hyperstreamdb 0.1.1__tar.gz → 0.1.2__tar.gz

{hyperstreamdb-0.1.1 → hyperstreamdb-0.1.2}/.github/workflows/release.yml

@@ -101,7 +101,7 @@ jobs:
           shared-key: "linux-x86_64"
           cache-targets: false
       - name: Install CUDA Toolkit
-        uses: Jimver/cuda-toolkit@v0.2.30
+        uses: Jimver/cuda-toolkit@v0.2.35
         id: cuda-toolkit
         with:
           cuda: '12.1.0'

{hyperstreamdb-0.1.1 → hyperstreamdb-0.1.2}/.gitignore

@@ -97,3 +97,4 @@ python/hyperstreamdb/*.so
 python/hyperstreamdb/*.pyd
 python/hyperstreamdb/*.dylib
 python/hyperstreamdb/*.dSYM/
+groq_api_key.txt

{hyperstreamdb-0.1.1 → hyperstreamdb-0.1.2}/Cargo.lock

@@ -3775,7 +3775,7 @@ dependencies = [
 
 [[package]]
 name = "hyperstreamdb"
-version = "0.1.1"
+version = "0.1.2"
 dependencies = [
  "ahash 0.8.12",
  "anyhow",

{hyperstreamdb-0.1.1 → hyperstreamdb-0.1.2}/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "hyperstreamdb"
-version = "0.1.1"
+version = "0.1.2"
 edition = "2021"
 license = "MIT AND Apache-2.0"
 description = "HyperStreamDB - Serverless Index-Streaming Database with Overlay Indexing and Vector Search"
@@ -14,6 +14,10 @@ categories = ["database-implementations", "data-structures", "concurrency"]
 name = "hyperstreamdb"
 crate-type = ["cdylib", "rlib"]
 
+[[test]]
+name = "test_connector_ffi"
+required-features = ["java"]
+
 [features]
 default = ["intel_gpu", "candle"]
 candle = []
@@ -67,7 +71,7 @@ opencl3 = { version = "0.9", optional = true } # Intel / OpenCL
 
 # indexing
 roaring = "0.10.2" # For scalar bitmaps
-hnsw_rs = { path = "vendor/hnsw_rs" } # For vector indexing
+hnsw_rs = { version = "0.1.19", path = "vendor/hnsw_rs" } # For vector indexing
 # lance removed due to missing protoc
 
 # error handling
@@ -144,7 +148,7 @@ optional = true
 criterion = { version = "0.5", features = ["async_tokio"] }
 rand = "0.8"
 rand_distr = "0.4"
-hnsw_rs = { path = "vendor/hnsw_rs" }
+hnsw_rs = { version = "0.1.19", path = "vendor/hnsw_rs" }
 tempfile = "3.10"
 proptest = "1.4"
 

{hyperstreamdb-0.1.1 → hyperstreamdb-0.1.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: hyperstreamdb
-Version: 0.1.1
+Version: 0.1.2
 Classifier: Development Status :: 3 - Alpha
 Classifier: Intended Audience :: Developers
 Classifier: Programming Language :: Rust

hyperstreamdb-0.1.2/demo/01_basics.py

@@ -0,0 +1,106 @@
+import hyperstreamdb as hdb
+import pandas as pd
+import numpy as np
+
+# Set pandas options to show full text in results
+pd.set_option('display.max_colwidth', None)
+pd.set_option('display.width', 1000)
+pd.set_option('display.max_rows', 50)
+import os
+import shutil
+from datasets import load_dataset
+from sentence_transformers import SentenceTransformer
+import warnings
+
+# Suppress warnings
+warnings.filterwarnings('ignore')
+
+def run_demo():
+    print("=== HyperStreamDB Basics Demo ===")
+
+    # 1. Setup
+    db_path = "news_db"
+    if os.path.exists(db_path):
+        shutil.rmtree(db_path)
+
+    # Initialize the Intel GPU context (or CPU if not available)
+    # Note: If "intel" fails, fallback to CPU
+    try:
+        ctx = hdb.ComputeContext("intel")
+        print("Using Intel GPU context")
+    except Exception as e:
+        print(f"Intel GPU not available ({e}), using CPU")
+        ctx = None
+
+    # 2. Load Data
+    print("\nLoading AG News dataset...")
+    dataset = load_dataset("ag_news", split="test[:500]")
+    df = pd.DataFrame(dataset)
+    label_map = {0: "World", 1: "Sports", 2: "Business", 3: "Sci/Tech"}
+    df["category"] = df["label"].map(label_map)
+
+    # 3. Embedding
+    print(f"Embedding {len(df)} articles using all-MiniLM-L6-v2...")
+    model = SentenceTransformer("all-MiniLM-L6-v2")
+    embeddings = model.encode(df["text"].tolist())
+    df["embedding"] = [list(e) for e in embeddings]
+
+    # 4. Ingest into HyperStreamDB
+    print("\nIngesting into HyperStreamDB...")
+    table = hdb.Table(db_path, context=ctx)
+    
+    # Enable Vector Indexing for the embedding column and Inverted Index for category
+    # Note: Use add_index_columns for multiple columns
+    table.add_index_columns(["embedding", "category", "text"])
+    
+    table.write(df)
+    table.commit()
+    print("Ingestion complete.")
+
+    # 5. Scalar Filtering
+    print("\n--- Scalar Filtering ---")
+    # Using the fluent API: table.filter(expr).to_pandas()
+    # Or table.to_pandas(filter=expr)
+    sports_query = table.filter("category = 'Sports'")
+    sports_df = sports_query.to_pandas()
+    print(f"Found {len(sports_df)} sports articles.")
+    if len(sports_df) > 0:
+        print(sports_df[["category", "text"]].head())
+
+    # 6. Vector Search
+    print("\n--- Vector Search ---")
+    query_text = "Winning medals in international sports competitions"
+    query_embedding = list(model.encode(query_text))
+
+    # Using the fluent API: table.filter(vector_filter=...).to_pandas()
+    # Or table.to_pandas(vector_filter=...)
+    # The current API seems to prefer vector_filter as a dict or list in to_pandas
+    
+    print(f"Searching for: '{query_text}'")
+    results = table.to_pandas(vector_filter=query_embedding, k=5)
+    
+    if "_distance" in results.columns:
+        print(results[["category", "text", "_distance"]])
+    elif "distance" in results.columns:
+        print(results[["category", "text", "distance"]])
+    else:
+        print(results[["category", "text"]])
+
+    # 7. Hybrid Search
+    print("\n--- Hybrid Search (Sci/Tech + Vector) ---")
+    query_text = "What is the latest news in Artificial Intelligence?"
+    query_embedding = list(model.encode(query_text))
+    # Use Fluent API (as requested by user)
+    hybrid_results = (table.filter("category = 'Sci/Tech'")
+                           .vector_search(query_embedding, k=5)
+                           .to_pandas())
+    if not hybrid_results.empty:
+        cols = ["category", "text"]
+        if "_distance" in hybrid_results.columns: cols.append("_distance")
+        elif "distance" in hybrid_results.columns: cols.append("distance")
+        print(hybrid_results[cols].to_string())
+    else:
+        print("No results found for hybrid search.")
+
+if __name__ == "__main__":
+    run_demo()