ceteri / pkg

Collaboration on description of personal knowledge graphs (PKG)

Wash Warm, Tumble Dry 👋

• 🔭☕🛠️ Currently working on ...

  • entity resolved knowledge graphs
  • natural language + entity resolution + graph technologies + machine learning, see https://huggingface.co/pacoid
  • devrel @ senzing.com
  • managing partner @ derwen.ai
  • advisor + board member @ argilla.io; advisor @ kungfu.ai; advisor @ dataspartan ; advisor @ kurve.ai
  • graph data science in general -- see: https://www.linkedin.com/groups/6725785/
  • enhancements and integrations for pytextrank, kglab, mkrefs, disparity_filter
  • API development with fastapi, pydantic, pydantic-aioredis
  • "Thinking Sparse and Dense" about how hardware evolution drives AI

• 💬

  • fluently: EN, Py, C++, JavaScript, Bash, LPG, RDF/OWL/SKOS/SHACL, NLP, ML, Neural Networks, Distrib Systems, Optimization
  • conversationally: FR, ES
  • attempted: FA, CY, JA, AR, RU, Abstract Algebra, Topology
  • etc: Scala, Clojure, R, Java, C, FORTRAN, LISP, Pascal, JCL, Spice, Verilog, various assembly
  • pronouns: he/him
  • 📫 reach me: https://derwen.ai/paco, @[email protected], Signal

• 🙌 collaborates on: integrating Python libraries that provide different kinds of language or graph capabilities

  • see also: https://stackoverflow.com/users/1698443/paco

• 👍 animist, bioregionalist, /|\

• 🌳 fun fact: lives on an apple orchard in the Redwoods near the coast, making gelato, sorbet, granita, cider, mead, limoncello, and so on

• 🎬 favorite animation series: Llamas with Hats

• 🏙️ member: ACM, PSF, NumFOCUS, ESIP

• orcid.org/0000-0003-3167-1539 | 📚

• 🚴 activities: cycling, kayaking, hiking

pkg's Issues

Layer 1

Related to the point of including data in addition to documents:

Should we only think about S3-like storage? What about local CSV/XML/JSON local files, or databases?

Tools -- add hypernote

It's maybe worth adding Hypernote to the list of tools.

My first impressions of Hypernote.

(I tried to commit this addition myself but don't have write privileges)

Layer 3

Markdown is document-centric, and serves well in this context.

Perhaps however this should be expanded / renamed to Markup, to include XML/HTML/JSON?

Sharing links to add context to group discussion Use Case

Could this be generalized to annotating, served by tools such as hypothes.is and dokieli?

PKG Use Cases/Requirements

What are the PKG group’s requirements/needs from a tool stack?

• Asynchronous project chat
• Shared document for note taking
• Creating semantic markups
• Shared storage
• Versioning
• Publishing on the web
• Knowledge graph

Current Use Cases for the Group

• Personal knowledge collection
• Sharing links to add context to group discussion
• Topical conversation on KG related subject
• Collaborating on specific project
• Publishing tutorials

Visualization

Recently Dan McCreary published this note on Auto-Generating KG Documents, which i found relevant.

In it he mentioned mkdocs, as well as Mermaid, a library for generating drawings using Markdown.

I was not aware of it, but it looks like it would make sense to use it, too, starting with the visualization of the Layers

Personal knowledge collection use case

I think this one could be refined further, as it can include both

1. Archiving knowledge items of interest (e.g. saving bookmarks), and
2. Generating new knowledge (e.g. writing notes)

Furthermore, both the above refer to documents. What about data, such as:

1. People's data held / generated by public services (e.g. SOLID / Raf Buyle use case)
2. People's data held / generated by various private services (e.g. Fitbit, Imdb, last.fm etc