Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Aug 11 2021 20:52
    @RubenVerborgh banned @mikeadams1
  • Jan 04 2021 20:23
    @RubenVerborgh banned @WebCivics_twitter
  • Jan 04 2021 20:18
    @RubenVerborgh banned @SailingDigital_twitter
  • May 27 2019 06:08
    User @Mitzi-Laszlo unbanned @in1t3r
  • May 23 2019 06:49
    @Mitzi-Laszlo banned @in1t3r
  • May 16 2019 09:49
    @Mitzi-Laszlo banned @mediaprophet
  • Feb 01 2019 22:04
    User @melvincarvalho unbanned @namedgraph_twitter
  • Feb 01 2019 21:49
    @melvincarvalho banned @namedgraph_twitter
William Collins-Broza
@mrhegemon
@konobi there is. Our system mainly runs on a search engine that we optimize to bring queries down to about 2 seconds. https://lifescope.io/getting-started/ I'm happy to go over all of the current data infrastructure.
Fred Gibson
@gibsonf1
2 seconds seems slow - what are you searching through in that time?
William Collins-Broza
@mrhegemon
We do searches with filters for "who, what, when, where" events happened in a person's history. and then surface related people, places, things, content. This is done by a series of MongoDB collections with graph relations. We compose and execute the query using a GraphQL ODM we built from exiting projects. The DB is a sharded and replicated kubernetes cluster running on a per-user isolated Mesos environment. https://github.com/LifeScopeLabs/lifescope-api These clusters currently run in the cloud but we are rebuilding the architecture so the nodes can exist on user devices. EG between a person's windows/mac/linux/ios/android devices. https://docs.mongodb.com/stitch/mongodb/mobile/mobile-features/
By no means married to MongoDB, just can't find ANYTHING else that checks all the boxes we want.
William Collins-Broza
@mrhegemon
We can get the speed up 10x from 2 seconds but that is our current tolerance for what we think the user will put up with. Once we get optimized, I'd like to do instant search under 0.3 seconds.
Fred Gibson
@gibsonf1
I think key to success there is a good conceptual index on top so you don't have to brute force search through everything
For example, Allegrograph is able to keep in memory indexing for 1 billion triples (on a big machine)
William Collins-Broza
@mrhegemon
Would love to give it a shot.
We currently use graphql-compose as our abstraction layer. It's not perfect, but gets the job done and has let us experiment with SQL/ElasticSearch backends as well. https://graphql-compose.github.io/ Up for writing a mapper to Allegrograph and/or generic Solid store/search with me?
It's an ODM (object document mapper). They have good and bad sides, but never been able to do any abstraction that was better.
Also, I'm looking for any work and documentation build to bridge GraphQL or any HTTP APIs w/ Solid.
Fred Gibson
@gibsonf1
I think if you make the leap to the conceptual "graph" way, use sparql etc, you will get results impossible to do with the more traditional appraoches
But it's a cultural change. It's much easier for people to move from 20th century databases to a property graph (like neo4J) than to a semantic graph - it's a very different way of thinking, and provides incredible advances
@RubenVerborgh gave a very good talk on the order of magnitude difference between graphql and sparql in this short talk: https://www.youtube.com/watch?v=LUF7plExdv8
William Collins-Broza
@mrhegemon
Yea, I do love me a semantic graph but I've never built one for production. Hard to sell on any projects. Everyone in the community I have built wants a GraphQL API w/ OAuth 2. I have thought of GraphQL to SPARQL, but my team isn't super strong on data principles, so I am stuck with all the engineering if I want to make the leap.
Fred Gibson
@gibsonf1
Remember when everyone wanted Java ?
There is no doubt that to fully capture the capability of the semantic way is a major cultural change - it took me a few years to make the transition myself. In any kind of engineering field, cultural change is hard. However, in the world of innovation, the people that make the leap first reap the rewards, and this technology is now ready for true scale
William Collins-Broza
@mrhegemon
LOL, well my current plan is to keep the GraphQL schema and datastore fairly clean so I can move the data around to better structures as it becomes viable. Apache Jena vs rdf4j looked like no winners to me. Down to learn about what you got going on.
Half my team is JavaScript hipsters I can't get to learn Python. Team, if you are reading this, I still love you.
Fred Gibson
@gibsonf1
No problem - one of our features is conceptually graphing the internals of pdfs (we get 50K statements out of set of drawings for example)
Javascript is fantastic on the browser side. For the seriously heavy lifting, not quite as much. Although the solid team has created the best javascript I've yet seen - so it's not so much the language as the approach to thinking about the problem that counts
William Collins-Broza
@mrhegemon
https://github.com/solid/solid-platform I have been trying to compare solid platform implementations, but it is hard to figure out the status of support of features. Especially performance, scalability, security, LDP Search capabilities, and data import from other platforms.
Still hoping someone can help me understand the status of current solid platform implementations!
Ted Thibodeau Jr
@TallTed

:point_up: December 29, 2018 2:34 PM
@mrhegemon - That question is hard to address in a chat-space. I'd suggest starting a page in the solid wiki (https://github.com/solid/node-solid-server/wiki), likely with some connection to the page of Solid compatible servers (https://github.com/solid/node-solid-server/wiki/Running-Test-Implementations-of-Solid-compatible-servers), and setting up a table of the sorts of features/functionality you want to compare (probably best as column headers) vs the implementations of which you're aware (probably best as row headers). Depending on the depth of discussion, It might make sense to treat each column as a section, and break such table into a textual stream...

