| 00:32:23 | * | mrinsane joined #nim |
| 01:17:21 | * | rockcavera quit (Remote host closed the connection) |
| 01:25:38 | * | sdw3911 joined #nim |
| 01:44:12 | * | SchweinDeBurg quit (Quit: WeeChat 4.8.0-dev) |
| 01:50:22 | * | hygo quit (Ping timeout: 256 seconds) |
| 01:52:27 | * | hygo joined #nim |
| 03:30:14 | * | sdw3911 quit (Quit: WeeChat 4.7.2) |
| 03:37:35 | * | SchweinDeBurg joined #nim |
| 04:49:50 | * | Goodbye_Vincent1 joined #nim |
| 05:29:48 | * | Goodbye_Vincent1 quit (Ping timeout: 264 seconds) |
| 05:31:13 | * | fallback quit (Quit: IRCNow and Forever!) |
| 05:32:14 | * | fallback joined #nim |
| 05:32:47 | * | Goodbye_Vincent1 joined #nim |
| 06:01:52 | * | mrinsane quit (Ping timeout: 268 seconds) |
| 07:05:29 | * | ormiret quit (Ping timeout: 252 seconds) |
| 07:05:51 | * | ormiret joined #nim |
| 08:59:50 | * | hygo quit (Read error: Connection reset by peer) |
| 13:44:30 | * | xet7 quit (Remote host closed the connection) |
| 13:49:20 | * | xet7 joined #nim |
| 14:12:20 | * | krunsh joined #nim |
| 14:19:07 | * | xet7 quit (Remote host closed the connection) |
| 14:20:50 | * | xet7 joined #nim |
| 14:21:57 | FromDiscord | <asviest> @anuke ↵I was reading the code of fau and came across this:↵https://github.com/Anuken/fau/blob/b935b17ff5bfcbde22d07cc544d5869d00813012/src/fau/mesh.nim#L4↵Actually, this macro hack is unnecessary. It occurs because ident is used instead of a symbol, and `quote do` probably doesn't bind the symbol. You should use `bindSym` or perhaps `genAst` (from std/genasts), or a combination of both, to bind symbols to the macro. This way, the |
| 14:23:13 | FromDiscord | <Zoom> pmunch\: https://nimble.directory/ is down, looks like |
| 14:29:12 | FromDiscord | <anuke> In reply to @asviest "<@123539225919488000> I was": If I were to just replace a block of `quote do` code containing non-exported GL calls with `genAst`, would that fix the problem? I've heard of genAst, but I wasn't aware of this use case |
| 14:36:54 | FromDiscord | <asviest> In reply to @anuke "If I were to": Yes https://media.discordapp.net/attachments/371759389889003532/1484561373064204358/IMG_20260320_173634_036.jpg?ex=69bead05&is=69bd5b85&hm=b596974c8d4d712fef2831afc8f4156da4685942e841d9bf3886ada747bba27d& |
| 14:41:02 | * | xet7 quit (Ping timeout: 256 seconds) |
| 14:44:11 | FromDiscord | <asviest> In reply to @anuke "If I were to": Also, sometimes you need to use `genAst do`, it's required in the same cases where `quote do` is necessary and you can't just use a simple quote (for example, `node.add quote:` invalid syntax but with `quote do` it's correct). Btw, I often move `genAst` into a separate variable since I'm not a big fan of the `do` operator. |
| 14:53:08 | * | xet7 joined #nim |
| 15:41:26 | FromDiscord | <mratsim> By the way @planetis_m for pdf-ocr, how hard would it be to extract images as well and insert them in the Markdown file?↵↵I started to do that in a PR for the papers I used because it helps LLMs with context but I'd like graphs as well and it seems in your case you extract text only https://media.discordapp.net/attachments/371759389889003532/1484577616378855546/image.png?ex=69bebc25&is=69bd6aa5&hm=b1d3313e4e043f32a3dc441a403d5f224187a |
| 15:42:19 | FromDiscord | <mratsim> in my case I can run the model locally. PaddleOCR for example is just 3GiB of VRAM or so |
