feat: add D-InSAR source bundle distribution

backend/app/copier.py

@@ -3,6 +3,8 @@ import shutil
 import asyncio
 import tempfile
 import zipfile
+import json
+import hashlib
 from datetime import datetime
 from typing import List, Tuple, Optional, Dict, Any
 
@@ -158,6 +160,724 @@ def _zip_task_directory(task_dir: str, zip_path: str) -> None:
                 pass
 
 
+def _directory_has_entries(path: str) -> bool:
+    if not os.path.isdir(path):
+        return False
+    try:
+        with os.scandir(path) as entries:
+            return any(True for _ in entries)
+    except OSError:
+        return False
+
+
+def _is_existing_dinsar_folder_complete(task_dir: str) -> bool:
+    return (
+        _directory_has_entries(os.path.join(task_dir, "master"))
+        and _directory_has_entries(os.path.join(task_dir, "slave"))
+    )
+
+
+def _is_existing_dinsar_zip_complete(zip_path: str) -> bool:
+    try:
+        return os.path.isfile(zip_path) and os.path.getsize(zip_path) > 0
+    except OSError:
+        return False
+
+
+def _build_dinsar_pair_metadata(
+    item: Dict[str, Any],
+    task_name: str,
+    task_alias: str,
+    package_format: str,
+    include_orbit_files: bool,
+    orbit_entries: List[Dict[str, Any]],
+) -> Dict[str, Any]:
+    return {
+        "pair_key": item.get("pair_key"),
+        "task_name": task_name,
+        "task_alias": task_alias,
+        "master_path": item.get("master_path"),
+        "slave_path": item.get("slave_path"),
+        "master_satellite": item.get("master_satellite"),
+        "slave_satellite": item.get("slave_satellite"),
+        "master_imaging_date": item.get("master_imaging_date"),
+        "slave_imaging_date": item.get("slave_imaging_date"),
+        "master_imaging_mode": item.get("master_imaging_mode"),
+        "slave_imaging_mode": item.get("slave_imaging_mode"),
+        "master_polarization": item.get("master_polarization"),
+        "slave_polarization": item.get("slave_polarization"),
+        "time_baseline_days": item.get("time_baseline_days"),
+        "spatial_baseline_meters": item.get("spatial_baseline_meters"),
+        "scene_center_distance_meters": item.get("scene_center_distance_meters"),
+        "package_format": package_format,
+        "include_orbit_files": bool(include_orbit_files),
+        "master_orbit_file_path": item.get("master_orbit_file_path"),
+        "slave_orbit_file_path": item.get("slave_orbit_file_path"),
+        "orbit_files": orbit_entries,
+        "scene_pair_uid": item.get("scene_pair_uid") or item.get("pair_uid"),
+        "pair_uid": item.get("pair_uid") or item.get("scene_pair_uid"),
+        "network_run_id": item.get("network_run_id"),
+        "network_edge_id": item.get("network_edge_id"),
+        "policy_version": item.get("policy_version"),
+        "selection_strategy": item.get("selection_strategy"),
+        "copied_at": datetime.utcnow().isoformat(timespec="seconds") + "Z",
+    }
+
+
+def _safe_bundle_entry_name(source_path: str, prefix: str) -> str:
+    base_name = os.path.basename(os.path.normpath(str(source_path or ""))) or "item"
+    digest = hashlib.sha1(os.path.normcase(os.path.abspath(source_path)).encode("utf-8")).hexdigest()[:10]
+    return f"{prefix}_{digest}_{base_name}"
+
+
+def _bundle_relative_scene_path(source_path: str) -> str:
+    normalized = os.path.normpath(os.path.abspath(str(source_path)))
+    return os.path.join("data", _safe_bundle_entry_name(normalized, "scene")).replace(os.sep, "/")
+
+
+def _normalize_bundle_relative_path(path: Any) -> str:
+    return str(path or "").strip().replace("\\", "/")
+
+
+def _source_bundle_id_number(value: Any, prefix: str) -> int:
+    text = str(value or "")
+    marker = f"{prefix}_"
+    if not text.startswith(marker):
+        return 0
+    try:
+        return int(text[len(marker):])
+    except ValueError:
+        return 0
+
+
+def _next_source_bundle_id(records: List[Dict[str, Any]], id_key: str, prefix: str) -> int:
+    max_number = 0
+    for record in records:
+        max_number = max(max_number, _source_bundle_id_number(record.get(id_key), prefix))
+    return max(max_number, len(records)) + 1
+
+
+def _source_bundle_pair_keys_from_item(item: Dict[str, Any]) -> List[str]:
+    keys: List[str] = []
+
+    def add(label: str, value: Any) -> None:
+        text = str(value or "").strip()
+        if text:
+            keys.append(f"{label}:{text}")
+
+    add("uid", item.get("scene_pair_uid") or item.get("pair_uid"))
+    add("pair_key", item.get("pair_key"))
+    network_run_id = str(item.get("network_run_id") or "").strip()
+    network_edge_id = str(item.get("network_edge_id") or "").strip()
+    if network_run_id and network_edge_id:
+        keys.append(f"network:{network_run_id}:{network_edge_id}")
+
+    master_path = item.get("master_path")
+    slave_path = item.get("slave_path")
+    if master_path and slave_path:
+        master_abs = os.path.normcase(os.path.normpath(os.path.abspath(str(master_path))))
+        slave_abs = os.path.normcase(os.path.normpath(os.path.abspath(str(slave_path))))
+        keys.append(f"source_paths:{master_abs}|{slave_abs}")
+        keys.append(
+            "bundle_paths:"
+            f"{_bundle_relative_scene_path(str(master_path))}|"
+            f"{_bundle_relative_scene_path(str(slave_path))}"
+        )
+    return keys
+
+
+def _source_bundle_pair_keys_from_pair(pair: Dict[str, Any]) -> List[str]:
+    keys: List[str] = []
+
+    identity_key = str(pair.get("identity_key") or "").strip()
+    if identity_key:
+        keys.append(identity_key)
+
+    def add(label: str, value: Any) -> None:
+        text = str(value or "").strip()
+        if text:
+            keys.append(f"{label}:{text}")
+
+    add("uid", pair.get("scene_pair_uid") or pair.get("pair_uid"))
+    add("pair_key", pair.get("pair_key"))
+    network_run_id = str(pair.get("network_run_id") or "").strip()
+    network_edge_id = str(pair.get("network_edge_id") or "").strip()
+    if network_run_id and network_edge_id:
+        keys.append(f"network:{network_run_id}:{network_edge_id}")
+
+    master_source_path = pair.get("master_source_path")
+    slave_source_path = pair.get("slave_source_path")
+    if master_source_path and slave_source_path:
+        master_abs = os.path.normcase(os.path.normpath(os.path.abspath(str(master_source_path))))
+        slave_abs = os.path.normcase(os.path.normpath(os.path.abspath(str(slave_source_path))))
+        keys.append(f"source_paths:{master_abs}|{slave_abs}")
+
+    master_data = _normalize_bundle_relative_path(pair.get("master_data"))
+    slave_data = _normalize_bundle_relative_path(pair.get("slave_data"))
+    if master_data and slave_data:
+        keys.append(f"bundle_paths:{master_data}|{slave_data}")
+
+    return keys
+
+
+def _read_json_file(path: str) -> Dict[str, Any]:
+    if not os.path.isfile(path):
+        return {}
+    with open(path, "r", encoding="utf-8-sig") as handle:
+        payload = json.load(handle)
+    if not isinstance(payload, dict):
+        return {}
+    return payload
+
+
+def _copy_source_into_bundle(source_path: str, dest_path: str, skip_existing: bool) -> str:
+    if skip_existing and os.path.exists(dest_path):
+        return "skipped"
+    if os.path.exists(dest_path):
+        if os.path.isdir(dest_path):
+            shutil.rmtree(dest_path)
+        else:
+            os.remove(dest_path)
+    parent_dir = os.path.dirname(dest_path)
+    os.makedirs(parent_dir, exist_ok=True)
+    if os.path.isdir(source_path):
+        shutil.copytree(source_path, dest_path)
+    else:
+        shutil.copy2(source_path, dest_path)
+    return "copied"
+
+
+def _write_json_file(path: str, payload: Dict[str, Any]) -> None:
+    os.makedirs(os.path.dirname(path), exist_ok=True)
+    temp_path = f"{path}.tmp"
+    try:
+        with open(temp_path, "w", encoding="utf-8") as handle:
+            json.dump(payload, handle, ensure_ascii=False, indent=2)
+        os.replace(temp_path, path)
+    finally:
+        if os.path.exists(temp_path):
+            try:
+                os.remove(temp_path)
+            except OSError:
+                pass
+
+
+async def run_dinsar_source_bundle_items(
+    task_id: str,
+    items: List[Dict[str, Any]],
+    dest_dir: str,
+    *,
+    include_orbit_files: bool = True,
+    skip_existing: bool = True,
+    max_items: Optional[int] = None,
+) -> None:
+    try:
+        if max_items is not None:
+            try:
+                max_items = int(max_items)
+            except (TypeError, ValueError):
+                max_items = None
+            if max_items is not None and max_items <= 0:
+                max_items = None
+
+        await task_service.start_task(task_id, message="Starting D-InSAR source bundle export...")
+        await _log_and_update(task_id, f"D-InSAR source bundle export started. Dest: {dest_dir}")
+        await _log_and_update(
+            task_id,
+            (
+                "D-InSAR source bundle options: "
+                f"include_orbit_files={include_orbit_files}, "
+                f"skip_existing={skip_existing}, "
+                f"max_items={max_items if max_items is not None else 'unlimited'}"
+            ),
+        )
+
+        await asyncio.to_thread(os.makedirs, dest_dir, exist_ok=True)
+        data_dir = os.path.join(dest_dir, "data")
+        orbit_dir = os.path.join(dest_dir, "orbit")
+        await asyncio.to_thread(os.makedirs, data_dir, exist_ok=True)
+        if include_orbit_files:
+            await asyncio.to_thread(os.makedirs, orbit_dir, exist_ok=True)
+
+        pairs_path = os.path.join(dest_dir, "pairs.json")
+        manifest_path = os.path.join(dest_dir, "manifest.json")
+        existing_pairs_payload = await asyncio.to_thread(_read_json_file, pairs_path)
+        existing_manifest_payload = await asyncio.to_thread(_read_json_file, manifest_path)
+
+        raw_existing_pairs = existing_pairs_payload.get("pairs")
+        existing_pairs: List[Dict[str, Any]] = [
+            dict(pair)
+            for pair in raw_existing_pairs
+            if isinstance(pair, dict)
+        ] if isinstance(raw_existing_pairs, list) else []
+
+        raw_existing_scenes = existing_manifest_payload.get("scenes")
+        existing_scene_list: List[Dict[str, Any]] = [
+            dict(scene)
+            for scene in raw_existing_scenes
+            if isinstance(scene, dict)
+        ] if isinstance(raw_existing_scenes, list) else []
+
+        raw_existing_orbits = existing_manifest_payload.get("orbits")
+        existing_orbit_list: List[Dict[str, Any]] = [
+            dict(orbit)
+            for orbit in raw_existing_orbits
+            if isinstance(orbit, dict)
+        ] if isinstance(raw_existing_orbits, list) else []
+
+        existing_pair_keys = {
+            key
+            for pair in existing_pairs
+            for key in _source_bundle_pair_keys_from_pair(pair)
+        }
+        if existing_pairs:
+            await _log_and_update(task_id, f"Existing source bundle pairs found: {len(existing_pairs)}")
+
+        candidate_tasks: List[Dict[str, Any]] = []
+        for item in items:
+            task_name = item.get("task_name") or item.get("task_alias") or "task"
+            task_alias = item.get("task_alias") or task_name
+            master_path = item.get("master_path")
+            slave_path = item.get("slave_path")
+            if not master_path or not slave_path:
