Clojurians Log v2

Hey @udit! Nice to see you here :simple_smile:

Hey everyone! 👋

I wonder how many people here write clojure as a part of their jobs vs not as a part of their jobs. Are there many "clojure companies" India? :simple_smile: I only know of helpshift, nilenso, quintype.

Hey @srihari. I Googled slack later that day. Ended up logging on to a couple of channels. It is pretty nifty.

Yup :simple_smile: We have a public #learn channel at nilenso. You should join! http://learn.nilenso.com.

@srihari: If an imaginary startup counts, I do clojure/script regularly. Not making any money yet. Like, no plans too as of right now.

@jaju: nice! :simple_smile: what are you working on?

Can’t exactly tell what I am building right now - not exactly a secret, but not quite well-defined too. But building a tool for curating and organizing information online. Standard stack on the server-side, electron-shell for the client. reagent, core.async amongst the usual suspects, and using cursive.

Oh, nice!

@jaju open source? :simple_smile:

Curating and organizing information of a particular kind? Or generic?

@jaju: How are you finding Reagent?

@srihari: Not open-source. At least not yet. :simple_smile: The focus is an individual user - so, would want it to be decided/defined by the user what topics/sources. But it’s in a state of flux. I abruptly moved from a server/client model to a fat-client (electron) because of the thought of deployment/distribution complexities that I could imagine. I was lucky to have found https://github.com/ttezel/twit - twitter streaming in JS (nodejs), so the decision was quite an impulsive one. It helped that the server code was in clojure, so the transition was not much of a pain. :simple_smile: core.async works almost the same way in clojure/script, which handled the streaming...

@steven: It’s a godsend. Especially for someone like me who was way afraid of UI programming and had never touched JS in any manner worthy of speaking, I jumped into browser stuff only because of the confidence that clojurescript gave - and then, react made good its promise of keeping the code so simple! I love the reagent experience - I never considered react-js directly, obviously, but reagent is way too easy for my small brain if I compare it with om.

@jaju interesting. I’ve been working on making an open database for indian classical music, and I’m getting the data by scraping the internet. So, I end up curating a lot of this information. If and when I open this out to the public, I expect people will curate the information already present, and add more. Before that though, I imagine a machine doing deduplication, de-noising, etc. It would be pretty hard for a person to comb through all the data. I wonder if there are libraries meant to do this. I don’t know of any. Also, right now I’m writing custom scrapers for each website. It would be really cool if the scraper/machine could automatically tag and categorize the information it scrapes. The input would maybe be a bunch of keywords and some rules. Again, I’m not sure if something like this exists out there already. Is anything that you’re doing remote related to any of this? :simple_smile:

@srihari: Sounds interesting! Automated/intelligent scraping is one of the problems I am looking to tackle. But the domain isn’t fixed, like music in your case. FWIW, using machine-learning, and corrective user-feedback, may be an option to try. I have a vague desire to try that. https://github.com/karpathy has some interesting JS stuff on ML - which also was a reason for me to jump to node-js, because I may like to try that out too some day. :simple_smile:

Yes! That. I’m just starting my ML course now though, and I have a long way to go before I can do this.

:simple_smile:

Are your scrapers public by any chance, @srihari?

@jaju: There’s some here: https://github.com/ssrihari/ragavardhini/tree/master/src/movertone/scripts

@jaju: But that repo has music synthesis too, so I’m moving the scrapers and the open database part out into https://github.com/ssrihari/kosha

With what information I was able to mine with my initial spike, I built http://r4g4.com. This works, and I use it on a daily basis, but it’s incomplete, and shoddily written.

Cool! I did skim over your func-conf talk video quite some time back - some day, I’d like to understand music better. Nice!

@jaju :simple_smile:

Looks pretty neat! For want of a better phrase, “it’s greek and latin” to me. (It’s all Carnatic to me wouldn’t have the same punch as much as the latin and greek phrase wouldn’t sound amusing to the Greeks et al). I just did a random search “bhairavi” and my guess is that the results must be pretty useful for folks who know that stuff.

Hehe, yeah