vector-inspector 0.2.6__tar.gz → 0.3.1__tar.gz

{vector_inspector-0.2.6 → vector_inspector-0.3.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vector-inspector
-Version: 0.2.6
+Version: 0.3.1
 Summary: A comprehensive desktop application for visualizing, querying, and managing vector database data
 Author-Email: Anthony Dawson <anthonypdawson+github@gmail.com>
 License: MIT
@@ -22,6 +22,8 @@ Requires-Dist: sentence-transformers>=2.2.0
 Requires-Dist: fastembed>=0.7.4
 Requires-Dist: pyarrow>=14.0.0
 Requires-Dist: pinecone>=8.0.0
+Requires-Dist: keyring>=25.7.0
+Requires-Dist: hf-xet>=1.2.0
 Description-Content-Type: text/markdown
 
 

{vector_inspector-0.2.6 → vector_inspector-0.3.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "vector-inspector"
-version = "0.2.6"
+version = "0.3.1"
 description = "A comprehensive desktop application for visualizing, querying, and managing vector database data"
 authors = [
     { name = "Anthony Dawson", email = "anthonypdawson+github@gmail.com" },
@@ -19,6 +19,8 @@ dependencies = [
     "fastembed>=0.7.4",
     "pyarrow>=14.0.0",
     "pinecone>=8.0.0",
+    "keyring>=25.7.0",
+    "hf-xet>=1.2.0",
 ]
 requires-python = ">=3.10,<3.13"
 readme = "README.md"

vector_inspector-0.3.1/src/vector_inspector/config/__init__.py

@@ -0,0 +1,4 @@
+"""Data package for Vector Inspector.
+
+Contains static data files like the model registry.
+"""

