Spaces:
Running
Running
File size: 8,662 Bytes
09801ca | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 | """
Multi-Model Configuration
=========================
Dynamic configuration for all AI models used in the application.
No hardcoding - models are configured via environment or this file.
"""
import os
from typing import Dict, List, Any, Optional
from dataclasses import dataclass, field
@dataclass
class ModelConfig:
"""Configuration for a single AI model"""
id: str
name: str
provider: str
api_id: str # OpenRouter model ID
description: str
context_length: int = 4096
supports_streaming: bool = True
supports_vision: bool = False
supports_charts: bool = True # All models can generate text that includes chart code
category: str = "general" # general, vision, embedding
badge: Optional[str] = None
# Define what this model is BEST at
strengths: List[str] = field(default_factory=lambda: ["text", "analysis"])
# ============================================================================
# MODEL REGISTRY - Best Models for DataVision
# Each model has specific strengths for different types of queries
# ============================================================================
MODELS: Dict[str, ModelConfig] = {
# ----------------------
# GENERAL MODELS - Best for Business Analysis
# ----------------------
"deepseek": ModelConfig(
id="deepseek",
name="DeepSeek V4 Pro",
provider="DeepSeek",
api_id="deepseek-ai/deepseek-v4-pro",
description="Fast • Accurate • Recommended",
context_length=32768,
supports_streaming=True,
supports_charts=True,
category="general",
badge="Best",
strengths=["quick_lookup", "data_analysis", "calculations", "summaries", "charts"]
),
"claude-3.5": ModelConfig(
id="claude-3.5",
name="Claude 3.5 Sonnet",
provider="Anthropic",
api_id="anthropic/claude-3.5-sonnet-20241022",
description="Advanced reasoning & deep analysis",
context_length=200000,
supports_streaming=True,
supports_charts=True,
category="general",
badge="Pro",
strengths=["complex_reasoning", "multi_step_analysis", "insights", "comparisons", "charts"]
),
"llama": ModelConfig(
id="llama",
name="Llama 3.3 70B",
provider="Meta",
api_id="meta-llama/llama-3.3-70b-instruct:free",
description="Comprehensive & detailed reports",
context_length=131072,
supports_streaming=True,
supports_charts=True,
category="general",
strengths=["comprehensive_reports", "detailed_explanations", "charts", "tables"]
),
# ----------------------
# VISION MODELS - For Charts, Reports & Images
# ----------------------
"gpt4v": ModelConfig(
id="gpt4v",
name="GPT-4 Vision",
provider="OpenAI",
api_id="openai/gpt-4-vision-preview",
description="Chart & image analysis",
context_length=128000,
supports_streaming=True,
supports_vision=True,
supports_charts=True,
category="vision",
badge="Vision",
strengths=["image_analysis", "chart_reading", "visual_data_extraction", "charts"]
),
"claude-vision": ModelConfig(
id="claude-vision",
name="Claude 3.5 Vision",
provider="Anthropic",
api_id="anthropic/claude-3.5-sonnet-20241022",
description="Document & report understanding",
context_length=200000,
supports_streaming=True,
supports_vision=True,
supports_charts=True,
category="vision",
badge="Vision",
strengths=["document_analysis", "report_reading", "table_extraction", "charts"]
),
"nemotron": ModelConfig(
id="nemotron",
name="Nemotron Ultra",
provider="NVIDIA",
api_id="nvidia/nemotron-3-ultra-550b-a55b",
description="Predictive & rigorous analysis",
context_length=128000,
supports_streaming=True,
supports_charts=True,
category="general",
strengths=["predictive", "reasoning", "charts"]
),
"glm": ModelConfig(
id="glm",
name="GLM 5.1",
provider="Zhipu",
api_id="z-ai/glm-5.1",
description="Agentic multi-tool execution",
context_length=128000,
supports_streaming=True,
supports_charts=True,
category="general",
strengths=["agentic", "tool_use", "charts"]
),
"kimi": ModelConfig(
id="kimi",
name="Kimi 2.6",
provider="Moonshot",
api_id="moonshotai/kimi-k2.6",
description="Fast multimodal inference",
context_length=32768,
supports_streaming=True,
supports_vision=True,
supports_charts=True,
category="vision",
strengths=["fast_inference", "vision", "charts"]
),
# ----------------------
# EMBEDDING MODELS (for RAG)
# ----------------------
"embedding": ModelConfig(
id="embedding",
name="Text Embedding",
provider="OpenAI",
api_id="openai/text-embedding-3-small",
description="Document embeddings",
context_length=8191,
supports_streaming=False,
supports_charts=False,
category="embedding",
strengths=["embeddings", "similarity_search"]
),
}
def get_model(model_id: str) -> Optional[ModelConfig]:
"""Get model configuration by ID"""
return MODELS.get(model_id)
def get_model_api_id(model_id: str, fallback: str = "deepseek/deepseek-chat") -> str:
"""Get OpenRouter API ID for a model"""
model = MODELS.get(model_id)
if model:
return model.api_id
return fallback
def get_models_by_category(category: str) -> List[ModelConfig]:
"""Get all models in a category"""
return [m for m in MODELS.values() if m.category == category]
def get_vision_models() -> List[ModelConfig]:
"""Get all models that support vision/images"""
return [m for m in MODELS.values() if m.supports_vision]
def get_free_models() -> List[ModelConfig]:
"""Get all free models"""
return [m for m in MODELS.values() if m.is_free]
def get_streaming_models() -> List[ModelConfig]:
"""Get all models that support streaming"""
return [m for m in MODELS.values() if m.supports_streaming]
def get_default_model() -> ModelConfig:
"""Get the default model for general use"""
# Check environment for override
default_id = os.getenv("DEFAULT_MODEL", "deepseek")
return MODELS.get(default_id, MODELS["deepseek"])
def get_default_vision_model() -> ModelConfig:
"""Get the default vision model"""
vision_id = os.getenv("DEFAULT_VISION_MODEL", "gpt4v")
return MODELS.get(vision_id, MODELS.get("llava"))
def model_to_frontend_format(model: ModelConfig) -> Dict[str, Any]:
"""Convert model to frontend-compatible format"""
return {
"id": model.id,
"label": model.name,
"description": model.description,
"badge": model.badge,
"isAI": True,
"supportsVision": model.supports_vision,
"supportsStreaming": model.supports_streaming,
"isFree": model.is_free,
"provider": model.provider,
}
def get_all_models_for_frontend() -> List[Dict[str, Any]]:
"""Get all models formatted for frontend dropdown"""
return [model_to_frontend_format(m) for m in MODELS.values()]
# API endpoint for frontend
def get_available_models_api() -> Dict[str, Any]:
"""API response format for available models"""
return {
"models": get_all_models_for_frontend(),
"default": get_default_model().id,
"default_vision": get_default_vision_model().id if get_default_vision_model() else None,
"categories": {
"general": [m.id for m in get_models_by_category("general")],
"vision": [m.id for m in get_vision_models()],
"code": [m.id for m in get_models_by_category("code")],
}
}
# Test
if __name__ == "__main__":
print("Available Models:")
print("=" * 50)
for model_id, model in MODELS.items():
vision = "👁️" if model.supports_vision else " "
free = "🆓" if model.is_free else " "
stream = "🔄" if model.supports_streaming else " "
print(f"{vision}{free}{stream} {model.name:20} ({model.provider})")
print("\n" + "=" * 50)
print(f"Default: {get_default_model().name}")
print(f"Vision models: {[m.name for m in get_vision_models()]}")
print(f"Free models: {[m.name for m in get_free_models()]}")
|