UltipaDocs
Try Playground
  • Introduction
  • Managing HDC Graphs
  • Managing Distributed Projections
  • Installing Algorithms
  • Running Algorithms
    • Degree Centrality
    • Closeness Centrality
    • Harmonic Centrality
    • Graph Centrality
    • Betweenness Centrality
    • Eigenvector Centrality
    • Katz Centrality
    • CELF
    • PageRank
    • ArticleRank
    • TextRank
    • HITS
    • SybilRank
    • Jaccard Similarity
    • Overlap Similarity
    • Cosine Similarity
    • Pearson Correlation Coefficient
    • Euclidean Distance
    • K-Hop All
    • Bipartite Graph
    • HyperANF
    • Connected Component
    • Triangle Counting
    • Induced Subgraph
    • k-Core
    • k-Truss
    • p-Cohesion
    • k-Edge Connected Components
    • Local Clustering Coefficient
    • Topological Sort
    • Schema Overview
    • Dijkstra's Single-Source Shortest Path
    • Delta-Stepping Single-Source Shortest Path
    • Shortest Path Faster Algorithm (SPFA)
    • Minimum Spanning Tree
    • Breadth-First Search (BFS)
    • Depth-First Search (DFS)
    • Adamic-Adar Index
    • Common Neighbors
    • Preferential Attachment
    • Resource Allocation
    • Total Neighbors
    • Louvain
    • Leiden
    • Label Propagation
    • HANP
    • k-Means
    • kNN (k-Nearest Neighbors)
    • K-1 Coloring
    • Conductance
      • Random Walk
      • Node2Vec Walk
      • Node2Vec
      • Struc2Vec Walk
      • Struc2Vec
      • LINE
      • Fast Random Projection
      • Summary of Graph Embedding
      • Gradient Descent
      • Backpropagation
      • Skip-gram
      • Skip-gram Optimization
  1. Docs
  2. /
  3. Graph Analytics & Algorithms

Installing Algorithms

Algorithms for HDC and distributed versions are managed differently, resulting in distinct installation methods.

HDC Algorithms

HDC algorithms are provided as hot-pluggable plugins, enabling installation, update, and removal without causing server downtime. Each package includes a .so file and a .yml configuration file, requiring independent installations on each HDC server.

Installation Methods

You can install HDC algorithms in different ways:

  • Ultipa Manager: If you're using the GDBMS of Ultipa, install algorithms directly via the HDC module interface.
  • Ultipa CLI: This cross-platform command line interface allows you to install algorithms using the install algo command.
  • SDKs: For programmatic access, refer to the relevant algorithm management section for your preferred SDK language:
    • Java
    • Python
    • Go
    • Node.js
    • C#

Showing HDC Algorithms

To view the algorithms installed on an HDC server like hdc-server-1:

SHOW HDC ALGO ON "hdc-server-1"

The returns include the _algoList table, which lists all algorithms installed on the specified HDC server. It includes the following fields:

Field
Description
nameName of the algorithm.
typeType of the algorithm, which can be algo or specific system algorithms like pathfind and khop. System algorithms come pre-installed on the HDC server, enabling essential graph queries and other functions.
write_support_typeSupported writeback execution modes.
can_rollbackIndicates true for algorithms with updates that allow rollback.
paramsIncludes parameters, write_to_file_parameters and write_to_db_parameters.

Distributed Algorithms

Distributed algorithms in Ultipa Powerhouse (v5) come pre-installed as part of the deployment package, so no additional installation is needed. These algorithms are readily available for use without further setup.