+                continue
+            candidate_tasks.append(
+                {
+                    **item,
+                    "task_name": task_name,
+                    "task_alias": task_alias,
+                    "master_path": master_path,
+                    "slave_path": slave_path,
+                }
+            )
+
+        skipped_existing_pair_count = 0
+        exportable_tasks: List[Dict[str, Any]] = []
+        for item in candidate_tasks:
+            pair_keys = _source_bundle_pair_keys_from_item(item)
+            item["identity_key"] = pair_keys[0] if pair_keys else None
+            if existing_pair_keys and any(key in existing_pair_keys for key in pair_keys):
+                skipped_existing_pair_count += 1
+                continue
+            exportable_tasks.append(item)
+
+        if max_items is not None:
+            deferred_count = max(0, len(exportable_tasks) - max_items)
+            tasks = exportable_tasks[:max_items]
+        else:
+            deferred_count = 0
+            tasks = exportable_tasks
+
+        total_pairs = len(tasks)
+        await _log_and_update(
+            task_id,
+            (
+                f"Found {len(candidate_tasks)} candidate pairs; "
+                f"{skipped_existing_pair_count} already exported; "
+                f"{total_pairs} new pairs selected for this run."
+            ),
+        )
+        if total_pairs == 0:
+            await task_service.update_task(
+                task_id,
+                status="COMPLETED",
+                message=(
+                    "No new source bundle pairs to export. "
+                    f"Existing pairs: {len(existing_pairs)}; "
+                    f"already exported in request: {skipped_existing_pair_count}."
+                ),
+                progress=100,
+            )
+            return
+
+        scene_records: List[Dict[str, Any]] = []
+        scene_entries_by_source: Dict[str, Dict[str, Any]] = {}
+        scene_entries_by_relative: Dict[str, Dict[str, Any]] = {}
+        scene_entries_by_id: Dict[str, Dict[str, Any]] = {}
+        orbit_records: List[Dict[str, Any]] = []
+        orbit_entries_by_source: Dict[str, Dict[str, Any]] = {}
+        orbit_entries_by_relative: Dict[str, Dict[str, Any]] = {}
+        orbit_entries_by_id: Dict[str, Dict[str, Any]] = {}
+        next_scene_number = _next_source_bundle_id(existing_scene_list, "scene_id", "scene")
+        next_orbit_number = _next_source_bundle_id(existing_orbit_list, "orbit_id", "orbit")
+        next_pair_number = _next_source_bundle_id(existing_pairs, "pair_id", "pair")
+        copy_units_by_target: Dict[str, Tuple[str, str, str, str]] = {}
+        pairs: List[Dict[str, Any]] = []
+        missing_sources: List[Dict[str, Any]] = []
+
+        def register_scene_entry(raw_entry: Dict[str, Any]) -> Dict[str, Any]:
+            nonlocal next_scene_number
+            entry = dict(raw_entry)
+            source_path = entry.get("source_path")
+            source_key = ""
+            if source_path:
+                source_path = os.path.normpath(os.path.abspath(str(source_path)))
+                source_key = os.path.normcase(source_path)
+                entry["source_path"] = source_path
+            relative_path = _normalize_bundle_relative_path(entry.get("relative_path"))
+            if not relative_path and source_path:
+                relative_path = _bundle_relative_scene_path(source_path)
+            if relative_path:
+                entry["relative_path"] = relative_path
+
+            existing = None
+            if source_key:
+                existing = scene_entries_by_source.get(source_key)
+            if existing is None and relative_path:
+                existing = scene_entries_by_relative.get(relative_path)
+            scene_id = str(entry.get("scene_id") or "").strip()
+            if existing is None and scene_id:
+                existing = scene_entries_by_id.get(scene_id)
+            if existing is not None:
+                if source_path and not existing.get("source_path"):
+                    existing["source_path"] = source_path
+                    scene_entries_by_source[source_key] = existing
+                if relative_path and not existing.get("relative_path"):
+                    existing["relative_path"] = relative_path
+                    scene_entries_by_relative[relative_path] = existing
+                return existing
+
+            if not scene_id:
+                scene_id = f"scene_{next_scene_number:04d}"
+                entry["scene_id"] = scene_id
+                next_scene_number += 1
+            else:
+                next_scene_number = max(next_scene_number, _source_bundle_id_number(scene_id, "scene") + 1)
+
+            scene_records.append(entry)
+            scene_entries_by_id[scene_id] = entry
+            if source_key:
+                scene_entries_by_source[source_key] = entry
+            if relative_path:
+                scene_entries_by_relative[relative_path] = entry
+            return entry
+
+        def register_orbit_entry(raw_entry: Dict[str, Any]) -> Dict[str, Any]:
+            nonlocal next_orbit_number
+            entry = dict(raw_entry)
+            source_path = entry.get("source_path")
+            source_key = ""
+            if source_path:
+                source_path = os.path.normpath(os.path.abspath(str(source_path)))
+                source_key = os.path.normcase(source_path)
+                entry["source_path"] = source_path
+            relative_path = _normalize_bundle_relative_path(entry.get("relative_path"))
+            if not relative_path and source_path:
+                relative_path = os.path.join("orbit", _safe_bundle_entry_name(source_path, "orbit")).replace(os.sep, "/")
+            if relative_path:
+                entry["relative_path"] = relative_path
+
+            existing = None
+            if source_key:
+                existing = orbit_entries_by_source.get(source_key)
+            if existing is None and relative_path:
+                existing = orbit_entries_by_relative.get(relative_path)
+            orbit_id = str(entry.get("orbit_id") or "").strip()
+            if existing is None and orbit_id:
+                existing = orbit_entries_by_id.get(orbit_id)
+            if existing is not None:
+                if source_path and not existing.get("source_path"):
+                    existing["source_path"] = source_path
+                    orbit_entries_by_source[source_key] = existing
+                if relative_path and not existing.get("relative_path"):
+                    existing["relative_path"] = relative_path
+                    orbit_entries_by_relative[relative_path] = existing
+                return existing
+
+            if not orbit_id:
+                orbit_id = f"orbit_{next_orbit_number:04d}"
+                entry["orbit_id"] = orbit_id
+                next_orbit_number += 1
+            else:
+                next_orbit_number = max(next_orbit_number, _source_bundle_id_number(orbit_id, "orbit") + 1)
+            if not isinstance(entry.get("used_by"), list):
+                entry["used_by"] = []
+
+            orbit_records.append(entry)
+            orbit_entries_by_id[orbit_id] = entry
+            if source_key:
+                orbit_entries_by_source[source_key] = entry
+            if relative_path:
+                orbit_entries_by_relative[relative_path] = entry
+            return entry
+
+        def add_copy_unit(kind: str, unit_id: str, source_path: Optional[str], relative_path: Optional[str]) -> None:
+            if not source_path or not relative_path:
+                return
+            target_path = os.path.join(dest_dir, _normalize_bundle_relative_path(relative_path))
+            target_key = os.path.normcase(os.path.abspath(target_path))
+            copy_units_by_target.setdefault(target_key, (kind, unit_id, source_path, target_path))
+
+        def add_orbit_usage(entry: Dict[str, Any], scene_id: str, role: str) -> None:
+            used_by = entry.setdefault("used_by", [])
+            usage = {"scene_id": scene_id, "role": role}
+            if not any(
+                str(item.get("scene_id")) == scene_id and str(item.get("role")) == role
+                for item in used_by
+                if isinstance(item, dict)
+            ):
+                used_by.append(usage)
+
+        for scene in existing_scene_list:
+            register_scene_entry(scene)
+
+        for orbit in existing_orbit_list:
+            register_orbit_entry(orbit)
+
+        for pair in existing_pairs:
+            for role in ("master", "slave"):
+                scene_relative = _normalize_bundle_relative_path(pair.get(f"{role}_data"))
+                if scene_relative:
+                    register_scene_entry(
+                        {
+                            "scene_id": pair.get(f"{role}_scene_id"),
+                            "source_path": pair.get(f"{role}_source_path"),
+                            "relative_path": scene_relative,
+                        }
+                    )
+                orbit_relative = _normalize_bundle_relative_path(pair.get(f"{role}_orbit"))
+                if orbit_relative:
+                    register_orbit_entry(
+                        {
+                            "orbit_id": pair.get(f"{role}_orbit_id"),
+                            "source_path": pair.get(f"{role}_orbit_source_path"),
+                            "relative_path": orbit_relative,
+                            "used_by": [
+                                {
+                                    "scene_id": pair.get(f"{role}_scene_id"),
+                                    "role": role,
+                                }
+                            ] if pair.get(f"{role}_scene_id") else [],
+                        }
+                    )
+
+        def ensure_scene_entry(scene_path: str, role_item: Dict[str, Any], prefix: str) -> Dict[str, Any]:
+            source_path = os.path.normpath(os.path.abspath(str(scene_path)))
+            source_key = os.path.normcase(source_path)
+            relative_path = _bundle_relative_scene_path(source_path)
+            existing = scene_entries_by_source.get(source_key) or scene_entries_by_relative.get(relative_path)
+            if existing:
+                if not existing.get("source_path"):
+                    existing["source_path"] = source_path
+                    scene_entries_by_source[source_key] = existing
+                if not existing.get("relative_path"):
+                    existing["relative_path"] = relative_path
+                    scene_entries_by_relative[relative_path] = existing
+                add_copy_unit("scene", existing["scene_id"], source_path, existing["relative_path"])
+                return existing
+            entry = register_scene_entry(
+                {
+                    "source_path": source_path,
+                    "relative_path": relative_path,
+                    "satellite": role_item.get("satellite"),
+                    "imaging_date": role_item.get("imaging_date"),
+                    "imaging_mode": role_item.get("imaging_mode"),
+                    "polarization": role_item.get("polarization"),
+                }
+            )
+            add_copy_unit("scene", entry["scene_id"], source_path, entry["relative_path"])
+            return entry
+
