work on OCR

tools/OcrDaemon/tessdata/generate-words.mjs

@@ -0,0 +1,166 @@
+#!/usr/bin/env node
+/**
+ * Fetches POE2 trade API data and generates Tesseract user-words and user-patterns
+ * files to improve OCR accuracy for tooltip text.
+ *
+ * Usage: node generate-words.mjs
+ * Output: poe2.user-words, poe2.user-patterns (in same directory)
+ */
+
+import { writeFileSync } from "fs";
+import { dirname, join } from "path";
+import { fileURLToPath } from "url";
+
+const __dirname = dirname(fileURLToPath(import.meta.url));
+const UA = "OAuth poe2trade/1.0 (contact: poe2trade@users.noreply.github.com)";
+
+async function fetchJson(path) {
+  const url = `https://www.pathofexile.com/api/trade2/data/${path}`;
+  const res = await fetch(url, { headers: { "User-Agent": UA } });
+  if (!res.ok) throw new Error(`${url}: ${res.status}`);
+  return res.json();
+}
+
+async function main() {
+  console.log("Fetching POE2 trade API data...");
+  const [items, stats, static_, filters] = await Promise.all([
+    fetchJson("items"),
+    fetchJson("stats"),
+    fetchJson("static"),
+    fetchJson("filters"),
+  ]);
+
+  const words = new Set();
+
+  // Helper: split text into individual words and add each
+  function addWords(text) {
+    if (!text) return;
+    // Remove # placeholders and special chars, split on whitespace
+    const cleaned = text
+      .replace(/#/g, "")
+      .replace(/[{}()\[\]]/g, "")
+      .replace(/[+\-]/g, " ");
+    for (const word of cleaned.split(/\s+/)) {
+      // Only keep words that are actual words (not numbers, not single chars)
+      const trimmed = word.replace(/^[^a-zA-Z]+|[^a-zA-Z]+$/g, "");
+      if (trimmed.length >= 2) words.add(trimmed);
+    }
+  }
+
+  // Helper: add a full phrase (multi-word item name) as-is
+  function addPhrase(text) {
+    if (!text) return;
+    addWords(text);
+  }
+
+  // Items: type names (base types like "Tribal Mask", "Leather Vest")
+  for (const cat of items.result) {
+    addPhrase(cat.label);
+    for (const entry of cat.entries) {
+      addPhrase(entry.type);
+      addPhrase(entry.name);
+      addPhrase(entry.text);
+    }
+  }
+
+  // Stats: mod text like "+#% to Chaos Resistance", "# to maximum Life"
+  for (const cat of stats.result) {
+    for (const entry of cat.entries) {
+      addPhrase(entry.text);
+    }
+  }
+
+  // Static: currency/fragment names like "Divine Orb", "Scroll of Wisdom"
+  for (const cat of static_.result) {
+    addPhrase(cat.label);
+    for (const entry of cat.entries) {
+      addPhrase(entry.text);
+    }
+  }
+
+  // Filters: filter labels and option texts
+  for (const cat of filters.result) {
+    addPhrase(cat.title);
+    if (cat.filters) {
+      for (const f of cat.filters) {
+        addPhrase(f.text);
+        if (f.option?.options) {
+          for (const opt of f.option.options) {
+            addPhrase(opt.text);
+          }
+        }
+      }
+    }
+  }
+
+  // Add common tooltip keywords not in trade API
+  const extraWords = [
+    // Section headers
+    "Quality", "Requires", "Level", "Asking", "Price",
+    "Corrupted", "Mirrored", "Unmodifiable",
+    "Twice", "Sockets",
+    // Attributes
+    "Strength", "Dexterity", "Intelligence", "Spirit",
+    // Defense types
+    "Armour", "Evasion", "Rating", "Energy", "Shield",
+    // Damage types
+    "Physical", "Elemental", "Lightning", "Cold", "Fire", "Chaos",
+    // Common mod words
+    "increased", "reduced", "more", "less",
+    "added", "converted", "regeneration",
+    "maximum", "minimum", "total",
+    "Resistance", "Damage", "Speed", "Duration",
+    "Critical", "Hit", "Chance", "Multiplier",
+    "Attack", "Cast", "Spell", "Minion", "Skill",
+    "Mana", "Life", "Rarity",
+    // Item classes
+    "Helmet", "Gloves", "Boots", "Body", "Belt",
+    "Ring", "Amulet", "Shield", "Quiver",
+    "Sword", "Axe", "Mace", "Dagger", "Wand", "Staff", "Bow",
+    "Sceptre", "Crossbow", "Flail", "Spear",
+    // Rarity
+    "Normal", "Magic", "Rare", "Unique",
+  ];
+  for (const w of extraWords) words.add(w);
+
+  // Sort and write user-words
+  const sortedWords = [...words].sort((a, b) => a.toLowerCase().localeCompare(b.toLowerCase()));
+  const wordsPath = join(__dirname, "poe2.user-words");
+  writeFileSync(wordsPath, sortedWords.join("\n") + "\n");
+  console.log(`Wrote ${sortedWords.length} words to ${wordsPath}`);
+
+  // Generate user-patterns for common tooltip formats
+  const patterns = [
+    // Stat values: "+12% to Chaos Resistance", "+3 to Level"
+    "\\+\\d+%",
+    "\\+\\d+",
+    "\\-\\d+%",
+    "\\-\\d+",
+    // Ranges: "10-20"
+    "\\d+-\\d+",
+    // Currency amounts: "7x Divine Orb", "35x Divine Orb"
+    "\\d+x",
+    // Quality: "+20%"
+    "\\d+%",
+    // Level requirements: "Level \\d+"
+    "Level \\d+",
+    // Asking Price section
+    "Asking Price:",
+    // Item level
+    "Item Level: \\d+",
+    // Requires line
+    "Requires:",
+    // Rating values
+    "Rating: \\d+",
+    "Shield: \\d+",
+    "Quality: \\+\\d+%",
+  ];
+  const patternsPath = join(__dirname, "poe2.user-patterns");
+  writeFileSync(patternsPath, patterns.join("\n") + "\n");
+  console.log(`Wrote ${patterns.length} patterns to ${patternsPath}`);
+}
+
+main().catch((e) => {
+  console.error(e);
+  process.exit(1);
+});