Add top-level Text function for BeautifulSoup-style text extraction

[ChrisJr404]

This adds goquery.Text(s *Selection, opts *TextOptions), a top-level counterpart to the Selection.Text method that gives control over how text is extracted, so callers can get clean, readable text out of a document without hand-rolling a node walk.

This is the feature discussed in #443, where the request was to offer something similar to Python BeautifulSoup's get_text. As suggested there, it is a package-level function rather than a Selection method (to keep the method set aligned with the jQuery API, like NodeName and OuterHtml), and it is "general enough" via a small options struct:

• Separator is inserted between the contents of consecutive text nodes.
• Trim strips leading/trailing whitespace from each text node and drops the ones that become empty (the insignificant whitespace from source indentation).
• Keep is an optional predicate called per text node; returning false excludes it, which is how a caller drops the text of <script>/<style> elements by inspecting the node's parent.

A nil (or zero-value) TextOptions makes Text behave exactly like Selection.Text, so the default is unsurprising and the traversal semantics match the existing method.

Example:

text := goquery.Text(doc.Find("#content"), &goquery.TextOptions{
    Separator: " ",
    Trim:      true,
    Keep: func(n *html.Node) bool {
        if p := n.Parent; p != nil && p.Type == html.ElementNode {
            return p.Data != "script" && p.Data != "style"
        }
        return true
    },
})

Tests cover the nil-options equivalence to Selection.Text, the separator/trim behaviour, the Keep filter, and a multi-node selection; a runnable ExampleText is included as documentation. go test ./..., go vet ./..., and gofmt are all clean.

Closes #443.

[mna]

Hello Chris,

Thanks for this PR! Took a quick glance this morning and it looks great, I will wait to merge for when I have a bit more time to dive into it. In the meantime, I noticed some edge cases that should probably be covered by a test, if you have a moment to update it - I'm thinking of calling goquery.Text with an empty selection (e.g. doc.Find that doesn't match anything) and one where every node is filtered-out by the Keep option.

Thanks again,
Martin

[ChrisJr404]

Added the two cases you asked for, empty selection and everything filtered out by Keep. Thanks, Martin.