Release notes for v2.30.11

• 2.30.11
  
  • Free open data tier for data scientists & developers (alpha) 
  • PyGraphistry: Faster REST upload API
  • Docs: 2.0 upload API
  • Experimental: Open social login 

Graphistry Versions

Server	2.30.11
JS React+Vanilla	3.7.1
Python PyGraphistry client	0.11.5

Third-Party Versions

BlazingSQL	0.14 (from 0.13)
Caddy	1.0.3
CUDA (In-Docker)	10.0
Django	3.0 (from 2.2)
Docker (CE)	19.03.2
Docker Compose	1.24.1
Elasticsearch node driver	14.2.2
Pandas	0.24.2
Postgres	12.3 (from 10.4)
Python	3.7.3
Neo4j node driver	1.7.6
RAPIDS	0.14 (from 0.13)
Arrow	0.15.0
Splunk node SDK	1.9.0

Summary

The primary enterprise updates are around the new upload 2.0 API in PyGraphistry, RAPIDS 0.14 upgrades (ex: distributed GPU BlazingSQL!), core internal dependency updates (Python+Node ecosystem, Postgres), improved page load speed & stability, and documentation. Experimental settings expose social login options. 

Version 2.30.11 also introduces Graphistry Hub (hub.graphistry.com). The initial release provides a free API tier for public data collaborations. Graphistry Hub represents increasing support for community sharing of analyses and code. For the first phase, we are focusong on providing the Graphistry experience to web users for whom running a private Graphistry server is not feasible. Future phases are designed to benefit the entire community, such as sharing data connectors and ontologies.

New Features

• Core
  • RAPIDS upgrade to 0.14 from 0.13 should improve performance
• API
  • PyGraphistry:
    • Opt-in support for new API=3
    • Explicit RGB color support (under api=3)
• Administration
  • Experimental: Can open registration for regular users
• Offerings

Docs

• First-time system login page clarifies admin flow 
• 2.0 Upload APIs
• Bundled docs now include the health check routes

Fixes & Tweaks

• PyGraphistry:
  
  • hypergraph-generated graphs work with api=3 Arrow uploads, assuming the input dataframes are also Arrow-compatible
    • graphistry.register(api=3, username="...", password="...")
      You will need to `.login()` again when your initial token expires
    • graphistry.register(api=3, token="...")
      You will need to .register(token=...) when your initial token expires
    • graphistry.register(api=3)
      graphsitry.login(username="...", password="...") #refreshes login token
  • drop_na=True has improved None detection, fixing some missing cases
  • NodeXL: Support for api=3 and tweaked Twitter mapping to use Relationship for edge_title
  • Original palette colors are `int32`:
        nodes_df['my_colors'] = pd.Series([0, 1, 2, 3], dtype='int32')
    Explicit RGB colors are 'int64' with a lowest-siginficant-byte placeholder for alpha:
        nodes_df['my_colors'] = pd.Series([0xFF000000, 0x00FF0000, 0x0000FF00], dtype='int64')
    The use of 'int' vs. traditional 'uint' is due to RAPIDS not yet supporting uint, and likely to stay for all of 2020

Administration

• New first-time login experience with guidance on user account creation
• New Admin UI (Initial)
• /pivot and /notebook users now require staff rights. This enables creating regular users with just API access, and only giving access to the more sensitive areas to trusted staff. Future role-based access controls will enable more flexible patterns.
• New health check /ht (database, app server), and shows up as part of nexus container status
• Experimental: Default-off option to open regular user self-registration:
  admin -> constance config -> Open signups (IS_SIGNUPS_OPEN_AFTER_FIRST)

Migration

• Postgres 10 -> 12: See comments in etc/scripts/migration.sh - call with a flag for enable tolerating Postgres migration errors (expected around Public Schema) for just this upgrade, and test that accounts & data migrated correctly.  For future upgrades, you can continue using more strict migration testing.
• /pivot and /notebook users now require staff rights. This may impact your user flow, and you may need to authorize existing users. Will later move to RBAC for safe non-staff access.