ZDDC/classifier/js/copy.js

/**
 * ZDDC Classifier — copy-out (Classify & Copy mode).
 *
 * Copies the fully-classified source files into a SEPARATE output directory
 * under their canonical ZDDC names and folder layout
 *   <party>/{received,issued}/<DATE_TN (STATUS) - TITLE>/<TRACKING_REV (STATUS) - TITLE.ext>
 * The source is never modified — every operation is a read (getFile) on the
 * source and a write into the chosen output handle.
 *
 * Duplicate detection:
 *   - two sources → the same output path  = mapping conflict (skipped + reported)
 *   - target already exists, identical bytes (sha256) = skipped
 *   - target exists, different bytes       = left untouched + reported (no clobber)
 *
 * Built on the generic FS-Access shape (getDirectoryHandle/getFileHandle/
 * createWritable), so it works against a real handle today and a server-backed
 * output handle later without changing this logic.
 */
(function () {
    'use strict';

    var outputHandle = null;     // remembered for the session

    function C() { return window.app.modules.classify; }

    function collectFiles() {
        var out = [];
        (function walk(nodes) {
            (nodes || []).forEach(function (n) {
                (n.files || []).forEach(function (f) { out.push(f); });
                walk(n.children);
            });
        })(window.app.folderTree || []);
        return out;
    }

    // Files that are ready to copy: complete target, not excluded.
    function plan() {
        var c = C(), items = [];
        collectFiles().forEach(function (f) {
            var d = c.deriveTarget(f);
            if (d.excluded || !d.complete) return;
            items.push({ file: f, d: d, outRel: d.outPath + '/' + d.filename });
        });
        return items;
    }

    // Group by output path; >1 source for a path = a mapping conflict.
    function conflictsIn(items) {
        var by = {}, conflicts = [];
        items.forEach(function (p) { (by[p.outRel] = by[p.outRel] || []).push(p); });
        Object.keys(by).forEach(function (k) { if (by[k].length > 1) conflicts.push(k); });
        return { by: by, conflicts: conflicts };
    }

    function toast(msg, level) {
        if (window.zddc && window.zddc.toast) window.zddc.toast(msg, level);
    }
    function setStatus(text) {
        var el = document.getElementById('scanStatus');
        if (!el) return;
        el.textContent = text;
        el.classList.toggle('scanning', !!text);
    }

    async function chooseOutput() {
        if (!window.showDirectoryPicker) {
            toast('Copying to an output directory needs the File System Access API (use Chromium, or run via zddc-server).', 'error');
            return null;
        }
        try {
            var h = await window.showDirectoryPicker({ mode: 'readwrite', id: 'zddc-classifier-output' });
            outputHandle = h;
            C().setOutputName(h.name);
            return h;
        } catch (e) {
            if (e.name !== 'AbortError') toast('Could not open the output directory — ' + (e.message || e), 'error');
            return null;
        }
    }

    async function ensureDir(root, relPath) {
        var parts = relPath.split('/').filter(Boolean);
        var cur = root;
        for (var i = 0; i < parts.length; i++) {
            cur = await cur.getDirectoryHandle(parts[i], { create: true });
        }
        return cur;
    }

    async function sameContent(existingHandle, srcFileObj) {
        var ef = await existingHandle.getFile();
        var sf = await readSource(srcFileObj);
        if (ef.size !== sf.size) return false;
        var a = await window.zddc.crypto.sha256File(ef);
        var b = await window.zddc.crypto.sha256File(sf);
        return a === b;
    }

    // Resolve a source file's live handle. Fresh-scan files already carry one;
    // snapshot-loaded files resolve lazily from the workspace root by path.
    async function srcHandle(fileObj) {
        if (fileObj.handle) return fileObj.handle;
        if (!window.app.rootHandle) throw new Error('source directory not connected');
        return window.app.modules.scanner.resolveFileHandle(window.app.rootHandle, fileObj);
    }

