Merge branch 'master' into sarc-371-api

config/sarc-dev.json

@@ -27,7 +27,7 @@
             "duc_inodes_command": null,
             "duc_storage_command": null,
             "diskusage_report_command": "beegfs-ctl --cfgFile=/etc/beegfs/home.d/beegfs-client.conf --getquota --uid $USER --csv",
-            "prometheus_url": "http://monitoring.server.mila.quebec:9090/",
+            "prometheus_url": "http://prometheus01.server.mila.quebec:9090/",
             "start_date": "2022-04-01"
         },
         "narval": {

config/sarc-prod.json

@@ -27,7 +27,7 @@
             "duc_inodes_command": null,
             "duc_storage_command": null,
             "diskusage_report_command": "beegfs-ctl --cfgFile=/etc/beegfs/home.d/beegfs-client.conf --getquota --uid $USER --csv",
-            "prometheus_url": "http://monitoring.server.mila.quebec:9090/",
+            "prometheus_url": "http://prometheus01.server.mila.quebec:9090/",
             "start_date": "2022-04-01"
         },
         "narval": {

sarc/ldap/acquire.py

@@ -42,11 +42,16 @@ def run(
         cache_policy=cache_policy,
     )
 
-    LD_users = fetch_mymila(
-        cfg,
-        LD_users,
-        cache_policy=cache_policy,
-    )
+    # MyMila scraping "NotImplementedError" is temporary ignored until we have a working fetching implementation,
+    # or a working workaround using CSV cache.
+    with using_trace(
+        "sarc.ldap.acquire", "fetch_mymila", exception_types=(NotImplementedError,)
+    ) as span:
+        LD_users = fetch_mymila(
+            cfg,
+            LD_users,
+            cache_policy=cache_policy,
+        )
 
     # For each supervisor or co-supervisor, look for a mila_email_username
     # matching the display name. If None has been found, the previous value remains

sarc/ldap/revision.py

@@ -27,7 +27,15 @@ def is_date_missing(date):
     return date is None or date == DEFAULT_DATE
 
 
-def has_changed(user_db, user_latest, excluded=("_id",)):
+def has_changed(
+    user_db,
+    user_latest,
+    excluded=(
+        "_id",
+        "record_start",
+        "record_end",
+    ),
+):
     keys = set(list(user_db.keys()) + list(user_latest.keys()))
 
     for k in keys:

scripts/systemd/scrapers.sh

@@ -2,4 +2,7 @@
 SCRIPT=$(readlink -f "$0")
 SCRIPTPATH=$(dirname "$SCRIPT")
 cd $SCRIPTPATH/../../
+# scraping jobs
 sudo -u sarc SARC_MODE=scraping SARC_CONFIG=$SCRIPTPATH/../../config/sarc-prod.json ../.local/bin/poetry run sarc acquire jobs -c narval cedar beluga graham mila -d auto
+# scraping users
+sudo -u sarc SARC_MODE=scraping SARC_CONFIG=$SCRIPTPATH/../../config/sarc-prod.json ../.local/bin/poetry run sarc acquire users

tests/functional/cli/acquire/test_acquire_users.py

@@ -124,17 +124,22 @@ def test_acquire_users(cli_main, patch_return_values, mock_file, captrace):
     # as everything goes well without corner cases.
     # We will test logging in test_acquire_users_prompt below.
     spans = captrace.get_finished_spans()
-    assert len(spans) == 1
-    assert spans[0].name == "match_drac_to_mila_accounts"
+    assert len(spans) == 2
+
+    assert spans[0].name == "fetch_mymila"
     assert spans[0].status.status_code == StatusCode.OK
-    assert len(spans[0].events) == 4
+    assert len(spans[0].events) == 0
+
+    assert spans[1].name == "match_drac_to_mila_accounts"
+    assert spans[1].status.status_code == StatusCode.OK
+    assert len(spans[1].events) == 4
     assert (
-        spans[0].events[0].name
+        spans[1].events[0].name
         == "Loading mila_ldap, drac_roles and drac_members from files ..."
     )
-    assert spans[0].events[1].name == "Loading matching config from file ..."
-    assert spans[0].events[2].name == "Matching DRAC/CC to mila accounts ..."
-    assert spans[0].events[3].name == "Committing matches to database ..."
+    assert spans[1].events[1].name == "Loading matching config from file ..."
+    assert spans[1].events[2].name == "Matching DRAC/CC to mila accounts ..."
+    assert spans[1].events[3].name == "Committing matches to database ..."
 