+        def ensure_orbit_entry(orbit_path: Optional[str], scene_id: str, role: str) -> Optional[Dict[str, Any]]:
+            if not include_orbit_files or not orbit_path:
+                return None
+            source_path = os.path.normpath(os.path.abspath(str(orbit_path)))
+            source_key = os.path.normcase(source_path)
+            relative_path = os.path.join("orbit", _safe_bundle_entry_name(source_path, "orbit")).replace(os.sep, "/")
+            existing = orbit_entries_by_source.get(source_key) or orbit_entries_by_relative.get(relative_path)
+            if existing:
+                if not existing.get("source_path"):
+                    existing["source_path"] = source_path
+                    orbit_entries_by_source[source_key] = existing
+                if not existing.get("relative_path"):
+                    existing["relative_path"] = relative_path
+                    orbit_entries_by_relative[relative_path] = existing
+                add_orbit_usage(existing, scene_id, role)
+                add_copy_unit("orbit", existing["orbit_id"], source_path, existing["relative_path"])
+                return existing
+            entry = register_orbit_entry(
+                {
+                    "source_path": source_path,
+                    "relative_path": relative_path,
+                }
+            )
+            add_orbit_usage(entry, scene_id, role)
+            add_copy_unit("orbit", entry["orbit_id"], source_path, entry["relative_path"])
+            return entry
+
+        for item in tasks:
+            task_name = item["task_name"]
+            task_alias = item["task_alias"]
+            master_entry = ensure_scene_entry(
+                item["master_path"],
+                {
+                    "satellite": item.get("master_satellite"),
+                    "imaging_date": item.get("master_imaging_date"),
+                    "imaging_mode": item.get("master_imaging_mode"),
+                    "polarization": item.get("master_polarization"),
+                },
+                "scene",
+            )
+            slave_entry = ensure_scene_entry(
+                item["slave_path"],
+                {
+                    "satellite": item.get("slave_satellite"),
+                    "imaging_date": item.get("slave_imaging_date"),
+                    "imaging_mode": item.get("slave_imaging_mode"),
+                    "polarization": item.get("slave_polarization"),
+                },
+                "scene",
+            )
+            master_orbit = ensure_orbit_entry(
+                item.get("master_orbit_file_path"),
+                master_entry["scene_id"],
+                "master",
+            )
+            slave_orbit = ensure_orbit_entry(
+                item.get("slave_orbit_file_path"),
+                slave_entry["scene_id"],
+                "slave",
+            )
+            master_source_path = os.path.normpath(os.path.abspath(str(item["master_path"])))
+            slave_source_path = os.path.normpath(os.path.abspath(str(item["slave_path"])))
+            pairs.append(
+                {
+                    "pair_id": f"pair_{next_pair_number:04d}",
+                    "identity_key": item.get("identity_key"),
+                    "task_name": task_name,
+                    "task_alias": task_alias,
+                    "pair_key": item.get("pair_key"),
+                    "scene_pair_uid": item.get("scene_pair_uid") or item.get("pair_uid"),
+                    "pair_uid": item.get("pair_uid") or item.get("scene_pair_uid"),
+                    "master_scene_id": master_entry["scene_id"],
+                    "slave_scene_id": slave_entry["scene_id"],
+                    "master_source_path": master_source_path,
+                    "slave_source_path": slave_source_path,
+                    "master_data": master_entry["relative_path"],
+                    "slave_data": slave_entry["relative_path"],
+                    "master_orbit_id": master_orbit.get("orbit_id") if master_orbit else None,
+                    "slave_orbit_id": slave_orbit.get("orbit_id") if slave_orbit else None,
+                    "master_orbit_source_path": master_orbit.get("source_path") if master_orbit else None,
+                    "slave_orbit_source_path": slave_orbit.get("source_path") if slave_orbit else None,
+                    "master_orbit": master_orbit.get("relative_path") if master_orbit else None,
+                    "slave_orbit": slave_orbit.get("relative_path") if slave_orbit else None,
+                    "master_imaging_date": item.get("master_imaging_date"),
+                    "slave_imaging_date": item.get("slave_imaging_date"),
+                    "time_baseline_days": item.get("time_baseline_days"),
+                    "scene_center_distance_meters": item.get("scene_center_distance_meters"),
+                    "spatial_baseline_meters": item.get("spatial_baseline_meters"),
+                    "network_run_id": item.get("network_run_id"),
+                    "network_edge_id": item.get("network_edge_id"),
+                    "policy_version": item.get("policy_version"),
+                    "selection_strategy": item.get("selection_strategy"),
+                }
+            )
+            next_pair_number += 1
+
+        scene_list = sorted(
+            scene_records,
+            key=lambda entry: (_source_bundle_id_number(entry.get("scene_id"), "scene"), str(entry.get("scene_id") or "")),
+        )
+        orbit_list = sorted(
+            orbit_records,
+            key=lambda entry: (_source_bundle_id_number(entry.get("orbit_id"), "orbit"), str(entry.get("orbit_id") or "")),
+        )
+        copy_units = list(copy_units_by_target.values())
+
+        copied_count = 0
+        skipped_count = 0
+        failed_count = 0
+        total_units = max(1, len(copy_units))
+        for index, (kind, unit_id, source_path, target_path) in enumerate(copy_units, start=1):
+            prog = int(((index - 1) / total_units) * 90)
+            await task_service.update_task(
+                task_id,
+                progress=prog,
+                message=f"Bundling {kind} ({index}/{total_units}): {unit_id}",
+            )
+            await _log_and_update(task_id, f"[{index}/{total_units}] Bundling {kind}: {unit_id}")
+            if not os.path.exists(source_path):
+                await _log_and_update(task_id, f" -> Missing source: {source_path}")
+                missing_sources.append({"kind": kind, "id": unit_id, "source_path": source_path})
+                failed_count += 1
+                continue
+            try:
+                action = await asyncio.to_thread(_copy_source_into_bundle, source_path, target_path, skip_existing)
+                if action == "skipped":
+                    skipped_count += 1
+                    await _log_and_update(task_id, " -> Skipped (already exists in bundle)")
+                else:
+                    copied_count += 1
+                    await _log_and_update(task_id, " -> Success")
+            except PermissionError:
+                await _log_and_update(task_id, " -> Failed: permission denied")
+                failed_count += 1
+            except Exception as exc:
+                await _log_and_update(task_id, f" -> Failed: {exc}")
+                failed_count += 1
+
+        combined_pairs = existing_pairs + pairs
+        final_msg = (
+            "D-InSAR source bundle export finished. "
+            f"Pairs {len(combined_pairs)} (+{len(pairs)}), Scenes {len(scene_list)}, "
+            f"Orbits {len(orbit_list)}, "
+            f"Copied {copied_count}, Skipped {skipped_count}, "
+            f"Already exported pairs {skipped_existing_pair_count}, "
+            f"Deferred {deferred_count}, Failed {failed_count}"
+        )
+        if failed_count > 0:
+            await task_service.update_task(task_id, status="FAILED", message=final_msg, progress=100)
+            raise CopyTaskExecutionError(final_msg)
+
+        exported_at = datetime.utcnow().isoformat(timespec="seconds") + "Z"
+        pairs_payload = {
+            "schema": "dinsar_source_bundle_pairs.v1",
+            "exported_at": exported_at,
+            "pairs": combined_pairs,
+        }
+        manifest_payload = {
+            "schema": "dinsar_source_bundle_manifest.v1",
+            "exported_at": exported_at,
+            "package_format": "source_bundle",
+            "destination": os.path.normpath(os.path.abspath(dest_dir)),
+            "pair_count": len(combined_pairs),
+            "new_pair_count": len(pairs),
+            "existing_pair_count": len(existing_pairs),
+            "candidate_pair_count": len(candidate_tasks),
+            "skipped_existing_pairs": skipped_existing_pair_count,
+            "scene_count": len(scene_list),
+            "orbit_count": len(orbit_list),
+            "include_orbit_files": bool(include_orbit_files),
+            "skip_existing": bool(skip_existing),
+            "max_items": max_items,
+            "deferred_pairs": deferred_count,
+            "copied_units": copied_count,
+            "skipped_units": skipped_count,
+            "failed_units": failed_count,
+            "missing_sources": missing_sources,
+            "directories": {
+                "data": "data",
+                "orbit": "orbit" if orbit_list else None,
+            },
+            "scenes": scene_list,
+            "orbits": orbit_list,
+        }
+        await asyncio.to_thread(_write_json_file, pairs_path, pairs_payload)
+        await asyncio.to_thread(_write_json_file, manifest_path, manifest_payload)
+
+        await _log_and_update(task_id, final_msg, progress=100)
+        await task_service.update_task(task_id, status="COMPLETED", message=final_msg, progress=100)
+    except CopyTaskExecutionError:
+        raise
+    except Exception as e:
+        fail_msg = f"D-InSAR source bundle fatal error: {e}"
+        try:
+            await task_service.update_task(task_id, status="FAILED", message=fail_msg)
+        except Exception:
+            pass
+        raise CopyTaskExecutionError(fail_msg) from e
+
+
 async def run_ps_copy_items(task_id: str, items: List[Dict[str, Any]], dest_dir: str) -> None:
     try:
         await task_service.start_task(task_id, message="Starting PS-InSAR copy task...")
@@ -247,8 +967,18 @@ async def run_dinsar_copy_items(
     *,
     include_orbit_files: bool = False,
     export_zip: bool = False,
+    skip_existing: bool = True,
+    max_items: Optional[int] = None,
 ) -> None:
     try:
+        if max_items is not None:
+            try:
+                max_items = int(max_items)
+            except (TypeError, ValueError):
+                max_items = None
+            if max_items is not None and max_items <= 0:
+                max_items = None
+
         await task_service.start_task(task_id, message="Starting D-InSAR copy task...")
         await _log_and_update(task_id, f"D-InSAR copy started. Dest: {dest_dir}")
         await _log_and_update(
@@ -256,7 +986,9 @@ async def run_dinsar_copy_items(
             (
                 "D-InSAR copy options: "
                 f"include_orbit_files={include_orbit_files}, "
-                f"export_zip={export_zip}"
+                f"export_zip={export_zip}, "
+                f"skip_existing={skip_existing}, "
+                f"max_items={max_items if max_items is not None else 'unlimited'}"
             ),
         )
 