vector_inspector-0.3.1/src/vector_inspector/config/known_embedding_models.json

@@ -0,0 +1,432 @@
+{
+    "models": [
+        {
+            "name": "all-MiniLM-L6-v2",
+            "type": "sentence-transformer",
+            "dimension": 384,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Fast, small-footprint text embeddings (good default for text search)"
+        },
+        {
+            "name": "openai/clip-vit-base-patch32",
+            "type": "clip",
+            "dimension": 512,
+            "modality": "multimodal",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Standard CLIP ViT-B/32 model; supports matching text ↔ images"
+        },
+        {
+            "name": "paraphrase-albert-small-v2",
+            "type": "sentence-transformer",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Smaller paraphrase-specialized model"
+        },
+        {
+            "name": "all-mpnet-base-v2",
+            "type": "sentence-transformer",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "High-quality text embeddings; recommended for semantic tasks"
+        },
+        {
+            "name": "all-roberta-large-v1",
+            "type": "sentence-transformer",
+            "dimension": 1024,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Large model — high quality, larger memory and compute"
+        },
+        {
+            "name": "gtr-t5-large",
+            "type": "sentence-transformer",
+            "dimension": 1536,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Very large embeddings useful for specialized high-recall tasks"
+        },
+        {
+            "name": "sentence-transformers/multi-qa-MiniLM-L6-cos-v1",
+            "type": "sentence-transformer",
+            "dimension": 384,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Optimized for semantic search and question-answering tasks"
+        },
+        {
+            "name": "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
+            "type": "sentence-transformer",
+            "dimension": 384,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Multilingual support for 50+ languages"
+        },
+        {
+            "name": "sentence-transformers/msmarco-distilbert-base-v4",
+            "type": "sentence-transformer",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Trained on MS MARCO dataset, good for passage retrieval"
+        },
+        {
+            "name": "sentence-transformers/all-distilroberta-v1",
+            "type": "sentence-transformer",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Distilled RoBERTa model, balance of speed and quality"
+        },
+        {
+            "name": "sentence-transformers/paraphrase-mpnet-base-v2",
+            "type": "sentence-transformer",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "High-quality paraphrase detection and semantic similarity"
+        },
+        {
+            "name": "BAAI/bge-small-en-v1.5",
+            "type": "sentence-transformer",
+            "dimension": 384,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Beijing Academy of AI model, strong performance for size"
+        },
+        {
+            "name": "BAAI/bge-base-en-v1.5",
+            "type": "sentence-transformer",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "High-quality English embeddings, MTEB benchmark leader"
+        },
+        {
+            "name": "BAAI/bge-large-en-v1.5",
+            "type": "sentence-transformer",
+            "dimension": 1024,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Large model with excellent retrieval performance"
+        },
+        {
+            "name": "thenlper/gte-small",
+            "type": "sentence-transformer",
+            "dimension": 384,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "General Text Embeddings (GTE) small variant"
+        },
+        {
+            "name": "thenlper/gte-base",
+            "type": "sentence-transformer",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "General Text Embeddings (GTE) base model"
+        },
+        {
+            "name": "thenlper/gte-large",
+            "type": "sentence-transformer",
+            "dimension": 1024,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "General Text Embeddings (GTE) large variant"
+        },
+        {
+            "name": "intfloat/e5-small-v2",
+            "type": "sentence-transformer",
+            "dimension": 384,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "E5 family small model, prefix with 'query: ' or 'passage: '"
+        },
+        {
+            "name": "intfloat/e5-base-v2",
+            "type": "sentence-transformer",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "E5 family base model, strong asymmetric retrieval"
+        },
+        {
+            "name": "intfloat/e5-large-v2",
+            "type": "sentence-transformer",
+            "dimension": 1024,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "E5 family large model, top MTEB performance"
+        },
+        {
+            "name": "intfloat/multilingual-e5-small",
+            "type": "sentence-transformer",
+            "dimension": 384,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Multilingual E5 model supporting 100+ languages"
+        },
+        {
+            "name": "intfloat/multilingual-e5-base",
+            "type": "sentence-transformer",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Multilingual E5 base model, excellent cross-lingual retrieval"
+        },
+        {
+            "name": "intfloat/multilingual-e5-large",
+            "type": "sentence-transformer",
+            "dimension": 1024,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Multilingual E5 large model, best-in-class multilingual embeddings"
+        },
+        {
+            "name": "openai/clip-vit-large-patch14",
+            "type": "clip",
+            "dimension": 768,
+            "modality": "multimodal",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Larger CLIP ViT-L/14 model, better quality than base"
+        },
+        {
+            "name": "openai/clip-vit-large-patch14-336",
+            "type": "clip",
+            "dimension": 768,
+            "modality": "multimodal",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Higher resolution (336x336) variant of ViT-L/14"
+        },
+        {
+            "name": "laion/CLIP-ViT-B-32-laion2B-s34B-b79K",
+            "type": "clip",
