**tl;dr** - how to search any sub-node for a particular string in order to extract the URL contained there?

---

I have a script that parses the [OPDS feed][o] for [Standard Ebooks][se], found at https://standardebooks.org/opds/all. I can get everything I basically need, and it produces [a simple cumulative index][ci] — fairly rough-and-ready, but that's what I'm after. So far, so good.

Sometimes, books appear in a series, however, and this is invariably noted and linked in the (long) "content" description of the ebook. **I would like to be able to extract the URL for the series, where such occurs.** The first example in the current listing is Agatha Christie's *The Murder on the Links*. The first paragraph of the long description has this:

> *The Murder on the Links* is [Agatha Christie’s](https://standardebooks.org/ebooks/agatha-christie) second [Poirot](https://standardebooks.org/collections/poirot) novel, featuring the brilliant Belgian detective and his sidekick, Captain Hastings.

The raw HTML:

```html
<i>The Murder on the Links</i> is <a href="https://standardebooks.org/ebooks/agatha-christie">Agatha Christie’s</a> second <a href="https://standardebooks.org/collections/poirot">Poirot</a> novel, featuring the brilliant Belgian detective and his sidekick, Captain Hastings.</p>
	
```

In the OPDS feed:

```xml
<content type="text/html">
  <p>
  <i>The Murder on the Links</i>
  is
  <a href="https://standardebooks.org/ebooks/agatha-christie">Agatha Christie’s</a>
  second
  <a href="https://standardebooks.org/collections/poirot">Poirot</a>
  novel, featuring the brilliant Belgian detective and his sidekick, Captain Hastings.
  </p>
  ...
</content>
```

And when parsed by PHP's SimpleXML gives this (extract only!):

```php
SimpleXMLElement Object
(
    [content] => SimpleXMLElement Object
        (
            [@attributes] => Array
                (
                    [type] => text/html
                )
            [p] => Array
                (
                    [0] => SimpleXMLElement Object
                        (
                            [i] => The Murder on the Links
                            [a] => Array
                                (
                                    [0] => SimpleXMLElement Object
                                        (
                                            [@attributes] => Array
                                                (
                                                    [href] => https://standardebooks.org/ebooks/agatha-christie
                                                )
                                            [0] => Agatha Christie’s
                                        )
                                    [1] => SimpleXMLElement Object
                                        (
                                            [@attributes] => Array
                                                (
                                                    [href] => https://standardebooks.org/collections/poirot
                                                )
                                            [0] => Poirot
                                        )
                                )
                        )
                    [1] => In this characteristic whodunit, Poirot is summoned to a seaside town in northern France by a desperate letter from a rich businessman, who fears that he is being stalked. Poirot arrives to find the businessman already dead, his body lying facedown in an open grave on a golf course, a knife in his back—the victim of a mysterious murder. Over the coming days Poirot clashes wits with an arrogant Parisian detective, Giraud, while Hastings finds himself pining after a beautiful but shadowy American expatriate known to him only as “Cinderella.” Together, Poirot and Hastings unravel the intricate web of mystery and deceit behind the murder.
                    [2] => Christie based this mystery after a real-life French murder case, and it’s believed that this is the first detective novel to use the phrase “the scene of the crime.”
                )
        )
)
```

I can pull out the URL for the "collection" when I know where it is, with something like this:  
`$content = $list[$i]->content->p[0]->a[1][href];` — where `$list[$i]` identifies the particular ebook "node" in a `foreach` loop: it produces the "SimpleXMLElement Object" extract, above. (Often in PHP docs it would be represented by `$xml`.)

But what I need to be able to do is search anywhere in the `content` node for a paragraph that has the text `standardebooks.org/collections` (or whatever is the most sensible string for this search), to then get the full URL for the collection.

I have tried things with `xpath()`, using `//` to search any sub-node, but I have not been able to work out the correct syntax for the search.

Any help with this would be greatly appreciated! (I can provide the complete "parsed" XML for the feed, if that would be a help.)

[o]: https://specs.opds.io/opds-1.2.html
[se]: https://standardebooks.org/
[ci]: https://www.sudalyph.org/seci/