Notes from the Trenches

Graphistry 2.32.4 is now live! The most requested features are around icons, badges, and RAPIDS. If you are new to Graphistry, try exploring relationships in your data with GPU visual graph analytics for free on Graphistry Hub and one-click launch a private server in your cloud provider (including our new GovCloud mode). As usual, check out the changelog for more information.

Launch

Custom glyphs and badges

Custom glyphs and badges help highlight context like an entity’s country of origin and highlighting risk. To go beyond the existing built-in icon support, we added two highly configurable capabilities:

• Advanced glyphs: Map data to built-in flags, custom linked images, custom embedded URIs, and custom text
• Badges: Map data to one or more badge icons placed around an entity

The image above shows setting text for the main icon area and attaching data-driven badges using built-in icons with data-driven coloring and built-in flags.

Setting advanced glyphs and custom badges via the various APIs is done similarly as sizes and colors. For example, PyGraphistry Python client users can quickly set advanced glyphs:

g.encode_point_icon(
  'some_column',
  shape="circle", #clip excess
  categorical_mapping={
      'macbook': 'laptop',             #icon
      'Canada': 'flag-icon-ca',          #flag
      'embedded_smile': 'data:svg...',    #embedded image
      'external_logo': 'http://..../img.png' #linked image
  },
  default_mapping="question")

Likewise, language-neutral REST API users can quickly tag nodes with a secondary flag badge:

{
  "graphType": "point",
  "encodingType": "badgeTopRight",
  "attribute": "type",
  "variation": "categorical",
  "mapping": {
    "categorical": {
      "fixed": {
        "America": "flags-icon-us",
        "Canada": "flags-icon-ca"
      },
      "other": "question"
    }
  },
  "shape": "circle",
  "border": { "width": 2, "color": "white", "stroke": "solid" }
}

For further examples and the full specification, head over to the complex encodings docs

RAPIDS 0.15

As usual, Graphistry releases follow the RAPIDS release schedule for stable and secured use. We’re especially excited about RAPIDS 0.15’s new graph algorithms and improved multi-GPU support.

Have an interesting dataset? Try for yourself on Graphistry Hub or one-click launch a private instance for your own private data!

More

The release comes with the usual set of fixes and tweaks. One of our favorites is PyGraphistry‘s graph bindings providing a new short form:

g.nodes(df, 'id_col').edges(df2, 'src_col', 'dst_col')

Launch