28:null 29:["$","title",null,{"children":"kglab in ObservableHQ | Thad Kerosky | Observable"}] 2e:T527,#### What to expect from ObservableHQ and Pyodide - Many of the major libraries `kglab` uses like `rdflib` are more directly supported in Javascript and may not need the Pyodide/`kglab` wrapper in all situations. - If you're not explicit with Pyodide cell dependencies, ObservableHQ may not be able to correctly [set the cell dataflow](https://observablehq.com/@observablehq/how-observable-runs). Unlike Jupyter, it is not linear by default. See the [Observable pyodide layer notebook](https://observablehq.com/@thadk/pyodide-18) for more. - Custom-built Pyodide dependencies and custom requirement specifications (`meta.yaml`) can be found [in this repository](https://github.com/thadk/pyodide_packages/tree/main/pyodide_18_1). Welcome to use any of this under an MIT license. These were built using the [Pyodide docker image](https://pyodide.org/en/stable/development/building-from-sources.html). The forthcoming Pyodide 0.19 release due next week will make custom-building easier. - Each browser constrains the Emscripten environment somewhat differently when it uses a lot of resources: if you are having problems on your second load of a notebook, copy the URL, close the tab, and open a new one with the same URL to work-around. *I have not taken efforts to constrain memory leaks yet as recommended by Pyodide.* 2f:Tfa0,

kglab in ObservableHQ

ex0_0 Data Sources ex1_0 Building a graph in RDF using rdflib ex1_1 Leveraging the kglab abstraction layer ex2_0 Build a medium size KG from a CSV dataset ex3_0 Interactive graph visualization with pyvis ex4_0 Running SPARQL queries ex5_0 SHACL validation with pySHACL ex6_0 Graph algorithms with networkx ex6_1 Discover community structure using igraph and leidenalg (*ref only)ex7_0 Measurement and inference ex7_1 Statistical Relational Learning with pslpython (*ref only)ex8_0 Vector embedding with gensim (*ref only)kglab generic

30:[{"id":"23cb4ac5afd2157ddcdf9ec30a6320190a2896025866634bb1ac5ed2b1e8eb178f789dbcfe3042f8310ec9d3b573a60f3e5244dd8ef5458c5d9d877eadc14ea6","href":"https://static.observableusercontent.com/files/23cb4ac5afd2157ddcdf9ec30a6320190a2896025866634bb1ac5ed2b1e8eb178f789dbcfe3042f8310ec9d3b573a60f3e5244dd8ef5458c5d9d877eadc14ea6","name":"image.png","type":"image/png","lastModified":1640381789712.82,"size":492811},{"id":"d74360307a26b04fb7155a5f8844e79a074aa3665d98ef00ae55450b93abb13f116f19f731253d1094cbb1bb64191aca8d95c255bc6f066468ccf672e0b5c0b3","href":"https://static.observableusercontent.com/files/d74360307a26b04fb7155a5f8844e79a074aa3665d98ef00ae55450b93abb13f116f19f731253d1094cbb1bb64191aca8d95c255bc6f066468ccf672e0b5c0b3","name":"kglab_visual_query_editor.gif","type":"image/gif","lastModified":1640843441997.485,"size":1757909}] 2c:{"id":"bd10e2614309b85a","slug":"kglab-in-observablehq","title":"kglab in ObservableHQ","thumbnail":"30e3a88839677b04d2953a570336393586ad2a94d76a7e4df57d0317537a2844","access_level":"public","create_time":"$D2021-12-24T21:35:04.066Z","update_time":"$D2022-01-03T18:02:21.467Z","view_time":null,"first_public_time":"$D2021-12-25T00:58:21.467Z","has_unpublished_edits":null,"fork":{"id":"662080314eb2cb2c","slug":"blog-post","title":"Blog Post","owner":{"id":"f35c755083683fe5","login":"observablehq","name":"Observable"}},"stars":7,"starred":false,"owner":{"type":"team","id":"5cd4a42aead2029f","login":"thadk","name":"Thad Kerosky","avatar_url":"https://avatars-next.observableusercontent.com/2f2413891eb0ff4669b49685946f849544a1e49e7047923d158638bc30666281","tier":"starter_2024"},"version":320,"publish_version":320,"role":null,"authored":false,"authors":[{"type":"individual","id":"be3e6e81af18a319","login":"thadk","name":"Thad Kerosky","avatar_url":"https://avatars-next.observableusercontent.com/2f2413891eb0ff4669b49685946f849544a1e49e7047923d158638bc30666281","tier":null}],"members":[{"type":"individual","id":"be3e6e81af18a319","login":"thadk","name":"Thad Kerosky","avatar_url":"https://avatars-next.observableusercontent.com/2f2413891eb0ff4669b49685946f849544a1e49e7047923d158638bc30666281","tier":null,"role":"editor"}],"comments":[],"custom_thumbnail":true,"trashed":false,"lapsed":null,"subscribed":null,"forks":1,"collection":{"id":"aa93231193d34e3a","slug":"kglab-js-alpha-example","title":"kglab.js (alpha) examples","owner":{"id":"5cd4a42aead2029f","login":"thadk","name":"Thad Kerosky"}},"type":"notebook","body":{"title":"kglab in ObservableHQ","cells":[{"id":0,"value":"# kglab in ObservableHQ\n## Graph Data Science toolkit in Javascript (unofficial)","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":4,"value":"\n

