llm_bindings_dir.iterdir() if d.is_dir() and not d.name.startswith("_")]
if not available_llms:
available_llms = ["ollama", "openai", "open_router", "claude", "gemini", "litellm", "vllm"]
print("\nSelect LLM Binding Provider:")
for idx, binding in enumerate(available_llms):
print(f" [{idx + 1}] {binding}")
llm_choice = input(f"Enter selection number [Current: {cfg.get('llm_binding_name')}]: ").strip()
if llm_choice:
try:
idx = int(llm_choice) - 1
if 0 <= idx < len(available_llms):
cfg["llm_binding_name"] = available_llms[idx]
except ValueError:
print("Invalid selection. Keeping current.")
if not cfg["llm_binding_name"]:
print("No LLM binding selected. Skipping parameters.")
continue
print(f"\nConfiguring '{cfg['llm_binding_name']}' Parameters:")
cur_model = cfg.get("llm_binding_config", {}).get("model_name", "")
cur_host = cfg.get("llm_binding_config", {}).get("host_address", "")
cur_key = cfg.get("llm_binding_config", {}).get("api_key", "")
# Fallbacks
default_model = cur_model or "gpt-4o-mini"
default_host = cur_host or ""
if cfg["llm_binding_name"] == "ollama":
default_model = cur_model or "llama3"
default_host = cur_host or "http://localhost:11434"
elif cfg["llm_binding_name"] == "open_router":
default_model = cur_model or "meta-llama/llama-3-8b-instruct:free"
default_host = cur_host or "https://openrouter.ai/api/v1"
elif cfg["llm_binding_name"] == "claude":
default_model = cur_model or "claude-3-5-sonnet-20240620"
elif cfg["llm_binding_name"] == "gemini":
default_model = cur_model or "gemini-1.5-flash"
model_name = input(f" Enter Model Name [{default_model}]: ").strip() or default_model
new_llm_cfg = {"model_name": model_name}
if cfg["llm_binding_name"] in ("ollama", "open_router", "vllm", "llama_cpp_server", "litellm"):
host_prompt = f" Enter Host Address [{default_host}]: " if default_host else " Enter Host Address: "
host_addr = input(host_prompt).strip() or default_host
if host_addr:
new_llm_cfg["host_address"] = host_addr
if cfg["llm_binding_name"] in ("openai", "open_router", "claude", "gemini", "litellm", "grok", "groq"):
key_prompt = " Enter API/Service Key (leave blank to keep current): " if cur_key else " Enter API/Service Key: "
api_key = input(key_prompt).strip()
if api_key:
new_llm_cfg["api_key"] = api_key
elif cur_key:
new_llm_cfg["api_key"] = cur_key
cfg["llm_binding_config"] = new_llm_cfg
elif choice == "2":
# --- SECTION 2: TTI BINDING ---
setup_tti = input(f"\nDo you want to enable Image Generation (TTI)? [Current: {'Yes' if cfg.get('tti_binding_name') else 'No'}] (y/N): ").strip().lower()
if setup_tti == 'y':
available_ttis = []
if tti_bindings_dir.exists():
available_ttis = [d.name for d in tti_bindings_dir.iterdir() if d.is_dir() and not d.name.startswith("_")]
if not available_ttis:
available_ttis = ["diffusers", "openai", "stability_ai", "gemini"]
print("\nSelect TTI Binding Provider:")
for idx, binding in enumerate(available_ttis):
print(f" [{idx + 1}] {binding}")
tti_choice = input(f"Enter selection number [Current: {cfg.get('tti_binding_name')}]: ").strip()
if tti_choice:
try:
idx = int(tti_choice) - 1
if 0 <= idx < len(available_ttis):
cfg["tti_binding_name"] = available_ttis[idx]
except ValueError:
print("Invalid selection. Keeping current.")
if not cfg["tti_binding_name"]:
print("No TTI binding selected. Skipping parameters.")
