printhive/lib/fingerprint.js

// Background worker that fingerprints stored files for the duplicate finder: a
// SHA-256 of the bytes for every file, plus a geometry fingerprint for mesh
// formats so the same model still matches after being renamed, re-exported
// (ASCII/binary STL) or converted (STL -> 3MF).

const fs = require('fs');
const path = require('path');
const crypto = require('crypto');
const { execFile } = require('child_process');
const db = require('./db');

const PYTHON = process.env.PYTHON_BIN || path.join(__dirname, '..', 'tools', 'venv', 'bin', 'python');
const SCRIPT = path.join(__dirname, '..', 'tools', 'fingerprint.py');
const GEOMETRY_FORMATS = ['stl', 'obj', '3mf'];
const BATCH = 10;

let pythonAvailable = true;
let running = false;
let rerun = false;

function sha256File(filePath) {
  return new Promise((resolve, reject) => {
    const hash = crypto.createHash('sha256');
    fs.createReadStream(filePath)
      .on('error', reject)
      .on('data', (chunk) => hash.update(chunk))
      .on('end', () => resolve(hash.digest('hex')));
  });
}

// n rounded to `digits` significant figures, as a stable string.
function sig(n, digits = 4) {
  if (!Number.isFinite(n) || n === 0) return '0';
  return String(Number(n.toPrecision(digits)));
}

// Hash of rounded shape measurements. Position and axis-aligned rotation don't
// change it, and rounding absorbs float noise between export formats.
function geometryKey(fp) {
  if (!fp || !fp.faces) return null;
  const extents = (fp.extents || []).map((e) => Math.abs(e)).sort((a, b) => a - b).map((e) => e.toFixed(1));
  const raw = `${fp.faces}|${sig(fp.area)}|${sig(Math.abs(fp.volume || 0))}|${extents.join('x')}`;
  return crypto.createHash('sha1').update(raw).digest('hex');
}

// Runs the trimesh sidecar over some files. Resolves Map(absPath -> {fp}|{error});
// output from a run that timed out or crashed part-way is still returned.
function measure(paths) {
  return new Promise((resolve, reject) => {
    execFile(
      PYTHON,
      [SCRIPT, ...paths],
      { maxBuffer: 4 * 1024 * 1024, timeout: 60000 * (paths.length + 1) },
      (err, stdout) => {
        if (err && err.code === 'ENOENT') return reject(err);
        const results = new Map();
        for (const line of String(stdout || '').split('\n')) {
          if (!line.trim()) continue;
          try {
            const r = JSON.parse(line);
            results.set(r.path, r);
          } catch { /* ignore non-JSON noise */ }
        }
        resolve(results);
      }
    );
  });
}

async function fingerprintGeometry(rows) {
  const abs = (row) => path.resolve(row.file_path);
  const results = await measure(rows.map(abs));
  for (const row of rows) {
    let r = results.get(abs(row));
    // A batch that died part-way: retry the missing file on its own.
    if (!r && rows.length > 1) r = (await measure([abs(row)])).get(abs(row));
    const key = r && r.fp ? geometryKey(r.fp) : null;
    await db.query('UPDATE files SET geometry_hash = $2, geometry_status = $3 WHERE id = $1',
      [row.id, key, key ? 'ok' : 'error']);
  }
}

async function run() {
  for (;;) {
    const { rows } = await db.query(
      `SELECT id, file_path, format, fingerprinted_at FROM files
       WHERE fingerprinted_at IS NULL OR (geometry_status IS NULL AND $1::boolean)
       ORDER BY id LIMIT $2`, [pythonAvailable, BATCH]);
    if (rows.length === 0) return;

    const geometry = [];
    for (const row of rows) {
      const abs = path.resolve(row.file_path);
      try {
        if (!fs.existsSync(abs)) {
          await db.query(`UPDATE files SET fingerprinted_at = now(), geometry_status = 'missing' WHERE id = $1`, [row.id]);
          continue;
        }
        if (!row.fingerprinted_at) {
          await db.query('UPDATE files SET sha256 = $2, fingerprinted_at = now() WHERE id = $1',
            [row.id, await sha256File(abs)]);
        }
        if (!GEOMETRY_FORMATS.includes(row.format)) {
          await db.query(`UPDATE files SET geometry_status = 'na' WHERE id = $1`, [row.id]);
        } else if (pythonAvailable) {
          geometry.push(row);
        }
      } catch (err) {
        console.warn(`Fingerprint failed for file ${row.id}:`, err.message);
        await db.query(`UPDATE files SET fingerprinted_at = now(), geometry_status = 'error' WHERE id = $1`, [row.id]);
      }
    }

    if (geometry.length) {
      try {
        await fingerprintGeometry(geometry);
      } catch (err) {
        if (err.code !== 'ENOENT') throw err;
        pythonAvailable = false;
        console.warn(`Geometry fingerprints disabled: no Python at ${PYTHON} (content hashes still run).`);
      }
    }
  }
}

// Start (or queue another pass of) the worker. Safe to call often.
function kick() {
  if (running) { rerun = true; return; }
  running = true;
  run()
    .catch((err) => console.error('Fingerprinting stopped:', err.message))
    .finally(() => {
      running = false;
      if (rerun) { rerun = false; kick(); }
    });
}

async function pendingCount() {
  const { rows } = await db.query(
    `SELECT COUNT(*)::int AS n FROM files
     WHERE fingerprinted_at IS NULL OR (geometry_status IS NULL AND $1::boolean)`, [pythonAvailable]);
  return rows[0].n;
}

module.exports = { kick, pendingCount, geometryKey, sha256File };