Deterministic swehockey.se schedule crawler

One HTTP GET per season, no browser: the page is server-rendered, so every row
arrives in one response and the extraction becomes a pure function of the HTML.

- layouts[] declares one profile per version of the site's table (modern and
  legacy), each identified by its header row, so a new version of the page is a
  new profile plus a fixture instead of a guess.
- The table is located by header signature, not by a CSS path.
- Every cell is shape-checked, and facts the site renders twice (desktop and
  mobile date, time in the tooltip) must agree.
- Cross-field invariants: periods sum to the score, ids are unique, played games
  always have an event link and unplayed ones never do.
- Output is sorted by date, time and id, so it does not inherit the site's row
  order, and it carries no fetch timestamp: same HTML, same bytes.
- Drift is a typed error and exit code 3, with the header it actually found.
- Tests run on committed HTML fixtures and golden JSON; TestLive is opt-in.

Verified over the full SHL archive: 27/27 seasons, 2119 games, 1779 played.

dom.go

@@ -0,0 +1,129 @@
+package main
+
+import (
+	"crypto/sha256"
+	"encoding/hex"
+	"strings"
+	"unicode"
+
+	"golang.org/x/net/html"
+)
+
+// The site writes the team separator as NBSP-hyphen-NBSP wrapped in newlines,
+// so every string that leaves the parser goes through normalize and downstream
+// comparisons and diffs stay stable.
+func normalize(s string) string {
+	return strings.Join(strings.FieldsFunc(s, isSpaceRune), " ")
+}
+
+func isSpaceRune(r rune) bool {
+	if unicode.IsSpace(r) {
+		return true
+	}
+	// No-break and zero-width variants the site emits around team names.
+	switch r {
+	case '\u00a0', '\u2007', '\u202f', '\u200b', '\ufeff':
+		return true
+	}
+	return false
+}
+
+// textOf returns the normalized text of a node subtree.
+func textOf(n *html.Node) string {
+	if n == nil {
+		return ""
+	}
+	var b strings.Builder
+	var walk func(*html.Node)
+	walk = func(n *html.Node) {
+		if n.Type == html.TextNode {
+			b.WriteString(n.Data)
+		}
+		for c := n.FirstChild; c != nil; c = c.NextSibling {
+			walk(c)
+		}
+	}
+	walk(n)
+	return normalize(b.String())
+}
+
+// textExcluding returns the normalized text of a node subtree with the first
+// descendant matching skipTag (and its subtree) left out. The schedule uses
+// this to read the mobile cell's start time without its <bold> date.
+func textExcluding(n *html.Node, skipTag string) string {
+	if n == nil {
+		return ""
+	}
+	var b strings.Builder
+	var walk func(*html.Node) bool
+	walk = func(n *html.Node) bool {
+		if n.Type == html.ElementNode && n.Data == skipTag {
+			return false
+		}
+		if n.Type == html.TextNode {
+			b.WriteString(n.Data)
+		}
+		for c := n.FirstChild; c != nil; c = c.NextSibling {
+			walk(c)
+		}
+		return true
+	}
+	walk(n)
+	return normalize(b.String())
+}
+
+// childElements returns the direct children of n with the given tag name.
+func childElements(n *html.Node, tag string) []*html.Node {
+	if n == nil {
+		return nil
+	}
+	var out []*html.Node
+	for c := n.FirstChild; c != nil; c = c.NextSibling {
+		if c.Type == html.ElementNode && c.Data == tag {
+			out = append(out, c)
+		}
+	}
+	return out
+}
+
+// descendants returns every element with the given tag name, in document order.
+func descendants(n *html.Node, tag string) []*html.Node {
+	var out []*html.Node
+	var walk func(*html.Node)
+	walk = func(n *html.Node) {
+		if n.Type == html.ElementNode && n.Data == tag {
+			out = append(out, n)
+		}
+		for c := n.FirstChild; c != nil; c = c.NextSibling {
+			walk(c)
+		}
+	}
+	walk(n)
+	return out
+}
+
+func attr(n *html.Node, key string) string {
+	if n == nil {
+		return ""
+	}
+	for _, a := range n.Attr {
+		if a.Key == key {
+			return a.Val
+		}
+	}
+	return ""
+}
+
+func hasClass(n *html.Node, class string) bool {
+	for _, f := range strings.Fields(attr(n, "class")) {
+		if f == class {
+			return true
+		}
+	}
+	return false
+}
+
+func sha256Hex(b []byte) string {
+	sum := sha256.Sum256(b)
+	return hex.EncodeToString(sum[:])
+}