| 15:56:25 | FromDiscord | <planetis_m> In reply to @mratsim "By the way <@915614219364368394>": Not at all hard but I don't know the specific pdfium api |
| 16:06:33 | * | mrinsane joined #nim |
| 16:12:40 | * | mrinsane quit (Ping timeout: 245 seconds) |
| 16:20:36 | FromDiscord | <mratsim> In reply to @planetis_m "Not at all hard": probably hidden somewhere in https://github.com/PaddlePaddle/PaddleOCR |
| 16:25:13 | * | xet7 quit (Ping timeout: 248 seconds) |
| 16:37:19 | * | xet7 joined #nim |
| 16:38:53 | FromDiscord | <bung8954> In reply to @mratsim "By the way <@915614219364368394>": ocr model like paddleocr contains layout infomation, it's easy extract pdf image with bounds info, not sure if paddleocr produce json or not, it would be even simpler. |
| 16:39:03 | * | mrinsane joined #nim |
| 16:40:28 | FromDiscord | <bung8954> I have fresh new qianfunocr(from baidu) model runs in pure nim today |
| 16:40:53 | FromDiscord | <bung8954> (edit) "pure" => "" | "nim ... today" added "(with binding)" |
| 16:56:30 | FromDiscord | <mratsim> In reply to @bung8954 "ocr model like paddleocr": yes paddleocr produces json if you ask for it |
| 16:56:41 | FromDiscord | <mratsim> at least Baidu's website allows either markdown or json download |
| 16:57:38 | FromDiscord | <bung8954> In reply to @mratsim "yes paddleocr produces json": yeah, qianfunocr does this too |
| 18:41:06 | * | xet7 quit (Remote host closed the connection) |
| 18:42:50 | * | xet7 joined #nim |
| 19:35:25 | FromDiscord | <planetis_m> Well you don't need complex layouting to extract images from a PDF, its a low level operation |
| 19:36:17 | FromDiscord | <planetis_m> No idea about inserting them |
| 19:37:33 | FromDiscord | <planetis_m> But I did check mineru and it was using another llm at some point for postprocessing, maybe it uses it for writing the image links to the final output |
| 20:15:48 | FromDiscord | <Zoom> planetis\: Man, couldn't you have chosen your own lib name prefix \:P ? |
| 20:18:23 | FromDiscord | <Zoom> You're also violating GPL as your code is obviously a derivative work, unless you make it GPL too |
| 20:33:55 | FromDiscord | <planetis_m> In reply to @Zoom "planetis\: Man, couldn't you": Well it's a through away experiment, don't really plan on using it or developing it further. |
| 20:34:51 | FromDiscord | <planetis_m> But is it a derivative? Does the code have any similarities? |
| 20:35:43 | FromDiscord | <planetis_m> I am going to use GPL if you want me to but it barely matters |
| 20:43:21 | * | xet7 quit (Ping timeout: 272 seconds) |
| 20:45:16 | * | beholders_eye joined #nim |
| 20:46:51 | FromDiscord | <bung8954> In reply to @planetis_m "Well you don't need": I know, but mratsim want insert to markdown, I assume he want insert to proper position |
| 20:55:15 | * | xet7 joined #nim |
| 20:55:18 | FromDiscord | <Zoom> Unfortunately, it the amount of similarities doesn't matter much, but there are plenty. I understand it's not a big deal, but it should be GPL. I suspect you used an LLM and as great of a whitewasher it is, it still read the code.↵(@planetis_m) |
| 20:55:36 | FromDiscord | <Zoom> In any way, I'm flattered you took any interest in the lib at all. |
| 21:04:07 | FromDiscord | <mratsim> @_araq when did Nim start supporting higher-kinded type? My LLM just vibecoded this and .... it works? https://media.discordapp.net/attachments/371759389889003532/1484658821765922956/image.png?ex=69bf07c6&is=69bdb646&hm=995f49dde17ad2096634fb1d60bb50e688cde42db011ca1312b5f805602eb80c& |