@@ -295,7 +1027,10 @@ async def run_dinsar_copy_items(
             return
 
         success_count = 0
+        skipped_count = 0
         failed_count = 0
+        deferred_count = 0
+        attempted_count = 0
         for i, item in enumerate(tasks, start=1):
             task_name = item["task_name"]
             task_alias = item["task_alias"]
@@ -308,6 +1043,28 @@ async def run_dinsar_copy_items(
 
             staging_root: Optional[str] = None
             try:
+                final_task_dir = os.path.join(dest_dir, task_alias)
+                zip_path = os.path.join(dest_dir, f"{task_alias}.zip") if export_zip else None
+
+                if skip_existing:
+                    if export_zip and zip_path and _is_existing_dinsar_zip_complete(zip_path):
+                        await _log_and_update(task_id, " -> Skipped (existing zip package)")
+                        skipped_count += 1
+                        continue
+                    if (not export_zip) and _is_existing_dinsar_folder_complete(final_task_dir):
+                        await _log_and_update(task_id, " -> Skipped (existing master/slave folders)")
+                        skipped_count += 1
+                        continue
+
+                if max_items is not None and attempted_count >= max_items:
+                    deferred_count = total - i + 1
+                    await _log_and_update(
+                        task_id,
+                        f"Reached max_items={max_items}; deferred {deferred_count} remaining candidates.",
+                    )
+                    break
+                attempted_count += 1
+
                 if export_zip:
                     staging_root = await asyncio.to_thread(
                         tempfile.mkdtemp,
@@ -315,10 +1072,13 @@ async def run_dinsar_copy_items(
                         dir=dest_dir,
                     )
                     task_dir = os.path.join(staging_root, task_alias)
-                    zip_path = os.path.join(dest_dir, f"{task_alias}.zip")
                 else:
-                    task_dir = os.path.join(dest_dir, task_alias)
+                    staging_root = await asyncio.to_thread(
-                    zip_path = None
+                        tempfile.mkdtemp,
+                        prefix="._dinsar_copy_",
+                        dir=dest_dir,
+                    )
+                    task_dir = os.path.join(staging_root, task_alias)
                 master_dir = os.path.join(task_dir, "master")
                 slave_dir = os.path.join(task_dir, "slave")
 