 
 @pytest.mark.parametrize(
@@ -398,15 +403,20 @@ def test_acquire_users_prompt(
 
     # Check traces
     spans = captrace.get_finished_spans()
-    assert len(spans) == 1
-    assert spans[0].name == "match_drac_to_mila_accounts"
+    assert len(spans) == 2
+
+    assert spans[0].name == "fetch_mymila"
     assert spans[0].status.status_code == StatusCode.OK
-    assert len(spans[0].events) == 5
+    assert len(spans[0].events) == 0
+
+    assert spans[1].name == "match_drac_to_mila_accounts"
+    assert spans[1].status.status_code == StatusCode.OK
+    assert len(spans[1].events) == 5
     assert (
-        spans[0].events[0].name
+        spans[1].events[0].name
         == "Loading mila_ldap, drac_roles and drac_members from files ..."
     )
-    assert spans[0].events[1].name == "Loading matching config from file ..."
-    assert spans[0].events[2].name == "Matching DRAC/CC to mila accounts ..."
-    assert spans[0].events[3].name == "Committing matches to database ..."
-    assert spans[0].events[4].name == "Saving 1 manual matches ..."
+    assert spans[1].events[1].name == "Loading matching config from file ..."
+    assert spans[1].events[2].name == "Matching DRAC/CC to mila accounts ..."
+    assert spans[1].events[3].name == "Committing matches to database ..."
+    assert spans[1].events[4].name == "Saving 1 manual matches ..."

tests/functional/ldap/test_acquire_ldap.py

@@ -5,7 +5,7 @@
 
 import sarc.account_matching.make_matches
 import sarc.ldap.acquire
-from sarc.client.ldap.api import get_user
+from sarc.client.ldap.api import get_user, get_users
 
 
 @pytest.mark.usefixtures("empty_read_write_db")
@@ -68,6 +68,75 @@ def test_acquire_ldap(patch_return_values, mock_file):
     assert js_user is None
 
 
+@pytest.mark.usefixtures("empty_read_write_db")
+def test_acquire_ldap_revision_change(patch_return_values, mock_file):
+    """
+    Test two LDAP acquisition, with a change in the LDAP data.
+    This should result in a new record in the database.
+    Then, one third acquisition, with no change in the LDAP data.
+    This should result in no change in the database.
+    """
+    nbr_users = 3
+
+    patch_return_values(
+        {
+            "sarc.ldap.read_mila_ldap.query_ldap": fake_raw_ldap_data(nbr_users),
+            "sarc.ldap.mymila.query_mymila_csv": [],
+        }
+    )
+
+    # Patch the built-in `open()` function for each file path
+    with patch("builtins.open", side_effect=mock_file):
+        sarc.ldap.acquire.run()
+
+    # inspect database to check the number of records
+    # should be nbr_users
+    users = get_users(latest=False)
+    nb_users_1 = len(users)
+    assert nb_users_1 == nbr_users
+
+    # re-acquire the same data
+    with patch("builtins.open", side_effect=mock_file):
+        sarc.ldap.acquire.run()
+
+    # inspect database to check the number of records
+    # should be the same
+    users = get_users(latest=False)
+    assert len(users) == nb_users_1
+
+    # change fake data
+    patch_return_values(
+        {
+            "sarc.ldap.read_mila_ldap.query_ldap": fake_raw_ldap_data(
+                nbr_users,
+                hardcoded_values_by_user={
+                    2: {  # The first user who is not a prof is the one with index 2
+                        "supervisor": "[email protected]"
+                    }
+                },
+            )
+        }
+    )
+
+    # re-acquire the new data
+    with patch("builtins.open", side_effect=mock_file):
+        sarc.ldap.acquire.run()
+
+    # inspect database to check the number of records
+    # should be incremented by 1
+    users = get_users(latest=False)
+    assert len(users) == nb_users_1 + 1
+
+    # re-acquire the same data
+    with patch("builtins.open", side_effect=mock_file):
+        sarc.ldap.acquire.run()
+
+    # inspect database to check the number of records
+    # should be the same
+    users = get_users(latest=False)
+    assert len(users) == nb_users_1 + 1
+
+
 @pytest.mark.usefixtures("empty_read_write_db")
 def test_merge_ldap_and_mymila(patch_return_values, mock_file):
     nbr_users = 10