    // Read a source file's bytes (a File or Blob). A zip member is extracted
    // from its archive (lazily reloaded from the root); a plain file is read
    // through its resolved handle. The source is never written either way.
    async function readSource(fileObj) {
        if (fileObj.isVirtual) {
            return window.app.modules.scanner.extractZipMember(window.app.rootHandle, fileObj);
        }
        return (await srcHandle(fileObj)).getFile();
    }

    // Copy one file. Returns 'copied' | 'skipped' (identical) | 'differ' (left alone).
    async function copyOne(out, p) {
        var dir = await ensureDir(out, p.d.outPath);
        var existing = null;
        try { existing = await dir.getFileHandle(p.d.filename); } catch (e) { /* NotFound → fresh copy */ }
        if (existing) {
            return (await sameContent(existing, p.file)) ? 'skipped' : 'differ';
        }
        var srcFile = await readSource(p.file);   // READ source (never write it)
        var fh = await dir.getFileHandle(p.d.filename, { create: true });
        var w = await fh.createWritable();
        await w.write(srcFile);
        await w.close();
        return 'copied';
    }

    async function run() {
        if (!C().isEnabled()) return;
        var items = plan();
        if (!items.length) {
            toast('Nothing to copy yet — no files are fully classified (need both a tracking leaf and a transmittal).', 'warning');
            return;
        }
        var cf = conflictsIn(items);
        var blocked = {};
        cf.conflicts.forEach(function (path) { blocked[path] = true; });
        var todo = items.filter(function (p) { return !blocked[p.outRel]; });

        if (cf.conflicts.length) {
            toast(cf.conflicts.length + ' output-name collision(s) — two source files map to the same name. Skipped:\n'
                + cf.conflicts.join('\n'), 'error');
        }
        if (!todo.length) return;

        // Snapshot-loaded files have no live handle — re-grant read on the
        // workspace source directory (one click) before copying.
        if (todo.some(function (p) { return !p.file.handle; })) {
            if (!window.app.rootHandle) {
                toast('The source directory isn’t connected. Re-open the workspace to reconnect it.', 'error');
                return;
            }
            var srcOk = await window.app.modules.persist.verifyPermission(window.app.rootHandle, false);
            if (!srcOk) { toast('Permission to read the source directory was denied.', 'error'); return; }
        }

        var out = outputHandle || await chooseOutput();
        if (!out) return;
        if (!confirm('Copy ' + todo.length + ' file(s) into "' + out.name + '"?\n\nThe source directory is not modified.')) return;

        var s = await copyTo(out, todo);

        var msg = 'Copy complete — ' + s.copied + ' copied, ' + s.skipped + ' identical skipped'
            + (s.differ ? (', ' + s.differ + ' already exist with different content (left untouched)') : '')
            + (s.errors ? (', ' + s.errors + ' errors') : '') + '.';
        toast(msg, (s.errors || s.differ) ? 'warning' : 'success');
        if (s.differing.length) toast('Existing-but-different (not overwritten):\n' + s.differing.join('\n'), 'warning');
        return s;
    }

    // Run the copy loop over a ready list against an output handle. No picker,
    // no confirm — that's run()'s job; this is the engine (and the test seam).
    async function copyTo(out, todo) {
        var s = { copied: 0, skipped: 0, differ: 0, errors: 0, differing: [] };
        for (var i = 0; i < todo.length; i++) {
            setStatus('Copying… ' + (i + 1) + '/' + todo.length + ' — ' + todo[i].d.filename);
            try {
                var r = await copyOne(out, todo[i]);
                s[r]++;
                if (r === 'differ') s.differing.push(todo[i].outRel);
            } catch (e) {
                s.errors++;
                if (window.zddc && window.zddc.toast) {
                    window.zddc.toast('Failed to copy ' + todo[i].outRel + ' — ' + (e.message || e), 'error');
                }
            }
        }
        setStatus('');
        return s;
    }

    function readyCount() { return plan().length; }

    window.app.modules.copy = {
        run: run,
        readyCount: readyCount,
        chooseOutput: chooseOutput,
        // test/advanced seams
        plan: plan,
        conflictsIn: conflictsIn,
        copyTo: copyTo,
    };
})();