@@ -345,40 +1105,26 @@ async def run_dinsar_copy_items(
                 await asyncio.to_thread(
                     write_pair_metadata,
                     task_dir,
-                    {
+                    _build_dinsar_pair_metadata(
-                        "pair_key": item.get("pair_key"),
+                        item,
-                        "task_name": task_name,
+                        task_name,
-                        "task_alias": task_alias,
+                        task_alias,
-                        "master_path": item.get("master_path"),
+                        "zip" if export_zip else "folder",
-                        "slave_path": item.get("slave_path"),
+                        include_orbit_files,
-                        "master_satellite": item.get("master_satellite"),
+                        orbit_entries,
-                        "slave_satellite": item.get("slave_satellite"),
+                    ),
-                        "master_imaging_date": item.get("master_imaging_date"),
-                        "slave_imaging_date": item.get("slave_imaging_date"),
-                        "master_imaging_mode": item.get("master_imaging_mode"),
-                        "slave_imaging_mode": item.get("slave_imaging_mode"),
-                        "master_polarization": item.get("master_polarization"),
-                        "slave_polarization": item.get("slave_polarization"),
-                        "time_baseline_days": item.get("time_baseline_days"),
-                        "spatial_baseline_meters": item.get("spatial_baseline_meters"),
-                        "scene_center_distance_meters": item.get("scene_center_distance_meters"),
-                        "package_format": "zip" if export_zip else "folder",
-                        "include_orbit_files": bool(include_orbit_files),
-                        "master_orbit_file_path": item.get("master_orbit_file_path"),
-                        "slave_orbit_file_path": item.get("slave_orbit_file_path"),
-                        "orbit_files": orbit_entries,
-                        "scene_pair_uid": item.get("scene_pair_uid") or item.get("pair_uid"),
-                        "pair_uid": item.get("pair_uid") or item.get("scene_pair_uid"),
-                        "network_run_id": item.get("network_run_id"),
-                        "network_edge_id": item.get("network_edge_id"),
-                        "policy_version": item.get("policy_version"),
-                        "selection_strategy": item.get("selection_strategy"),
-                        "copied_at": datetime.utcnow().isoformat(timespec="seconds") + "Z",
-                    },
                 )
                 if export_zip and zip_path:
                     await asyncio.to_thread(_zip_task_directory, task_dir, zip_path)
                     await _log_and_update(task_id, f" -> ZIP: {zip_path}")
+                elif not export_zip:
+                    if os.path.exists(final_task_dir):
+                        raise CopyTaskExecutionError(
+                            "Destination folder already exists and cannot be atomically replaced: "
+                            f"{final_task_dir}"
+                        )
+                    await asyncio.to_thread(os.replace, task_dir, final_task_dir)
+                    await _log_and_update(task_id, f" -> Folder: {final_task_dir}")
 
                 await _log_and_update(task_id, " -> Success")
                 success_count += 1
@@ -394,7 +1140,8 @@ async def run_dinsar_copy_items(
 
         final_msg = (
             f"D-InSAR copy finished. Mode {'zip' if export_zip else 'folder'}. "
-            f"Success {success_count}, Failed {failed_count}"
+            f"Copied {success_count}, Skipped {skipped_count}, "
+            f"Deferred {deferred_count}, Failed {failed_count}"
         )
         await _log_and_update(task_id, final_msg, progress=100)
         if failed_count > 0:

backend/app/routers/tools.py

@@ -27,6 +27,13 @@ COPY_BATCH_MAX_STATUS_COUNT = read_int_env(
     minimum=1,
     maximum=64,
 )
+COPY_BATCH_MAX_COPY_ITEMS = read_int_env(
+    "COPY_BATCH_MAX_COPY_ITEMS",
+    5000,
+    minimum=1,
+    maximum=200000,
+)
+COPY_DINSAR_PACKAGE_MODES = {"task_folder", "task_zip", "source_bundle"}
 
 
 class CopyBatchRequest(BaseModel):
@@ -35,6 +42,9 @@ class CopyBatchRequest(BaseModel):
     copy_statuses: Optional[List[str]] = None
     include_orbit_files: bool = False
     export_zip: bool = False
+    package_mode: str = "task_folder"
+    skip_existing: bool = True
+    max_items: Optional[int] = None
 
     @field_validator("batch_id", "dest_dir", mode="before")
     @classmethod
@@ -57,6 +67,33 @@ class CopyBatchRequest(BaseModel):
             )
         return value
 
+    @field_validator("max_items", mode="before")
+    @classmethod
+    def _normalize_max_items(cls, value):
+        if value in (None, ""):
+            return None
+        try:
+            parsed = int(value)
+        except (TypeError, ValueError) as exc:
+            raise ValueError("max_items must be an integer.") from exc
+        if parsed <= 0:
+            return None
+        if parsed > COPY_BATCH_MAX_COPY_ITEMS:
+            raise ValueError(
+                f"max_items exceeds max item count ({COPY_BATCH_MAX_COPY_ITEMS})."
+            )
+        return parsed
+
+    @field_validator("package_mode", mode="before")
+    @classmethod
+    def _normalize_package_mode(cls, value):
+        normalized = str(value or "task_folder").strip().lower()
+        if normalized not in COPY_DINSAR_PACKAGE_MODES:
+            raise ValueError(
+                f"package_mode must be one of: {sorted(COPY_DINSAR_PACKAGE_MODES)}."
+            )
+        return normalized
+
 
 def _normalize_copy_batch_statuses(copy_statuses: Optional[List[str]]) -> List[str]:
     if not copy_statuses:
@@ -135,13 +172,19 @@ async def copy_dinsar_pairs_endpoint(
     _validate_export_path(request.dest_dir, "dest_dir")
     try:
         copy_statuses = _normalize_copy_batch_statuses(request.copy_statuses)
+        package_mode = request.package_mode
+        if bool(request.export_zip) and package_mode == "task_folder":
+            package_mode = "task_zip"
         params = {
             "dest_dir": request.dest_dir,
             "file_type": "DINSAR_PAIRS",
             "batch_id": request.batch_id,
             "copy_statuses": copy_statuses,
             "include_orbit_files": bool(request.include_orbit_files),
-            "export_zip": bool(request.export_zip),
+            "export_zip": package_mode == "task_zip",
+            "package_mode": package_mode,
+            "skip_existing": bool(request.skip_existing),
+            "max_items": request.max_items,
         }
         task_id = await task_service.create_task("COPY_DATA", f"D-InSAR 数据分发: {request.dest_dir}", params=params)
 