continue
print(f"\nConfiguring '{cfg['tti_binding_name']}' Parameters:")
cur_tmodel = cfg.get("tti_binding_config", {}).get("model_name", "")
cur_tkey = cfg.get("tti_binding_config", {}).get("api_key", "")
default_tmodel = cur_tmodel or ("stabilityai/sdxl-turbo" if cfg["tti_binding_name"] == "diffusers" else "dall-e-3")
tti_model_name = input(f" Enter TTI Model Name [{default_tmodel}]: ").strip() or default_tmodel
new_tti_cfg = {"model_name": tti_model_name}
if cfg["tti_binding_name"] in ("openai", "stability_ai", "gemini"):
key_prompt = " Enter TTI API/Service Key (leave blank to keep current): " if cur_tkey else " Enter TTI API/Service Key: "
tti_key = input(key_prompt).strip()
if tti_key:
new_tti_cfg["api_key"] = tti_key
elif cur_tkey:
new_tti_cfg["api_key"] = cur_tkey
cfg["tti_binding_config"] = new_tti_cfg
elif setup_tti == 'n':
cfg["tti_binding_name"] = ""
cfg["tti_binding_config"] = {}
=======
if choice == "1":
# --- SECTION 1: LLM BINDING ---
available_llms = []
if llm_bindings_dir.exists():
available_llms = [d.name for d in llm_bindings_dir.iterdir() if d.is_dir() and not d.name.startswith("_")]
if not available_llms:
available_llms = ["ollama", "openai", "open_router", "claude", "gemini", "litellm", "vllm"]
print("\nSelect LLM Binding Provider:")
for idx, binding in enumerate(available_llms):
print(f" [{idx + 1}] {binding}")
llm_choice = input(f"Enter selection number [Current: {cfg.get('llm_binding_name')}]: ").strip()
if llm_choice:
try:
idx = int(llm_choice) - 1
if 0 <= idx < len(available_llms):
cfg["llm_binding_name"] = available_llms[idx]
except ValueError:
print("Invalid selection. Keeping current.")
if not cfg["llm_binding_name"]:
print("No LLM binding selected. Skipping parameters.")
continue
selected_llm = cfg["llm_binding_name"]
# Dynamic parameter discovery for the selected binding
import yaml
lollms_client_dir = Path(lollms_client.__file__).parent
binding_dir = lollms_client_dir / "llm_bindings" / selected_llm
desc_file = binding_dir / "description.yaml"
discovered_params = []
if desc_file.exists():
try:
with open(desc_file, "r", encoding="utf-8") as f:
desc_data = yaml.safe_load(f) or {}
for key_name in ("global_input_parameters", "input_parameters", "binding_parameters"):
if isinstance(desc_data.get(key_name), list):
discovered_params.extend(desc_data[key_name])
except Exception:
pass
# Fallback generic parameters if description.yaml was missing or failed to parse
if not discovered_params:
discovered_params = [
{"name": "host_address", "type": "str", "description": "The host address of the service.", "default": "http://localhost:11434" if selected_llm == "ollama" else ""},
{"name": "api_key", "type": "str", "description": "The API/Service Key for authentication.", "default": ""}
]
# Load last known configuration for this specific selected binding
saved_configs = cfg.get("llm_bindings_configs", {})
saved_config = saved_configs.get(selected_llm, {})
print(f"\nConfiguring '{selected_llm}' Parameters:")
llm_config = {}
for p in discovered_params:
p_name = p.get("name")
if p_name == "model_name":
continue # We will select this after validation
p_type = p.get("type", "str")
# Look up priority: saved_config value -> default value -> empty string
default_val = saved_config.get(p_name)
if default_val is None:
default_val = p.get("default", "")
label = p_name.replace("_", " ").title()
desc = f" ({p.get('description')})" if p.get("description") else ""
user_val = input(f" Enter {label}{desc} [{default_val}]: ").strip()
if not user_val:
val = default_val
else:
try:
if p_type in ("int", "integer"):
val = int(user_val)
elif p_type in ("float", "double"):
val = float(user_val)
elif p_type in ("bool", "boolean"):
val = user_val.lower() in ("true", "1", "yes", "y")
else:
val = user_val
except ValueError:
val = user_val
llm_config[p_name] = val
# Connect to binding to query and list available models
print(f"\n🔌 Connecting to '{selected_llm}' to retrieve available models...")
