@preventive/compare 0.0.0-stage → 1.0.1

package/src/words.js

@@ -0,0 +1,57 @@
+// Changed lines read word by word: what changed within a line and the one it
+// pairs with, as one shortest edit script between their words finds it.
+
+import { script } from './script.js'
+
+// Words as a word diff compares them, so `foo(a)` -> `foo(b)` marks `a` alone.
+const TOKEN = /[\p{L}\p{N}_$]+|\s+|[^\p{L}\p{N}_$\s]/gu
+// Past these a word diff is noise (minified lines) or too slow to be worth it.
+const MAX_CHARS = 2000
+const MAX_TOKENS = 800
+// Lines whose words are mostly replaced are not one line edited.
+export const MAX_CHANGED = 0.6
+
+// A line as a word diff reads it, each distinct word numbered in `ids` as the
+// script compares them; null past the size limits.
+export function tokenized(line, ids) {
+  if (line.length > MAX_CHARS) return null
+  const tokens = line.match(TOKEN) ?? []
+  if (tokens.length > MAX_TOKENS) return null
+  return { line, tokens, words: tokens.map((token) => {
+    if (!ids.has(token)) ids.set(token, ids.size)
+    return ids.get(token)
+  }) }
+}
+
+// A line's words the script changed as ranges `[from, to)`, those split by blanks alone merged.
+function changedRanges({ tokens, line }, kept) {
+  const ranges = []
+  for (let at = 0, i = 0; i < tokens.length; at += tokens[i++].length) {
+    if (kept[i]) continue
+    const last = ranges.at(-1)
+    if (last && /^\s*$/u.test(line.slice(last[1], at))) last[1] = at + tokens[i].length
+    else ranges.push([at, at + tokens[i].length])
+  }
+  return ranges
+}
+
+const part = (ranges, line) => ranges.reduce((sum, [from, to]) => sum + to - from, 0) / Math.max(1, line.length)
+
+// What a word view marks in two lines read: the ranges, the smaller share of
+// either they cover (all of it where the script keeps no word), and the work.
+export function wordView(a, b) {
+  const kept = { a: a.words.map(() => 0), b: b.words.map(() => 0) }
+  const steps = script(a.words, b.words, kept)
+  const ranges = { a: changedRanges(a, kept.a), b: changedRanges(b, kept.b) }
+  return { ranges, share: kept.a.includes(1) ? Math.min(part(ranges.a, a.line), part(ranges.b, b.line)) : 1, steps }
+}
+
+// Lines the same have no ranges, not null: null is a line with no word view.
+export function wordRanges(left, right) {
+  if (left === right) return { a: [], b: [] }
+  const ids = new Map()
+  const a = tokenized(left, ids), b = tokenized(right, ids)
+  if (a === null || b === null) return null
+  const { ranges, share } = wordView(a, b)
+  return share <= MAX_CHANGED ? ranges : null
+}

package/README.md

@@ -1,3 +0,0 @@
-# Temporary Holding Version
-
-This version is a temporary placeholder for this package. An operational version to replace this has been submitted for review and is awaiting a staged release.