@@ -151,7 +194,10 @@ async def copy_dinsar_pairs_endpoint(
             "batch_id": request.batch_id,
             "copy_statuses": copy_statuses,
             "include_orbit_files": bool(request.include_orbit_files),
-            "export_zip": bool(request.export_zip),
+            "export_zip": package_mode == "task_zip",
+            "package_mode": package_mode,
+            "skip_existing": bool(request.skip_existing),
+            "max_items": request.max_items,
         }
         await job_queue_service.create_job("COPY_DATA", payload=payload, task_id=task_id)
         await _add_operation_audit_log(
@@ -165,7 +211,10 @@ async def copy_dinsar_pairs_endpoint(
                 "dest_dir": request.dest_dir,
                 "copy_statuses": copy_statuses,
                 "include_orbit_files": bool(request.include_orbit_files),
-                "export_zip": bool(request.export_zip),
+                "export_zip": package_mode == "task_zip",
+                "package_mode": package_mode,
+                "skip_existing": bool(request.skip_existing),
+                "max_items": request.max_items,
             },
         )
         await db.commit()

backend/app/services/job_handlers.py

@@ -50,7 +50,7 @@ from .timeseries_service import (
     timeseries_service,
 )
 from .unpack_service import run_unpack_task
-from ..copier import run_ps_copy_items, run_dinsar_copy_items
+from ..copier import run_ps_copy_items, run_dinsar_copy_items, run_dinsar_source_bundle_items
 from ..ai_service import (
     train_quality_model,
     predict_quality,
@@ -116,6 +116,16 @@ def _normalize_copy_statuses(raw_statuses: Any) -> List[str]:
     return normalized or ["COMPLETED"]
 
 
+def _normalize_positive_int(value: Any) -> Optional[int]:
+    if value in (None, ""):
+        return None
+    try:
+        parsed = int(value)
+    except (TypeError, ValueError):
+        return None
+    return parsed if parsed > 0 else None
+
+
 def _dedupe_existing_dirs(paths: Any) -> List[str]:
     ordered: List[str] = []
     for raw_path in paths or []:
@@ -289,6 +299,9 @@ async def _handle_copy_data(job: SystemJobORM) -> None:
     copy_statuses = _normalize_copy_statuses(payload.get("copy_statuses"))
     include_orbit_files = bool(payload.get("include_orbit_files"))
     export_zip = bool(payload.get("export_zip"))
+    package_mode = str(payload.get("package_mode") or ("task_zip" if export_zip else "task_folder")).strip().lower()
+    skip_existing = payload.get("skip_existing") is not False
+    max_items = _normalize_positive_int(payload.get("max_items"))
 
     if not batch_id:
         raise ValueError("COPY_DATA requires batch_id payload.")
@@ -377,12 +390,24 @@ async def _handle_copy_data(job: SystemJobORM) -> None:
                 raise ValueError(
                     f"No D-InSAR items matched copy statuses: {', '.join(copy_statuses)}"
                 )
+            if package_mode in {"source_bundle", "bundle", "dedupe_source"}:
+                await run_dinsar_source_bundle_items(
+                    job.task_id,
+                    items,
+                    dest_dir,
+                    include_orbit_files=include_orbit_files,
+                    skip_existing=skip_existing,
+                    max_items=max_items,
+                )
+            else:
                 await run_dinsar_copy_items(
                     job.task_id,
                     items,
                     dest_dir,
                     include_orbit_files=include_orbit_files,
-                export_zip=export_zip,
+                    export_zip=(package_mode == "task_zip" or export_zip),
+                    skip_existing=skip_existing,
+                    max_items=max_items,
                 )
             return
 

docs/DINSAR_PAIRING_DISTRIBUTION_LOGIC_20260508.md

@@ -279,7 +279,10 @@ score =
 - `dest_dir`
 - `copy_statuses`，为空时默认 `["COMPLETED"]`
 - `include_orbit_files`，默认 `false`；为 `true` 时把 master/slave 精轨复制到 Task 内的 `orbit/`
-- `export_zip`，默认 `false`；为 `true` 时每个 Task 输出为一个 `.zip` 包
+- `package_mode`，支持 `task_folder`、`task_zip`、`source_bundle`
+- `export_zip`，兼容旧参数；为 `true` 且 `package_mode=task_folder` 时等价于 `task_zip`
+- `skip_existing`，默认 `true`
+- `max_items`，每次最多处理的新 Task 或新 pair 数量；为空或 0 表示不限制
 
 后端动作：
 
@@ -288,7 +291,9 @@ score =
 3. 创建 `SystemJob`，job_type 也是 `COPY_DATA`。
 4. worker 领取 job 后进入 `job_handlers._handle_copy_data()`。
 5. `_handle_copy_data()` 根据 `batch_id` 查询 `dinsar_task_items`，只取 `copy_statuses` 命中的条目。
-6. 调用 [backend/app/copier.py](../backend/app/copier.py) 的 `run_dinsar_copy_items()`。
+6. 根据 `package_mode` 调用 [backend/app/copier.py](../backend/app/copier.py) 的 `run_dinsar_copy_items()` 或 `run_dinsar_source_bundle_items()`。
+
+### 9.1 Task 文件夹 / ZIP 模式
 
 `run_dinsar_copy_items()` 对每个 item 执行：
 
@@ -298,8 +303,10 @@ score =
 - slave 目录：`<task_alias>/slave`
 - 如果启用 `include_orbit_files`，从 `radar_data.orbit_file_path` 找 master/slave 精轨并复制到 `<task_alias>/orbit/`
 - 直接复制配对时保存的原始产品目录；D-InSAR 分发不再优先使用 `envi_import/`
-- 使用 `shutil.copytree(..., dirs_exist_ok=True)` 复制 master/slave
+- 先复制到临时目录，完成后再替换为最终 Task 目录或 ZIP，避免留下半成品
 - 写入 `<task_alias>/.dinsar_pair.json`
+- `skip_existing=true` 时，文件夹模式检查 `<task_alias>/master` 和 `<task_alias>/slave` 非空即跳过；ZIP 模式只检查 `<task_alias>.zip` 存在且大小大于 0，不打开 ZIP 做深度校验
+- `max_items` 限制本次新复制数量；已跳过的既有 Task 不消耗本次额度
 
 `.dinsar_pair.json` 是后续生产追踪的关键 sidecar，包含：
 
@@ -319,7 +326,32 @@ score =
 - `selection_strategy`
 - `copied_at`
 
-当前实现不会清空已有 Task 目录，而是合并复制；如果目标已有旧文件，需要人工确认目录状态。
+当前实现不再合并写入已有 Task 目录。目标 Task 已完整存在时跳过；目标同名目录存在但不完整时，为避免误覆盖，会报错并要求人工处理。
+
+### 9.2 去重源数据包模式
+
+`package_mode=source_bundle` 时，分发不生成每个 `Task_*`，而是在同一个目标目录内维护：
+
+```text
+<dest_dir>/
+  data/
+  orbit/
+  pairs.json
+  manifest.json
+```
+
+规则：
+
+- `data/` 只复制唯一源影像目录或文件，命名为 `scene_<source_path_hash>_<source_name>`。
+- `orbit/` 只复制唯一精密轨道文件，命名为 `orbit_<source_path_hash>_<orbit_name>`。
+- `pairs.json` 记录每个 pair 的 master/slave 数据相对路径、轨道相对路径、pair 元数据和 `identity_key`。
+- `manifest.json` 记录 package 统计信息、场景清单、轨道清单和本次追加统计。
+- 同一目标目录再次分发时，系统先读取已有 `pairs.json/manifest.json`，根据 `identity_key`、`scene_pair_uid/pair_uid`、`pair_key`、`network_run_id + network_edge_id`、源路径或 bundle 相对路径识别已导出的 pair。
+- `max_items` 在跳过已导出 pair 后生效。因此 500 个 pair 第一次限制 100，第二次同一目录限制 100，会追加下一批未导出的 100 个 pair。
+- `data/` / `orbit/` 仍按文件存在性跳过重复复制；pair 级续跑以 `pairs.json` 为准。
+- `pairs.json` 和 `manifest.json` 写入时先写临时文件，再原子替换。
+
+这个模式面向外部分发和后续离线还原，不直接作为本系统生产输入。反向还原工具任务书见 [DINSAR_SOURCE_BUNDLE_REVERSE_TOOL_TASK_20260511.md](DINSAR_SOURCE_BUNDLE_REVERSE_TOOL_TASK_20260511.md)。
 