selected_model = None
try:
import importlib
module_path = f"lollms_client.llm_bindings.{selected_llm}"
module = importlib.import_module(module_path)
binding_class = getattr(module, getattr(module, "BindingName"))
# Instantiate binding temporarily with entered config
instance = binding_class(**llm_config)
models = []
if hasattr(instance, "list_models") and callable(instance.list_models):
models = instance.list_models()
if models:
print("\nSelect an LLM Model from the list of available models:")
for m_idx, m in enumerate(models):
m_val = m.get("model_name") if isinstance(m, dict) else m
print(f" [{m_idx + 1}] {m_val}")
while True:
try:
m_choice = input(f"\nEnter selection number [1-{len(models)}]: ").strip()
if m_choice:
m_idx = int(m_choice) - 1
if 0 <= m_idx < len(models):
selected_model = models[m_idx].get("model_name") if isinstance(models[m_idx], dict) else models[m_idx]
break
else:
# Default to the saved model or first available model
current_saved_model = saved_config.get("model_name")
if current_saved_model and any((m.get("model_name") if isinstance(m, dict) else m) == current_saved_model for m in models):
selected_model = current_saved_model
else:
selected_model = models[0].get("model_name") if isinstance(models[0], dict) else models[0]
break
except ValueError:
pass
print("Invalid selection. Please try again.")
else:
print("\n⚠️ No models returned by the binding. Please enter model name manually.")
except Exception as e:
print(f"\n❌ Connection failed: {e}")
if not selected_model:
current_saved_model = saved_config.get("model_name") or ""
man_model = input(f" Enter Model Name manually [{current_saved_model}]: ").strip()
selected_model = man_model if man_model else current_saved_model
llm_config["model_name"] = selected_model
cfg["llm_binding_config"] = llm_config
# Save right away to memory state
cfg.setdefault("llm_bindings_configs", {})[selected_llm] = llm_config
elif choice == "2":
# --- SECTION 2: TTI BINDING ---
setup_tti = input(f"\nDo you want to enable Image Generation (TTI)? [Current: {'Yes' if cfg.get('tti_binding_name') else 'No'}] (y/N): ").strip().lower()
if setup_tti == 'y':
available_ttis = []
if tti_bindings_dir.exists():
available_ttis = [d.name for d in tti_bindings_dir.iterdir() if d.is_dir() and not d.name.startswith("_")]
if not available_ttis:
available_ttis = ["diffusers", "openai", "stability_ai", "gemini"]
print("\nSelect TTI Binding Provider:")
for idx, binding in enumerate(available_ttis):
print(f" [{idx + 1}] {binding}")
tti_choice = input(f"Enter selection number [Current: {cfg.get('tti_binding_name')}]: ").strip()
if tti_choice:
try:
idx = int(tti_choice) - 1
if 0 <= idx < len(available_ttis):
cfg["tti_binding_name"] = available_ttis[idx]
except ValueError:
print("Invalid selection. Keeping current.")
if not cfg["tti_binding_name"]:
print("No TTI binding selected. Skipping parameters.")
continue
selected_tti = cfg["tti_binding_name"]
# Dynamic parameter discovery for the selected TTI binding
import yaml
lollms_client_dir = Path(lollms_client.__file__).parent
binding_dir = lollms_client_dir / "tti_bindings" / selected_tti
desc_file = binding_dir / "description.yaml"
discovered_params = []
if desc_file.exists():
try:
with open(desc_file, "r", encoding="utf-8") as f:
desc_data = yaml.safe_

Xet Storage Details

Size:
15.2 kB
·
Xet hash:
d0f262073cb51af7fb224be9d75c630aff083a6562543ad8913c04d67bd1c874

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.