Weekly update – 20 May 2022

[ryan]

We had a smooth rollout of the new main/master version 3.0.200 last week (read all about it here). If you haven't upgraded yet, consider doing so, this new version is a great upgrade.  I'm going to add the 3.0.200 tag shortly after I finish this post, which should trigger other services (like packagist) to upgrade. 

I've had a new client project in the pipeline that I've been waiting to start till the new main/master version was out, so this week I started that project. Pete and I are working together on it, like we've worked on others before. It involves taking a popular WordPress site and rebuilding it completely in ProcessWire. I've done this a couple times before, but this time it's bigger and broader in scope. I always find the large site conversions to be great learning experiences, as well as great opportunities to show how ProcessWire can achieve many things relative to WordPress, in this case.

At this stage, I'm having to spend a lot of time in WordPress just to get familiar with the content, fields, etc., as well as in the theme files (php and twig). The more time I spend in these, the more excited I get about moving it into ProcessWire. For this particular site, moving from WordPress into ProcessWire is going to result in a big boost in efficiency, maintainability, and performance. Part of that is just the nature of PW relative to the nature of WP. But part of it is also that the WP version of the site is kind of a disorganized patchwork of plugins, code files, and 3rd party services, all kind of duct taped together in an undeniably confused, undisciplined and fragile manner. (Though you'd never know it by looking at the front-end of the site, which is quite nice). This has been a common theme among WordPress sites I've dug into. Though to be fair I don't think that's necessarily the fault of WordPress itself. 

I always enjoy taking a hodgepodge and turning it into an efficient, performant and secure ProcessWire site. I love seeing the difference it makes to clients and their future by taking something perceived as a "necessary liability to run the business", and then turning it into the most trusted asset. I think the same is true for a lot of us here. We love to develop sites because it's an opportunity to make a big difference to our clients… and it's fun. 

Ironically, if past history is any indicator, I seem to get the most done on the core (and modules) when I'm actively developing a site. Needs just pop up a lot more. I don't know if that'll be the case this time or not, but I do expect to have weeks with lots of core updates and some weeks with no core updates, just depending on where we are in the project. This particular project has to launch phase 1 by sometime in July, which is kind of a tight schedule, and that may slow core updates temporarily, but who knows. I'll share more on this project and what we learn in this WP-to-PW conversion in the coming weeks. Thanks for reading and have a great weekend! 
 

[nbcommunication]

31 minutes ago, ryan said:

I always enjoy taking a hodgepodge and turning it into an efficient, performant and secure ProcessWire site.

Me too ?

[MarkE]

2 hours ago, ryan said:

I'll share more on this project and what we learn in this WP-to-PW conversion in the coming weeks.

I look forward to that. I have one to do coming up, although much simpler than your’s sounds. Previously, I’ve just started from scratch, using the WP site as the spec, but maybe there are better ways. 

[ryan]

Quote

 

Previously, I’ve just started from scratch, using the WP site as the spec, but maybe there are better ways. 

 

I always start from scratch when it comes to creating the fields and templates, and basic site structure, but rarely when it comes to the pages/content. On previous conversions I've gone into WP’s database and exports to migrate content. This new case is a little bit unique in that WordPress couldn't do everything they needed, at least not in a manner that was affordable or reasonable to implement, from what I understand.

While some of the content is in WP, other content is coming from authenticated web services. Lots of the pages have some content, custom fields and repeatable blocks in WP, but there are also custom fields with IDs for external services that it pulls other content from at runtime. They are paying a couple other companies quite a bit of money every month to host their content through these web services and make it searchable. They pull from those services at runtime and output it all from the WP templates. It strikes me as really inefficient, but it all works fine and you'd never know it from the front-end. But it's also expensive for them and fragile to have these kinds of external dependencies, not to mention a pain to have to manage content (even for the same page) across different services.

This is all stuff that ProcessWire does well on its own, so none of those external services are needed on the new site. Since the front-end of the site is already compiling them all into single-page output,  I built a scraper to pull it directly from the existing site and put it into the right places in ProcessWire. Scraping might seem like a last resort, but in this case it was the fastest, most direct way to pull everything out since the content is split among different databases and services. Basically using WP as an adaptor for the services it is pulling from. Luckily I can edit the WP templates to modify the markup (adding wrapping tags or custom html id attributes, etc.) where needed make the scraper relatively simple. 

@MarkE

[Arklogic]

On 5/22/2022 at 2:57 PM, ryan said:

I built a scraper to pull it directly from the existing site and put it into the right places in ProcessWire.

Ryan, this tool alone would be very nice to share with the community, so any kind of site could be migrated to ProcessWire more efficiently which would increase the number of community members of ProcessWire.