Joss

@ryan - On a slightly different note, I have just installed the qwen-2.5-32b instruct, and it is running fine on my system via ollama (better at balancing tham LM Studio, it appears). But I have been setting it up for gramma checking, using modelfile to customise it. Basically, it tells it to act as an editor, not to get too creative, then gives some specific instructions over strange words, grammar, etc. It then creates a model for just that purpose. Could something similar be run for AgentTools so that the actual model has specific hard-wired PW knowledge in advance? I don't know enough to say if this would speed things up or anything like that. But it certainly helps with Word as it knows in advance character names etc, without it having to be prompted at the time.

Head north about a thousand miles.... 😀

I select an image, click on submit. It sends me to this screen: So I choose a file again (this is a nice small file) and click on Continue and end up: It doesn't like me!!! This is what I was trying to upload:

The error that came up during install was that makedir failed to create /site/assets/at/jobs/done, and failed to create the htaccess within "done", and the same with "failed", "pending," and "running." Note - I am using laragon on windows 11 rather than a unix server.

Well, I might just be back a bit too. But I have a problem. I decided to update my little profile photo. It happily got rid of my old one, but won't let me upload my new one!!!!! Grrr. And so good to see my brother in fantasy @Soma back amongst the brave!

@ryan that is exactly the one I have been playing with today. The 14b version. Unfortunately, qwen 3 coder only comes in 30b and 40b, which is pushing it for me. I might try it again with Ollama. It suffered with LM Studio, but Ollama is meant to be better at dealing with local problems like not having a huge stack of GPUs lying around! The hard bit is the prompting. I was trying to get it to do a very simple page. Create a data set, then search it and spit out a single result. It was determined to ignore the search result and just loop the dataset. Grr! EDIT: By the way, when I installed, it failed to create the subdirectories in jobs. I have hand created them and copied the htaccess file from jobs into each one. Is that right? Anyway, here is the reason for the large box, and the view out of my studio window! Let me know when you start a podcast and need an editor!

Thank you, kind sir! (The Pirate bows!) I will have a play. I have been using qwen instruct but also have a few other models. I am a little limited because I only have an RTX 5080. Google's AI swore blind I could run a 31b model with a fair amount of offset (I have 128gb Ram), but I last saw the machine crawling out from under the desk, sweating and groaning, so I am playing with the 14b instead with 1.5b for autocomplete! I will get there. Honest! I still guess at all this stuff.

Hi Ryan! Will this, or something, talk to Visual Studio Code via Continue and Ollama? (I have LLMs locally). I am still trying to get my head around what is possible for site development with AI and Processwire - not helped that I am so out of date now!!! Joss

Just for fun, I asked....

@ryan Ta! Since I am in very early dev, I added the line to the config (I didn't have a dbCharset in there) and then duplicated the field. That did it.

Hi Ryan! Good to see you too. This is on my local machine using Laragon running MySQL 8. I ran the query: SELECT * FROM information_schema.SCHEMATA S WHERE schema_name = "aelwenjones"; And it returned: CATALOG_NAME def SCHEMA_NAME aelwenjones DEFAULT_CHARACTER_SET_NAME utf8mb4 DEFAULT_COLLATION_NAME utf8mb4_general_ci But if on phpMyAdmin I select one table and click on show create I get: CREATE TABLE `field_article_text` ( `pages_id` int unsigned NOT NULL, `data` mediumtext NOT NULL, PRIMARY KEY (`pages_id`), KEY `data_exact` (`data`(250)), FULLTEXT KEY `data` (`data`) ) ENGINE=MyISAM DEFAULT CHARSET=utf8mb3 I notice on my host, however, that most of my schemas are latin1, so I need to make sure they aren't, as it were, for this table.

Okay, solved it. I put this in the JSON field: { "quickbars_insert_toolbar": "pwimage h2 h3 emoticons charmap" } However, I have another problem now. If I insert and emoticon then save the page, I lose any text AFTER the emoticon and the emoticon doesn't display.

How can I customise the three Quick Toolbars? The new line context bar has two buttons - Image and Table. However, I don't want the table and the image button just opens up my local file system rather than the normal pwimage dialogue. I also want to add a couple of extra buttons on there ideally like embed or heading etc.

Eleven years later, but thanks for that, Ryan! It was exactly the explanation that I needed.

Yep! And I think I did it just before I had all these backup woes.