+            "dimension": 512,
+            "modality": "multimodal",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "LAION's CLIP trained on 2B image-text pairs"
+        },
+        {
+            "name": "laion/CLIP-ViT-H-14-laion2B-s32B-b79K",
+            "type": "clip",
+            "dimension": 1024,
+            "modality": "multimodal",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "LAION's huge CLIP model, excellent quality"
+        },
+        {
+            "name": "text-embedding-ada-002",
+            "type": "openai",
+            "dimension": 1536,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "openai-api",
+            "description": "OpenAI's production embedding model (legacy). Requires API key."
+        },
+        {
+            "name": "text-embedding-3-small",
+            "type": "openai",
+            "dimension": 1536,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "openai-api",
+            "description": "OpenAI's newer small model, better than ada-002. Requires API key."
+        },
+        {
+            "name": "text-embedding-3-large",
+            "type": "openai",
+            "dimension": 3072,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "openai-api",
+            "description": "OpenAI's large embedding model, highest quality. Requires API key."
+        },
+        {
+            "name": "embed-english-v3.0",
+            "type": "cohere",
+            "dimension": 1024,
+            "modality": "text",
+            "normalization": "none",
+            "source": "cohere-api",
+            "description": "Cohere's English embedding model. Requires API key."
+        },
+        {
+            "name": "embed-english-light-v3.0",
+            "type": "cohere",
+            "dimension": 384,
+            "modality": "text",
+            "normalization": "none",
+            "source": "cohere-api",
+            "description": "Cohere's lightweight English model. Requires API key."
+        },
+        {
+            "name": "embed-multilingual-v3.0",
+            "type": "cohere",
+            "dimension": 1024,
+            "modality": "text",
+            "normalization": "none",
+            "source": "cohere-api",
+            "description": "Cohere's multilingual model supporting 100+ languages. Requires API key."
+        },
+        {
+            "name": "embed-multilingual-light-v3.0",
+            "type": "cohere",
+            "dimension": 384,
+            "modality": "text",
+            "normalization": "none",
+            "source": "cohere-api",
+            "description": "Cohere's lightweight multilingual model. Requires API key."
+        },
+        {
+            "name": "textembedding-gecko@003",
+            "type": "vertex-ai",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "google-cloud",
+            "description": "Google's Gecko model for text embeddings. Requires Google Cloud credentials."
+        },
+        {
+            "name": "text-embedding-004",
+            "type": "vertex-ai",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "google-cloud",
+            "description": "Google's latest text embedding model. Requires Google Cloud credentials."
+        },
+        {
+            "name": "text-multilingual-embedding-002",
+            "type": "vertex-ai",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "google-cloud",
+            "description": "Google's multilingual embedding model. Requires Google Cloud credentials."
+        },
+        {
+            "name": "multimodalembedding@001",
+            "type": "vertex-ai",
+            "dimension": 1408,
+            "modality": "multimodal",
+            "normalization": "l2",
+            "source": "google-cloud",
+            "description": "Google's multimodal embedding model. Requires Google Cloud credentials."
+        },
+        {
+            "name": "voyage-large-2",
+            "type": "voyage",
+            "dimension": 1536,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "voyage-api",
+            "description": "Voyage AI's large model. Requires API key."
+        },
+        {
+            "name": "voyage-code-2",
+            "type": "voyage",
+            "dimension": 1536,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "voyage-api",
+            "description": "Voyage AI's code-optimized model. Requires API key."
+        },
+        {
+            "name": "voyage-2",
+            "type": "voyage",
+            "dimension": 1024,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "voyage-api",
+            "description": "Voyage AI's general-purpose model. Requires API key."
+        },
+        {
+            "name": "jinaai/jina-embeddings-v2-base-en",
+            "type": "sentence-transformer",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Jina AI's 8k context length model, good for long documents"
+        },
+        {
+            "name": "jinaai/jina-embeddings-v2-small-en",
+            "type": "sentence-transformer",
+            "dimension": 512,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Jina AI's small model with 8k context length"
+        },
+        {
+            "name": "nomic-ai/nomic-embed-text-v1",
+            "type": "sentence-transformer",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Nomic's open-source text embedding model with 8k context"
+        },
+        {
+            "name": "nomic-ai/nomic-embed-text-v1.5",
+            "type": "sentence-transformer",
+            "dimension": 768,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Nomic's improved model with better performance"
+        },
+        {
+            "name": "Alibaba-NLP/gte-Qwen2-7B-instruct",
+            "type": "sentence-transformer",
+            "dimension": 3584,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "Very large instruction-following embedding model, SOTA on many benchmarks"
+        },
+        {
+            "name": "nvidia/NV-Embed-v1",
+            "type": "sentence-transformer",
+            "dimension": 4096,
+            "modality": "text",
+            "normalization": "l2",
+            "source": "huggingface",
+            "description": "NVIDIA's embedding model, excellent for retrieval tasks"
+        }
+    ],
+    "metadata": {
+        "version": "1.0.0",
+        "last_updated": "2026-01-24",
+        "description": "Known embedding models registry for Vector Inspector"
+    }
+}