As to MongoDB vs Allegrograph vs TDB (the storage component of Apache Jena) vs RDF4J vs Virtuoso vs other back-end stores. It may be worthwhile to also consider their basic architectures, vis-a-vis single-model vs multi-model and single-query-language vs multi-query-languages -- https://community.openlinksw.com/t/tabulated-relational-database-management-systems-rdbms-comparison/274

Mark Hughes (happybeing/theWebalyst)
@happybeing
I'm interested in tools suitable for client side / static apps for use will Solid, SPARQL etc, and wonder if anyone has experience of or comments on React-static and gatsbyjs.org. Both seem suitable at first glance and could be used with @RubenVerborgh's React components I think (so LDflex very neatly) but I'm new to these frameworks having not done any serious front end stuff myself. Gatsby.js seems state of the art and trendy, and oriented towards GraphQL but I'm not sure if this is an advantage or taking us away from easier integration with RDF/SPARQL, and any other pros and cons. Any thoughts? I'm going to repost this to the forum, so reply there if you like: https://forum.solidproject.org/t/react-static-and-gatsby-js-frameworks/964?u=happybeing
Jean Lucas
@aleattorium_twitter

can I ask for help here in this room?

Just created an account and I get this message:

No permission to access this resource
You are currently logged in as jeanlucas, but do not have permission to access https://inrupt.net/jeanlucas.
Jean Lucas
@aleattorium_twitter
This happened on Windows with Chrome latest version and also on Firefox. Also on an Android with Firefox and Chrome. Looks like it's not a client error.
William Collins-Broza
@mrhegemon
@TallTed Thanks for the guideposts. Will attack this next week.
konobi
@konobi
@theWebalyst there's an npm module called graphql-to-sparql.js that's been created by someone here that could be used in conjunction with solid to provide an endpoint for gatsby
Hariom Dwivedi
@Hari987_gitlab
Happy New Year to Everyone in Advance.
Mark Hughes (happybeing/theWebalyst)
@happybeing
Thanks @konobi, noted.
Arne Hassel
@megoth_twitter
@Hari987_gitlab happy New year! ^_^
Eduardo Ibacache Rodriguez
@eduardoinnorway
GOTT NYTT ÅR, FELIZ AÑO NUEVO, สวัสดีปีใหม่, НОВИЙ РІК, SZCZĘŚLIWEGO NOWEGO ROKU, HAPPY NEW YEAR. May 2019 be a awesome Solid year.
tag42git
@tag42git
Happy New Year :pig:
proimages
@proimages_twitter
Hi All! New to the solid framework and just diving in..I come from self taught LAMP and mostly to code solve issues with open source apps going back many years. Not a coder more UI and concept developer. I have an older project that needs a complete revamp and used a server model similar to PODS for members personal info. ( pods are also a wordpress plugin FYI) Hoping I can covert to PODS and reuse some of my older code. Any road maps -videos etc for non coders yet ? and is Solid ready for a complete music site system..? Thanks for any info links and guidance!
Fred Gibson
@gibsonf1
Is there any ontology to differentiate account based information that, if edited by the user, will disrupt the functioning of the account? For example, to prevent a users from editing the solid:account link between their web id and root storage.
Is it possible to put ACL on a triple as opposed to a node? So that system based triples, like solid:account would be read only even to the account owner?
Fred Gibson
@gibsonf1
An ACL on a property rather than on the triple that is, perhaps a class that properties could belong to on which an acl placed would control the ACL on the object of the instantiated property
Matthias Evering
@ewingson
hey there fellow coders, new year here in germany, I wish everyone a good one !
Fred Gibson
@gibsonf1
Frohes Neues Jahr!
Matthias Evering
@ewingson
danke @gibsonf1 dir auch ! I' m trying to set up a solid server at solidweb.org and having issues: 1. mod_rewrite to force nonwww and https works but doesn' t seem to work with the reverse proxy needed. solution: this can also be done in apache2.conf vhost-section 2. accidentally I have touched fullchain.pem. started manually the server works but not as systemd-service. debugging complains and claims not found fullchain.pem although its there....... no solution yet... will dive into the solid-thang and this chat and help where I can because the project is FANTASTIC
Justin Bingham
@justinwb
@ewingson jump into the node-solid-server chat at https://gitter.im/solid/node-solid-server and someone should be able to help you track it down. you can also post some more detail about your issue at https://forum.solidproject.org
Tim McIver
@tmciver
Hey folks! Been away from solid for a while but trying to get back into it. I've been looking at Apache Marmotta and have been wondering if it is Solid-compatible. It appears to use WebID and has a notion of profiles and ACLs but I'm not sure if they're the same as solid. Anyone know if it is?
BTW, Happy New Year!
Dirk Roeleveld
@dirkesquire
Happy New Year!
Question: Is yarn add rdflib an okay way to install rdflib.js? Problem is I see it's install 85 other library dependencies
Eduardo Ibacache Rodriguez
@eduardoinnorway
@dirkesquire That is how JS ecosystem works today, I can’t say why 85 dependencies on rdflib.js but usually you get a ton of dependencies adding a js library, It has nothing to say if you have a propper build, where you shake off anything not used on production.
Dirk Roeleveld
@dirkesquire
Yes its okay, just wanted to check 'rdflib' was the correct name to use as even capitalisation matters with npm install (example jquery vs jQuery).
I did a compare against npm install rdflib, and it installs the exact same amount of dependencies, so I am confident that 'yarn add rdflib' is correct now.
Eduardo Ibacache Rodriguez
@eduardoinnorway
Webpack has become really good this past year, Tree shaking etc.. is awesome and you can get really slim packages if you do some configuration.
I also recommend to check out https://github.com/prettier/prettier