","pinned":false,"hidden":false,"mode":"html","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":140,"value":"html`${run()}`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":41,"value":"### What is `kglab`?","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":8,"value":"The [kglab package](https://derwen.ai/docs/kgl/) ([github](https://github.com/DerwenAI/kglab)) provides a simple abstraction layer in Python [and Javascript] for building knowledge graphs.\n\nThe main goal is to leverage idiomatic Python [within Pyodide] for common use cases in data science and data engineering work that require graph data, presenting graph data science as an emerging practice.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":78,"value":"### What is this notebook collection? ⚗️","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"","database":"$undefined"},{"id":35,"value":"I brought `kglab` from Python to Javascript using [Pyodide](https://pyodide.org) in these example tutorial notebooks, which had originally been [Jupyter notebooks](https://github.com/DerwenAI/kglab/tree/main/examples). \nNow apply `kglab` in the browser!\n\nI have made small changes in the example notebooks to support ObservableHQ and the [Pyodide Observable layer](https://observablehq.com/@thadk/pyodide-18) and the [Pyodide type conversions](https://pyodide.org/en/stable/usage/type-conversions.html).\n\nFork an example to follow it today.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":100,"value":"### What are the limitations of `kglab` in Pyodide here?","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"","database":"$undefined"},{"id":118,"value":"- First off: **SPARQL queries, SHACL constraints, graph closure, and `networkx` work great.** They'll start as wrapped python like:\n\n~~~py\npy`\nimport kglab\nkg = kglab.KnowledgeGraph()\nkg\n`\n~~~\n\n- The **startup time is significant** to download and load the Pyodide python modules behind `kglab`. Expect 40-120 seconds and 105mb of data transfer.\n- Safari 15.2 seems to have trouble with Pyodide and is not recommended. Try **non-mobile Firefox or a Chrome-based browser** until the memory leaks are addressed. Mobile browsers likely do not allot sufficient memory for kglab.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"","database":"$undefined"},{"id":207,"value":"html`${run()}`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":247,"value":"#### Some example notebooks and pieces of `kglab` don't run fully: These are marked as \"reference only\" with an `*`:\n- PSL and `pslpython` need a Java Virtual Machine to run and do their statistical relational learning. The JVM is not yet available in Emscripten.\n- `gensim`, a vector embedding library, needs attention by Pyodide developers to be supported.\n- pytorch and `igraph` rely on GPU libraries and cannot yet be supported in Pyodide.\n- pytorch references have been removed from the Pyodide-compiled version of `kglab` v4.2. The `pyvis` whl file was re-zipped to remove references to IPython which crashed.\n- Though they are available in Observable, Arrow library and Parquet format are [not yet supported](https://github.com/pyodide/pyodide/issues/802) in Pyodide.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"","database":"$undefined"},{"id":251,"value":"$2e","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"","database":"$undefined"},{"id":216,"value":"html`${run()}`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":291,"value":"# So What?\nIt should allow more interaction than straight python. For instance, here's a bonus [SPARQL code mirror](https://www.npmjs.com/package/@triply/yasgui) paired with one of kglab's APIs to actively [visualize a SPARQL query as a graph in ex4_0](https://observablehq.com/@thadk/ex4_0-running-sparql-queries?collection=@thadk/kglab-js-alpha-example).","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"","database":"$undefined"},{"id":289,"value":"FileAttachment(\"kglab_visual_query_editor.gif\").image()","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":133,"value":"run = () => html`\n `","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":268,"value":"## Raw List of Notebooks in Series:","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"","database":"$undefined"},{"id":264,"value":"$2f","pinned":false,"hidden":false,"mode":"html","author":"$undefined","summary":"$undefined","output":"","database":"$undefined"}],"files":"$W30","stdlib":"1","titling":"auto"}} 2d:"$2c:body"