 ## 10. 生产提交与运行分发
 

docs/DINSAR_SOURCE_BUNDLE_REVERSE_TOOL_TASK_20260511.md

@@ -0,0 +1,168 @@
+# D-InSAR 去重源数据包反向还原工具任务书
+
+日期：2026-05-11
+
+## 背景
+
+本系统新增“去重源数据包”分发模式。该模式不直接生成每个干涉对的 `Task_*` 目录，而是只分发唯一源影像、唯一精密轨道文件和配对关系文件，减少外部分发时的重复复制量。
+
+反向还原工具由任务接收方本地运行，将去重源数据包还原为传统 D-InSAR `Task_*` 目录结构。
+
+## 输入目录结构
+
+```text
+BundleRoot/
+  data/
+    scene_<hash>_<source_name>/
+    ...
+  orbit/
+    orbit_<hash>_<orbit_name>.txt
+    ...
+  pairs.json
+  manifest.json
+```
+
+`orbit/` 可能不存在，或 `pairs.json` 内某些配对的轨道字段为空。
+
+## 输出目录结构
+
+```text
+OutputRoot/
+  Task_YYYYMMDD_YYYYMMDD/
+    master/
+      <master source product content>
+    slave/
+      <slave source product content>
+    orbit/
+      <master/slave orbit files, if present>
+    .dinsar_pair.json
+```
+
+输出目录名称优先使用 `pairs.json` 内的 `task_alias`，若为空则使用 `task_name`，再为空则使用 `pair_id`。
+
+## pairs.json 关键字段
+
+```json
+{
+  "schema": "dinsar_source_bundle_pairs.v1",
+  "exported_at": "2026-05-11T00:00:00Z",
+  "pairs": [
+    {
+      "pair_id": "pair_0001",
+      "identity_key": "uid:<scene_pair_uid>",
+      "task_name": "Task_20250101_20250113",
+      "task_alias": "Task_20250101_20250113",
+      "master_source_path": "D:/Source/master",
+      "slave_source_path": "D:/Source/slave",
+      "master_scene_id": "scene_0001",
+      "slave_scene_id": "scene_0002",
+      "master_data": "data/scene_xxx_master",
+      "slave_data": "data/scene_yyy_slave",
+      "master_orbit_id": "orbit_0001",
+      "slave_orbit_id": "orbit_0002",
+      "master_orbit_source_path": "D:/Orbit/master.EOF",
+      "slave_orbit_source_path": "D:/Orbit/slave.EOF",
+      "master_orbit": "orbit/orbit_xxx.txt",
+      "slave_orbit": "orbit/orbit_yyy.txt",
+      "master_imaging_date": "20250101",
+      "slave_imaging_date": "20250113",
+      "time_baseline_days": 12
+    }
+  ]
+}
+```
+
+## 本系统分发续跑规则
+
+去重源数据包支持向同一个 `BundleRoot` 多次分发：
+
+- 每次启动时先读取目标目录内已有的 `pairs.json` 和 `manifest.json`。
+- 已导出的 pair 通过 `identity_key`、`scene_pair_uid/pair_uid`、`pair_key`、`network_run_id + network_edge_id`、`master/slave_source_path` 或 `master_data + slave_data` 识别。
+- 开启“每次最多追加新配对”时，系统会先跳过已导出的 pair，再从剩余 pair 中取下一批追加；例如 500 个 pair 第一次限制 100，第二次同一目录仍限制 100 时，会追加第 101-200 个未导出的 pair。
+- `data/` 和 `orbit/` 按源路径哈希命名，已有文件或目录在 `skip_existing` 开启时不会重复复制。
+- `pairs.json` 和 `manifest.json` 采用临时文件写入后原子替换，避免中途失败留下半写 JSON。
+
+注意：如果用户手动删除了 `pairs.json`，系统无法再根据记录判断哪些 pair 已经分发，只能根据重新生成的 `data/` 路径做源数据级去重，pair 级续跑能力会丢失。
+
+## 还原规则
+
+1. 读取 `pairs.json`。
+2. 对每个 pair 创建目标 `Task` 目录。
+3. 将 `master_data` 指向的数据复制到 `Task/master/`。
+4. 将 `slave_data` 指向的数据复制到 `Task/slave/`。
+5. 如 `master_orbit` / `slave_orbit` 存在，将轨道文件复制到 `Task/orbit/`。
+6. 生成 `.dinsar_pair.json`，至少保留：
+   - `pair_id`
+   - `identity_key`
+   - `task_name`
+   - `task_alias`
+   - `master_scene_id`
+   - `slave_scene_id`
+   - `master_data`
+   - `slave_data`
+   - `master_orbit_id`
+   - `slave_orbit_id`
+   - `master_orbit`
+   - `slave_orbit`
+   - `master_imaging_date`
+   - `slave_imaging_date`
+   - `time_baseline_days`
+   - `restored_at`
+7. 每个 Task 应采用临时目录还原，全部成功后再重命名为最终目录，避免半成品。
+
+## 覆盖策略
+
+工具应提供参数：
+
+- `--skip-existing`：默认开启。若目标 `Task/master` 和 `Task/slave` 均存在且非空，则跳过。
+- `--overwrite`：删除并重建已存在的目标 Task。
+- `--limit N`：最多还原 N 个 pair，便于分批执行。
+- `--dry-run`：只打印计划，不复制。
+
+`--skip-existing` 与 `--overwrite` 同时出现时应报错。
+
+## 校验要求
+
+启动前：
+
+- 检查 `pairs.json` 是否存在且可解析。
+- 检查 `data/` 是否存在。
+- 检查每个 pair 的 `master_data` / `slave_data` 是否存在。
+- 轨道缺失不应阻断还原，但要记录 warning。
+
+还原后：
+
+- `Task/master/` 非空。
+- `Task/slave/` 非空。
+- `.dinsar_pair.json` 存在。
+
+## 日志与报告
+
+工具结束后输出 `restore_report.json`：
+
+```json
+{
+  "started_at": "...",
+  "finished_at": "...",
+  "input_root": "...",
+  "output_root": "...",
+  "total_pairs": 20,
+  "restored": 18,
+  "skipped": 2,
+  "failed": 0,
+  "warnings": []
+}
+```
+
+同时建议输出人类可读日志 `restore.log`。
+
+## 建议实现
+
+建议使用 Python 3.10+：
+
+- `argparse` 处理命令行参数。
+- `pathlib.Path` 处理路径。
+- `shutil.copytree(..., dirs_exist_ok=True)` / `shutil.copy2()` 处理复制。
+- Windows 下注意长路径和权限异常。
+
+该工具不需要连接本系统数据库，也不需要调用本系统 API。

docs/INDEX.md

@@ -68,6 +68,8 @@
   Sentinel-1 D-InSAR 在不使用 ENVI + SARscape 核心时，基于 Gamma/PyINT 与 ISCE2 的可行性、接入边界和推荐实施顺序。
 - [DINSAR_PAIRING_DISTRIBUTION_LOGIC_20260508.md](DINSAR_PAIRING_DISTRIBUTION_LOGIC_20260508.md)
   2026-05-08 源码走读记录，梳理 D-InSAR 配对缓存、策略筛选、批次保存、数据分发和生产 worker 执行链路。
+- [DINSAR_SOURCE_BUNDLE_REVERSE_TOOL_TASK_20260511.md](DINSAR_SOURCE_BUNDLE_REVERSE_TOOL_TASK_20260511.md)
+  D-InSAR 去重源数据包的目录协议、续分发规则和外部反向还原工具任务书。
 - [PAIRING_ENHANCEMENT_DESIGN.md](PAIRING_ENHANCEMENT_DESIGN.md)
 - [FRONTEND_NAVIGATION_ARCHITECTURE.md](FRONTEND_NAVIGATION_ARCHITECTURE.md)
 