vector_inspector-0.3.1/src/vector_inspector/core/cache_manager.py

@@ -0,0 +1,159 @@
+"""
+Cache manager for storing databrowser and search panel state by database and collection.
+Provides fast switching between collections with automatic invalidation on refresh or settings changes.
+"""
+from typing import Dict, Any, Optional, Tuple
+from dataclasses import dataclass, field
+from datetime import datetime
+
+
+@dataclass
+class CacheEntry:
+    """Represents a cached state for a specific database and collection."""
+    data: Any
+    timestamp: datetime = field(default_factory=datetime.now)
+    
+    # Browser state
+    scroll_position: int = 0
+    selected_indices: list = field(default_factory=list)
+    
+    # Search panel state
+    search_query: str = ""
+    search_filters: Dict[str, Any] = field(default_factory=dict)
+    search_results: Optional[Any] = None
+    
+    # User inputs
+    user_inputs: Dict[str, Any] = field(default_factory=dict)
+
+
+class CacheManager:
+    """
+    Manages cache for databrowser and search panel by (database, collection) key.
+    Supports invalidation on refresh or settings changes.
+    """
+    
+    def __init__(self):
+        self._cache: Dict[Tuple[str, str], CacheEntry] = {}
+        self._enabled = True
+        
+    def get(self, database: str, collection: str) -> Optional[CacheEntry]:
+        """Retrieve cached entry for a database and collection."""
+        if not self._enabled:
+            return None
+        
+        key = (database, collection)
+        return self._cache.get(key)
+    
+    def set(self, database: str, collection: str, entry: CacheEntry) -> None:
+        """Store a cache entry for a database and collection."""
+        if not self._enabled:
+            return
+        
+        key = (database, collection)
+        entry.timestamp = datetime.now()
+        self._cache[key] = entry
+    
+    def update(self, database: str, collection: str, **kwargs) -> None:
+        """Update specific fields in an existing cache entry."""
+        key = (database, collection)
+        if key in self._cache:
+            entry = self._cache[key]
+            for field_name, value in kwargs.items():
+                if hasattr(entry, field_name):
+                    setattr(entry, field_name, value)
+            entry.timestamp = datetime.now()
+        else:
+            # Create new entry with provided fields
+            entry = CacheEntry(data=None)
+            for field_name, value in kwargs.items():
+                if hasattr(entry, field_name):
+                    setattr(entry, field_name, value)
+            self._cache[key] = entry
+    
+    def invalidate(self, database: Optional[str] = None, collection: Optional[str] = None) -> None:
+        """
+        Invalidate cache entries.
+        - If both database and collection are provided, invalidate that specific entry.
+        - If only database is provided, invalidate all collections in that database.
+        - If neither is provided, invalidate all entries (global refresh).
+        """
+        if database is None and collection is None:
+            # Clear all cache
+            self._cache.clear()
+        elif collection is None and database is not None:
+            # Clear all collections in the specified database
+            keys_to_remove = [key for key in self._cache.keys() if key[0] == database]
+            for key in keys_to_remove:
+                del self._cache[key]
+        elif database is not None and collection is not None:
+            # Clear specific database/collection combination
+            key = (database, collection)
+            if key in self._cache:
+                del self._cache[key]
+    
+    def clear(self) -> None:
+        """Clear all cached entries."""
+        self._cache.clear()
+    
+    def enable(self) -> None:
+        """Enable caching."""
+        self._enabled = True
+    
+    def disable(self) -> None:
+        """Disable caching and clear all entries."""
+        self._enabled = False
+        self._cache.clear()
+    
+    def is_enabled(self) -> bool:
+        """Check if caching is enabled."""
+        return self._enabled
+    
+    def get_cache_info(self) -> Dict[str, Any]:
+        """Get information about the current cache state."""
+        return {
+            "enabled": self._enabled,
+            "entry_count": len(self._cache),
+            "entries": [
+                {
+                    "database": db,
+                    "collection": coll,
+                    "timestamp": entry.timestamp.isoformat(),
+                    "has_data": entry.data is not None,
+                    "has_search_results": entry.search_results is not None,
+                }
+                for (db, coll), entry in self._cache.items()
+            ]
+        }
+
+
+# Global cache manager instance
+_cache_manager: Optional[CacheManager] = None
+
+
+def get_cache_manager() -> CacheManager:
+    """Get or create the global cache manager instance."""
+    global _cache_manager
+    if _cache_manager is None:
+        _cache_manager = CacheManager()
+        # Initialize from settings
+        try:
+            from vector_inspector.services.settings_service import SettingsService
+            settings = SettingsService()
+            if not settings.get_cache_enabled():
+                _cache_manager.disable()
+        except Exception:
+            # If settings can't be loaded, default to enabled
+            pass
+    return _cache_manager
+
+
+def invalidate_cache_on_settings_change() -> None:
+    """Invalidate all cache when settings change."""
+    cache = get_cache_manager()
+    cache.invalidate()
+
+
+def invalidate_cache_on_refresh(database: Optional[str] = None, collection: Optional[str] = None) -> None:
+    """Invalidate cache on refresh action."""
+    cache = get_cache_manager()
+    cache.invalidate(database, collection)