Merge branch 'data-export' into 'develop'

CLI utility to export archive data

See merge request nomad-lab/nomad-FAIR!1452

nomad/archive/query_reader.py

@@ -713,8 +713,7 @@ def load_archive(self, upload_id: str, entry_id: str) -> ArchiveDict:
             self.upload_pool[upload_id] = upload.upload_files
 
         try:
-            with self.upload_pool[upload_id].read_archive(entry_id) as archive:
-                return archive[entry_id]
+            return self.upload_pool[upload_id].read_archive(entry_id)[entry_id]
         except KeyError:
             raise ArchiveError(f'Archive {entry_id} does not exist in upload {entry_id}.')
 

nomad/archive/required.py

@@ -30,6 +30,7 @@
     Package
 from .query import ArchiveQueryError, to_json, _query_archive_key_pattern, _extract_key_and_index, _extract_child
 from .storage import ArchiveReader, ArchiveList, ArchiveError, ArchiveDict
+from .storage_v2 import ArchiveDict as NewArchiveDict
 from ..datamodel.context import parse_path, ServerContext
 
 
@@ -443,10 +444,13 @@ def _apply_required(
         if archive_item is None:
             return None  # type: ignore
 
-        archive_item = to_json(archive_item)
         result: dict = {}
 
-        if isinstance(archive_item, dict) and 'm_def' in archive_item:
+        # avoid the bug in the old reader that primitive key-value is not included in toc
+        if isinstance(archive_item, ArchiveDict):
+            archive_item = to_json(archive_item)
+
+        if isinstance(archive_item, (dict, NewArchiveDict)) and 'm_def' in archive_item:
             dataset = dataset.replace(definition=self._resolve_definition(
                 dataset.upload_id, archive_item['m_def'].split('@')[0], dataset.archive_root))
             result['m_def'] = archive_item['m_def']
@@ -459,7 +463,7 @@ def _apply_required(
                 return self._resolve_refs(dataset.definition, archive_item, dataset)
 
             if directive in ['*', 'include']:
-                return archive_item
+                return to_json(archive_item)
 
             raise ArchiveQueryError(f'unknown directive {required}')
 

nomad/archive/storage.py

@@ -15,7 +15,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 #
-import os
 from typing import Iterable, Any, Tuple, Dict, BinaryIO, Union, List, cast
 from io import BytesIO, BufferedReader
 from collections.abc import Mapping, Sequence
@@ -26,7 +25,7 @@
 import struct
 import json
 
-from nomad import utils, config
+from nomad import utils
 from nomad.config import archive
 
 __packer = msgpack.Packer(autoreset=True, use_bin_type=True)
@@ -552,11 +551,6 @@ def read_archive(file_or_path: Union[str, BytesIO], **kwargs) -> ArchiveReader:
     '''
     from .storage_v2 import ArchiveWriter as ArchiveWriterNew, ArchiveReader as ArchiveReaderNew
 
-    # if the file is smaller than the threshold, just read it into memory to avoid multiple small reads
-    if isinstance(file_or_path, str) and os.path.getsize(file_or_path) < 4 * config.archive.read_buffer_size:
-        with open(file_or_path, 'rb') as f:
-            file_or_path = BytesIO(f.read())
-
     if isinstance(file_or_path, str):
         with open(file_or_path, 'rb') as f:
             magic = f.read(ArchiveWriterNew.magic_len)

nomad/cli/admin/uploads.py

@@ -268,6 +268,83 @@ def _query_uploads(
     return final_query, proc.Upload.objects(final_query)
 
 
+@uploads.command(help='List selected uploads')
+@click.argument('UPLOADS', nargs=-1)
+@click.option('--required', type=str, help='The required in JSON format')
+@click.option('-o', '--output', type=str, help='The file to write data to')
+@click.pass_context
+def export(ctx, uploads, required, output: str):
+    import sys
+    from nomad.processing import Entry
+    from nomad.utils import get_logger
+    from nomad.files import UploadFiles
+    from nomad.archive import ArchiveQueryError, RequiredReader
+    import time
+    import zipfile
+
+    logger = get_logger(__name__)
+
+    if not output:
+        logger.error('no output given')
+        sys.exit(1)
+
+    if not output.endswith('.zip'):
+        logger.error('only zip output is supported')
+        sys.exit(1)
+
+    output_file = zipfile.ZipFile(output, 'w', allowZip64=True)
+
+    def write(entry_id, archive_data):
+        archive_json = json.dumps(archive_data)
+        output_file.writestr(f'{entry_id}.json', archive_json, compress_type=zipfile.ZIP_DEFLATED)
+
+    _, uploads = _query_uploads(uploads, **ctx.obj.uploads_kwargs)
+
+    try:
+        required_data = json.loads(required)
+    except Exception as e:
+        logger.error('could not parse required', exc_info=e)
+        sys.exit(1)
+
+    try:
+        required_reader = RequiredReader(required_data)
+    except Exception as e:
+        logger.error('could not validate required', exc_info=e)
+        sys.exit(1)
+
+    def get_rss():
+        return time.time()
+
+    start_time = get_rss()
+
+    upload_count = 0
+    total_count = 0
+    for upload in uploads:
+        upload_id = upload.upload_id
+        upload_files = UploadFiles.get(upload_id)
+        upload_count += 1
+        entry_ids = list(entry.entry_id for entry in Entry.objects(upload_id=upload_id))
+        entry_count = 0
+        for entry_id in entry_ids:
+            entry_count += 1
+            total_count += 1
+            try:
+                archive = upload_files.read_archive(entry_id, use_blocked_toc=False)
+                archive_data = required_reader.read(archive, entry_id, upload_id)
+                write(entry_id, archive_data)
+            except ArchiveQueryError as e:
+                logger.error('could not read archive', exc_info=e, entry_id=entry_id)
+            except KeyError as e:
+                logger.error('missing archive', exc_info=e, entry_id=entry_id)
+
+            if total_count % 100 == 0:
+                print(f'{upload_count:5}/{len(uploads)} {entry_count:5}/{len(entry_ids)} {total_count:5} {((get_rss() - start_time))} {upload_id}')
+
+        upload_files.close()
+
+    output_file.close()
+
+
 @uploads.command(help='List selected uploads')
 @click.argument('UPLOADS', nargs=-1)
 @click.option('-e', '--entries', is_flag=True, help='Show details about entries.')