How to get taxonomy terms for the current node from a block

Is there a solution for situations where you don't have URLs like "node/10". Speciflcally I use pathauto for all my nodes - I'm trying to think of a way to adapt this for the situation where the url is something like

"category/title-of-node"

Any ideas?
Greg

--
Growing Venture Solutions
Drupal Implementation and Support in Denver, CO

I found a nice way too, and that is to implement a hook_nodeapi function, intercepting a 'load' with a node of matching 'type' (or other condition), and then sets a global variable describing the node.

The block can simply check if the global variable is set, and can get whatever data it likes from the node through the variable.

Advantage here is that you don't need to read the url - which is nice when the url path is mapped by another module.

The only caveat I see is that it relies on drupal loading the node before building the blocks. This is how it works in Drupal 6.x.

Hi all,

I've been working through the same issue myself today, but I'm fairly new to drupal so I don't know if my approach is considered "kosher" or not.

From what I can tell, the approach shown above (checking arg(0) = node and arg(1) contains a numeric ID) is very much tied to the path /node/X, or any alias that refers to it.

On the site I'm working on, I built myself a custom "catalog" module and registered a couple of menu callbacks - one of which is /resources/item/%node. That path displays nodes of a particular type and does a bit of custom tweaking of the breadcrumb (to include parent taxonomy terms etc). The site still supports the /node/X path because it seems unfeasible to remove it - particularly since various modules, like the developer module rely on it. Also editing nodes always sends a user back to /node/X/edit when the save or preview buttons are clicked.

All of this means I need an approach (in my "node terms" block) that isn't tied to a particular path structure. So, I came up with this (for Drupal 6 - not sure if it will work in Drupal 5):

	// STEP 1: Get current menu item (includes path arguments converted to 
	//         objects, as loaded by xxx_load functions);
		$menuitem = menu_get_item();	
		if ($menuitem && is_array($menuitem['map'])) {
			// Walk over each element of the path (e.g. /node/%node, where 
			// wildcard path items will have been replaced with their fully 
			// loaded objects). Stop at the first NODE we find.
			foreach ($menuitem['map'] as $item) {
				if (isset($item->nid)) {
					// We have found a node
					$node = $item;
					break;
				}
			}
		}

When the current path is of the form /somename/someothername/%node, some drupal menurouter magic invokes node_load on %nodeid and stores the returned node object in the map element of the menu object. My code just walks over each part of the path looking for anything that is a node.

It seems to work nicely - and on my site it works automatically with both /resource/item/X and /node/X paths. I would assume this also saves unnecessary processing since node_load doesn't need to be called again, and the node object has already been "processed" by various modules such as CCK, which add all kinds of useful properties.

So my question is this: is there any reason I shouldn't be doing this? Is this infact a better approach than the if (arg(0) == 'node' && is_numeric(arg(1)) approach I keep seeing in other articles discussing how to establish the "current node id"?

Any thoughts?