frontend/src/DataCopierPanel.jsx

@@ -16,8 +16,10 @@ const DataCopierPanel = ({ apiEndpoint, readOnly = false, onJobQueued }) => {
   const [activeTab, setActiveTab] = useState('dinsar');
   const [destDir, setDestDir] = useState('');
   const [copyStatuses, setCopyStatuses] = useState(['COMPLETED']);
-  const [includeDinsarOrbitFiles, setIncludeDinsarOrbitFiles] = useState(false);
+  const [includeDinsarOrbitFiles, setIncludeDinsarOrbitFiles] = useState(true);
-  const [dinsarExportZip, setDinsarExportZip] = useState(false);
+  const [dinsarPackageMode, setDinsarPackageMode] = useState('task_folder');
+  const [skipExistingDinsarTasks, setSkipExistingDinsarTasks] = useState(true);
+  const [dinsarMaxItems, setDinsarMaxItems] = useState('200');
   const [batches, setBatches] = useState([]);
   const [selectedBatchId, setSelectedBatchId] = useState('');
   const [isUploading, setIsUploading] = useState(false);
@@ -84,7 +86,10 @@ const DataCopierPanel = ({ apiEndpoint, readOnly = false, onJobQueued }) => {
   const fetchLogs = async () => {
     if (!taskId) return;
     try {
-      const response = await axios.get(`${apiEndpoint}/tools/copy-status/${taskId}`, { withCredentials: true });
+      const response = await axios.get(`${apiEndpoint}/tools/copy-status/${taskId}`, {
+        withCredentials: true,
+        params: { limit: 1000 },
+      });
       setLogs(response.data.logs);
       const nextStatus = normalizeStatus(response.data.status);
       if (nextStatus && nextStatus !== 'UNKNOWN') {
@@ -126,13 +131,19 @@ const DataCopierPanel = ({ apiEndpoint, readOnly = false, onJobQueued }) => {
       };
       if (activeTab === 'dinsar') {
         payload.include_orbit_files = includeDinsarOrbitFiles;
-        payload.export_zip = dinsarExportZip;
+        payload.package_mode = dinsarPackageMode;
+        payload.export_zip = dinsarPackageMode === 'task_zip';
+        payload.skip_existing = skipExistingDinsarTasks;
+        const parsedMaxItems = Number.parseInt(dinsarMaxItems, 10);
+        if (Number.isFinite(parsedMaxItems) && parsedMaxItems > 0) {
+          payload.max_items = parsedMaxItems;
+        }
       }
       const response = await axios.post(endpoint, payload, { withCredentials: true });
       const taskId = response.data.task_id;
       setTaskId(taskId);
 
-      // 触发全局锁定
+      // 通知全局任务状态；COPY_DATA 在全局控制里按非阻塞处理。
       if (onJobQueued) {
         onJobQueued(taskId);
       }
@@ -196,6 +207,41 @@ const DataCopierPanel = ({ apiEndpoint, readOnly = false, onJobQueued }) => {
             }}
           >
             <label>D-InSAR 分发设置：</label>
+            <div style={{ display: 'grid', gap: '6px', marginTop: '8px' }}>
+              <label style={{ display: 'inline-flex', alignItems: 'center', gap: '6px' }}>
+                <input
+                  type="radio"
+                  name="dinsar-package-mode"
+                  value="task_folder"
+                  checked={dinsarPackageMode === 'task_folder'}
+                  onChange={(event) => setDinsarPackageMode(event.target.value)}
+                  disabled={status === 'RUNNING' || readOnly}
+                />
+                <span>生产 Task 文件夹</span>
+              </label>
+              <label style={{ display: 'inline-flex', alignItems: 'center', gap: '6px' }}>
+                <input
+                  type="radio"
+                  name="dinsar-package-mode"
+                  value="task_zip"
+                  checked={dinsarPackageMode === 'task_zip'}
+                  onChange={(event) => setDinsarPackageMode(event.target.value)}
+                  disabled={status === 'RUNNING' || readOnly}
+                />
+                <span>生产 Task ZIP</span>
+              </label>
+              <label style={{ display: 'inline-flex', alignItems: 'center', gap: '6px' }}>
+                <input
+                  type="radio"
+                  name="dinsar-package-mode"
+                  value="source_bundle"
+                  checked={dinsarPackageMode === 'source_bundle'}
+                  onChange={(event) => setDinsarPackageMode(event.target.value)}
+                  disabled={status === 'RUNNING' || readOnly}
+                />
+                <span>去重源数据包（data / orbit / pairs.json）</span>
+              </label>
+            </div>
             <div style={{ display: 'flex', gap: '16px', flexWrap: 'wrap', marginTop: '8px' }}>
               <label style={{ display: 'inline-flex', alignItems: 'center', gap: '6px' }}>
                 <input
@@ -204,20 +250,32 @@ const DataCopierPanel = ({ apiEndpoint, readOnly = false, onJobQueued }) => {
                   onChange={(event) => setIncludeDinsarOrbitFiles(event.target.checked)}
                   disabled={status === 'RUNNING' || readOnly}
                 />
-                <span>复制精密轨道到 Task/orbit</span>
+                <span>{dinsarPackageMode === 'source_bundle' ? '复制精密轨道到 orbit/' : '复制精密轨道到 Task/orbit'}</span>
               </label>
               <label style={{ display: 'inline-flex', alignItems: 'center', gap: '6px' }}>
                 <input
                   type="checkbox"
-                  checked={dinsarExportZip}
+                  checked={skipExistingDinsarTasks}
-                  onChange={(event) => setDinsarExportZip(event.target.checked)}
+                  onChange={(event) => setSkipExistingDinsarTasks(event.target.checked)}
                   disabled={status === 'RUNNING' || readOnly}
                 />
-                <span>导出为 ZIP 压缩包</span>
+                <span>{dinsarPackageMode === 'source_bundle' ? '复用已存在的 data/orbit' : '跳过目标目录中已存在的 Task'}</span>
               </label>
             </div>
+            <div style={{ display: 'flex', alignItems: 'center', gap: '8px', flexWrap: 'wrap', marginTop: '8px' }}>
+              <label style={{ fontSize: '13px' }}>{dinsarPackageMode === 'source_bundle' ? '每次最多追加新配对:' : '每次最多分发新 Task:'}</label>
+              <input
+                type="number"
+                min="0"
+                value={dinsarMaxItems}
+                onChange={(event) => setDinsarMaxItems(event.target.value)}
+                disabled={status === 'RUNNING' || readOnly}
+                style={{ width: '110px', padding: '5px 7px' }}
+              />
+              <span style={{ fontSize: '12px', color: '#64748b' }}>0 或留空表示不限制</span>
+            </div>
             <div style={{ fontSize: '12px', color: '#475569', marginTop: '6px' }}>
-              未勾选 ZIP 时直接导出 Task 文件夹；勾选后每个 Task 输出一个 .zip。
+              去重源数据包只复制唯一影像和精轨，并写出 pairs.json；再次分发到同一目录时会接着追加未导出的配对。
             </div>
           </div>
         )}

frontend/src/components/app/AppSidePanel.jsx

@@ -262,7 +262,11 @@ export default function AppSidePanel({
                         <LazyDataCopierPanel
                             apiEndpoint={apiEndpoint}
                             readOnly={isReadOnlyUser}
-                            onJobQueued={(taskId) => taskPanel.onTaskStart(taskId, '数据分发任务已入队，正在处理...')}
+                            onJobQueued={(taskId) => taskPanel.onTaskStart(
+                                taskId,
+                                '数据分发任务已入队，正在处理...',
+                                { taskType: 'COPY_DATA', nonBlocking: true },
+                            )}
                         />
                     </Suspense>
                 </div>

frontend/src/hooks/useDinsarOperations.js

@@ -11,7 +11,7 @@ import { normalizePagePayload } from '../utils/appHelpers';
 import { normalizeTaskStatus } from '../utils/appUiHelpers';
 import { DEFAULT_LIST_PAGE_SIZE } from '../config/appConstants';
 
-const NON_BLOCKING_TASK_TYPES = new Set(['UNPACK_ARCHIVES']);
+const NON_BLOCKING_TASK_TYPES = new Set(['UNPACK_ARCHIVES', 'COPY_DATA']);
 
 export default function useDinsarOperations({
     onCleanupDinsarLayers,

frontend/src/hooks/useGlobalTaskControl.js

@@ -2,7 +2,7 @@ import { useCallback, useEffect, useRef, useState } from 'react';
 import apiClient from '../api/client';
 import { normalizeTaskStatus } from '../utils/appUiHelpers';
 
-const NON_BLOCKING_TASK_TYPES = new Set(['UNPACK_ARCHIVES']);
+const NON_BLOCKING_TASK_TYPES = new Set(['UNPACK_ARCHIVES', 'COPY_DATA']);
 
 const isTaskNonBlocking = (taskId, taskType, nonBlockingTaskIds = []) => (
   NON_BLOCKING_TASK_TYPES.has(String(taskType || '').toUpperCase())

nginx/nginx.conf

@@ -18,6 +18,8 @@ http {
 
     sendfile        on;
     keepalive_timeout  65;
+    # Large D-InSAR batch saves can exceed nginx's default 1m request body limit.
+    client_max_body_size 32m;
 
     server {
         listen       80;