cronjobs: RMST-459: publish a manifest.json in the bid/cid folder

cronjobs/src/commands/build_compression_dictionaries.py

@@ -5,6 +5,7 @@
 """
 
 import hashlib
+import json
 import os
 import tempfile
 import typing
@@ -232,3 +233,24 @@ def build_compression_dictionaries():
                         content_type=dictpair.mimetype,
                         client=gcs_client,
                     )
+
+    # Publish a `manifest.json` for each impacted collection.
+    # It shows the list of all pairs, giving the list of possible
+    # sources for each target.
+    impacted_cids = {(p.bid, p.cid) for p in missing_dictionaries}
+    all_pairs_of_impacted_cids = [
+        p for p in all_pairs if (p.bid, p.cid) in impacted_cids
+    ]
+    pairs_by_cid_by_target = {}
+    for d in all_pairs_of_impacted_cids:
+        # Strip `bid/cid` out of `bid/cid/yyyymmdd--rid--filename.ext`
+        pairs_by_cid_by_target.setdefault((d.bid, d.cid), {}).setdefault(
+            d.new.split("/")[2], []
+        ).append(d.old.split("/")[2])
+    for (bid, cid), pairs_by_target in pairs_by_cid_by_target.items():
+        manifest_path = f"{DESTINATION_FOLDER}/{bid}/{cid}/manifest.json"
+        storage.Blob(bucket=dicts_bucket, name=manifest_path).upload_from_string(
+            json.dumps(pairs_by_target),
+            content_type="application/json",
+            client=gcs_client,
+        )

cronjobs/tests/commands/test_build_compression_dictionaries.py

@@ -260,6 +260,13 @@ def test_build_compression_dictionaries(
         client=mock.ANY,
     )
 
+    # Assert that a manifest.json file was published.
+    mock_blob.return_value.upload_from_string.assert_called_once_with(
+        """{"20260601--rid1--file.txt": ["20260101--rid1--file.txt"]}""",
+        content_type="application/json",
+        client=mock.ANY,
+    )
+
 
 def test_build_compression_dictionaries_up_to_date(
     mock_kinto_client, mock_fetch_all_changesets, mock_storage_bucket, mock_blob
@@ -285,3 +292,4 @@ def test_build_compression_dictionaries_up_to_date(
 
     mock_blob.return_value.download_to_file.assert_not_called()
     mock_blob.return_value.upload_from_file.assert_not_called()
+    mock_blob.return_value.upload_from_string.assert_not_called()