YDA-5698: add modified date option GL report

Add option to show last modified date of research
and vault collections to the group lifecycle report.

CHANGELOG.md

@@ -2,6 +2,8 @@
 
 ## Unreleased
 
+- Group lifecycle report: add options to show size and last modified date of research
+  and vault groups
 - Fix prompt for password if no .irodsA file is available
 
 ## 2024-05-01 v1.1.2

README.md

@@ -365,19 +365,23 @@ optional arguments:
 ### yreport\_grouplifecycle
 
 ```
-usage: yreport_grouplifecycle [-h] [-q] [-s] [-y {1.7,1.8,1.9,1.10}]
+usage: yreport_grouplifecycle [-h] [-q] [-s] [-m] [-y {1.7,1.8,1.9,1.10}]
 
 Generates a list of research groups, along with their creation date,
 expiration date (if available), lists of group managers, regular members, and
 readonly members. The report also shows whether each research compartment
-contains data, as well as whether its vault compartment contains data.
+contains data, as well as whether its vault compartment contains data. The
+report can optionally include size and last modified date of both the research
+and vault collection.
 
 optional arguments:
   -h, --help            show this help message and exit
   -q, --quasi-xml       Enable Quasi-XML parser in order to be able to parse
                         characters not supported by regular XML parser
   -s, --size            Include size of research collection and vault
                         collection in output
+  -m, --modified        Include last modified date research collection and
+                        vault collection in output
   -y {1.7,1.8,1.9,1.10}, --yoda-version {1.7,1.8,1.9,1.10}
                         Override Yoda version on the server
 ```

yclienttools/common_queries.py

@@ -1,6 +1,7 @@
+import datetime
 from itertools import chain
 import os
-from typing import Dict
+from typing import Dict, Union
 
 from irods.column import Like
 from irods.models import Collection, DataObject, Resource, User, UserGroup
@@ -97,6 +98,45 @@ def get_collection_size(session: iRODSSession,
     return result
 
 
+def get_collection_contents_last_modified(session: iRODSSession, collection_name: str) -> Union[None, datetime.datetime]:
+    """Returns datetime of last modification of collection or its contents
+
+       :param session:         iRODS session
+       :param collection_name: collection name
+
+       :returns: datetime of last modification of the collection or its contents
+                 (data objects or subcollections, recursively), or None if no last
+                 modified datetime could be determined
+    """
+    last_timestamp = None
+
+    dataobjects_root = (session.query(DataObject.modify_time)
+                        .filter(Collection.name == collection_name)
+                        .get_results())
+    dataobjects_sub  = (session.query(DataObject.modify_time)
+                        .filter(Like(Collection.name, collection_name + "/%"))
+                        .get_results())
+    collection_root  = (session.query(Collection.modify_time)
+                        .filter(Collection.name == collection_name)
+                        .get_results())
+    collections_sub  = (session.query(Collection.modify_time)
+                        .filter(Like(Collection.name, collection_name + "/%"))
+                        .get_results())
+
+    all_collection_data = chain(collection_root, collections_sub)
+    all_dataobject_data = chain(dataobjects_root, dataobjects_sub)
+
+    for collection in all_collection_data:
+        if last_timestamp is None or collection[Collection.modify_time] > last_timestamp:
+            last_timestamp = collection[Collection.modify_time]
+
+    for dataobject in all_dataobject_data:
+        if last_timestamp is None or dataobject[DataObject.modify_time] > last_timestamp:
+            last_timestamp = dataobject[DataObject.modify_time]
+
+    return last_timestamp
+
+
 def get_revision_collection_name(session, collection_name):
     '''Returns the revision collection name of a collection if it exists, otherwise None. '''
     expected_prefix = "/{}/home/".format(session.zone)

yclienttools/reportgrouplifecycle.py

@@ -1,6 +1,10 @@
 '''Generates a list of research groups, along with their creation date, expiration date (if available),
    lists of group managers, regular members, and readonly members. The report also shows whether each
-   research compartment contains data, as well as whether its vault compartment contains data.'''
+   research compartment contains data, as well as whether its vault compartment contains data.
+
+   The report can optionally include size and last modified date of both the research and
+   vault collection.
+'''
 
 import argparse
 import csv
@@ -15,7 +19,7 @@
 from irods.models import Collection, DataObject, User
 from irods.session import iRODSSession
 from yclienttools import common_args, common_config, session as s
-from yclienttools.common_queries import collection_exists, get_collection_size
+from yclienttools.common_queries import collection_exists, get_collection_contents_last_modified, get_collection_size
 from yclienttools.options import GroupByOption
 
 
@@ -41,6 +45,8 @@ def _get_args() -> argparse.Namespace:
                         help='Enable Quasi-XML parser in order to be able to parse characters not supported by regular XML parser')
     parser.add_argument("-s", "--size", default=False, action='store_true',
                         help='Include size of research collection and vault collection in output')
+    parser.add_argument("-m", "--modified", default=False, action='store_true',
+                        help='Include last modified date research collection and vault collection in output')
     common_args.add_default_args(parser)
     return parser.parse_args()
 
@@ -182,6 +188,10 @@ def _get_columns(args: argparse.Namespace) -> List[str]:
     else:
         extra_cols = []
 
+    if args.modified:
+        extra_cols.append("Research last modified")
+        extra_cols.append("Vault last modified")
+
     result = base_cols
     result.extend(extra_cols)
     return result
@@ -201,6 +211,10 @@ def _size_to_str(value: Union[int, None]) -> str:
         return humanize.naturalsize(value, binary=True)
 
 
+def _timestamp_to_date_str(value: Union[datetime.datetime, None]) -> str:
+    return "N/A" if value is None else value.strftime("%Y-%m-%d")
+
+
 def report_groups_lifecycle(args: argparse.Namespace, session: iRODSSession):
     output = csv.writer(sys.stdout, delimiter=',')
     output.writerow(_get_columns(args))
@@ -234,4 +248,10 @@ def _has_data_to_string(value):
             rowdata.append(_size_to_str(_get_research_size(session, group)))
             rowdata.append(_size_to_str(_get_vault_size(session, group)))
 
+        if args.modified:
+            rowdata.append(_timestamp_to_date_str(
+                get_collection_contents_last_modified(session, _get_research_group_collection(session, group))))
+            rowdata.append(_timestamp_to_date_str(
+                get_collection_contents_last_modified(session, _get_vault_group_collection(session, group))))
+
         output.writerow(rowdata)