| 21:09:29 | FromDiscord | <shalokshalom> ~ The power of friendship ~ |
| 21:09:57 | FromDiscord | <shalokshalom> ✨ 🦄 |
| 21:13:49 | Amun-Ra | mratsim: https://play.nim-lang.org/#pasty=WHvTskhN |
| 21:15:01 | * | mrinsane quit (Ping timeout: 272 seconds) |
| 21:16:25 | Amun-Ra | it's just a template type keyword |
| 21:16:30 | Amun-Ra | name* |
| 21:34:23 | FromDiscord | <mratsim> In reply to @Amun-Ra "it's just a template": for a long time I'm sure it didn't work with generics |
| 21:36:08 | FromDiscord | <mratsim> In reply to @Amun-Ra "it's just a template": Yeah can confirm: https://github.com/nim-lang/RFCs/issues/5 |
| 21:38:50 | FromDiscord | <mratsim> In Arraymancer I couldn't declare `myProc[TT: Tensor[T]](foo: TT)` so I had just TT everywhere https://media.discordapp.net/attachments/371759389889003532/1484667558148968704/image.png?ex=69bf0fe9&is=69bdbe69&hm=499cbc76e1df28eb9a1545ef9b831909740c8e6c04fb5173683f128ae2679c93& |
| 21:39:07 | FromDiscord | <planetis_m> In reply to @Zoom "Unfortunately, it the amount": Seriously there are no similarities. You are getting hang up on the cozy- prefix |
| 21:39:44 | FromDiscord | <planetis_m> Read the source code |
| 21:42:02 | FromDiscord | <planetis_m> As I said I don't plan to develop it further, I rather delete it. |
| 21:53:38 | * | mrinsane joined #nim |
| 21:54:30 | FromDiscord | <planetis_m> The original thought was to extent your lib to work with env and config.json files, but then decided I want something as simple as possible. Is it worth replacing 100 loc of using std/parseopt directly? IMO no its not worth it |
| 21:56:52 | FromDiscord | <planetis_m> The name stuck |
| 22:10:45 | * | xet7 quit (Ping timeout: 272 seconds) |
| 22:22:46 | * | xet7 joined #nim |
| 22:42:14 | FromDiscord | <Zoom> I read the code, both files and both commits. Do you insist it's not derivative and you/your agent didn't read the lib at all? Similarity is not the point. Anyway, I get you just experimented, but respecting the license would be nice.↵(@planetis_m) |
| 22:47:16 | FromDiscord | <Zoom> Regarding the name, it's not like I have any special rights to it, but chose my own little namespace and would like to retain it. It's not even that presentable, so I hoped no one would mind. I hope we could maintain mutual respect and camaraderie in our small community here. Like, me calling my lib `parsy` or `configgy` wouldn't be nice to Treeform so I won't do it. |
| 22:50:00 | FromDiscord | <Zoom> Regarding the name, it's not like I have any special rights to it, but I chose my own little namespace and would like to retain it. It's not even that presentable, so I hoped no one would mind. I believe we can maintain mutual respect and camaraderie in our small community here. Like, me calling my lib `parsy` or `configgy` wouldn't be nice to Treeform so I won't do it. |
| 22:53:45 | FromDiscord | <choltreppe> how would you interact with a db in prologue? Im currently just using the db_connector. does that behave well with the async model of prologue? should I maybe switch to mummy? |
| 23:11:37 | * | xet7 quit (Ping timeout: 248 seconds) |
| 23:23:11 | * | xet7 joined #nim |
| 23:45:09 | * | mrinsane quit (Ping timeout: 255 seconds) |