add checksum timestamp changes (#371)

* add checksum timestamp changes * style: black reformat * style: DRY * add unit test --------- Co-authored-by: Donny Winston <[email protected]>
microbiomedata · Nov 13, 2023 · fd3aee9 · fd3aee9
1 parent 903f5c3
commit fd3aee9
Show file tree

Hide file tree

Showing 4 changed files with 32 additions and 4 deletions.
diff --git a/nmdc_runtime/api/core/util.py b/nmdc_runtime/api/core/util.py
@@ -28,10 +28,13 @@ def hash_from_str(s: str, algo="sha256") -> str:
     return getattr(hashlib, algo)(s.encode("utf-8")).hexdigest()
 
 
-def sha256hash_from_file(file_path: str):
+def sha256hash_from_file(file_path: str, timestamp: str):
     # https://stackoverflow.com/a/55542529
     h = hashlib.sha256()
 
+    timestamp_bytes = timestamp.encode("utf-8")
+    h.update(timestamp_bytes)
+
     with open(file_path, "rb") as file:
         while True:
             # Reading is buffered, so we can read smaller chunks.

diff --git a/nmdc_runtime/api/endpoints/util.py b/nmdc_runtime/api/endpoints/util.py
@@ -431,17 +431,21 @@ def persist_content_and_get_drs_object(
         filepath = str(Path(save_dir).joinpath(filename))
         with open(filepath, "w") as f:
             f.write(content)
+        now_to_the_minute = datetime.now(tz=ZoneInfo("America/Los_Angeles")).isoformat(
+            timespec="minutes"
+        )
         object_in = DrsObjectIn(
             **drs_metadata_for(
                 filepath,
                 base={
                     "description": (
                         description
                         + f" (created by/for {username}"
-                        + f" at {datetime.now(tz=ZoneInfo('America/Los_Angeles')).isoformat(timespec='minutes')})"
+                        + f" at {now_to_the_minute})"
                     ),
                     "access_methods": [{"access_id": drs_id}],
                 },
+                timestamp=now_to_the_minute,
             )
         )
     self_uri = f"drs://{HOSTNAME_EXTERNAL}/{drs_id}"

diff --git a/nmdc_runtime/util.py b/nmdc_runtime/util.py
@@ -82,7 +82,7 @@ def put_object(filepath, url, mime_type=None):
         return requests.put(url, data=f, headers={"Content-Type": mime_type})
 
 
-def drs_metadata_for(filepath, base=None):
+def drs_metadata_for(filepath, base=None, timestamp=None):
     """given file path, get drs metadata
 
     required: size, created_time, and at least one checksum.
@@ -96,7 +96,7 @@ def drs_metadata_for(filepath, base=None):
         )
     if "checksums" not in base:
         base["checksums"] = [
-            {"type": "sha256", "checksum": sha256hash_from_file(filepath)}
+            {"type": "sha256", "checksum": sha256hash_from_file(filepath, timestamp)}
         ]
     if "mime_type" not in base:
         base["mime_type"] = mimetypes.guess_type(filepath)[0]

diff --git a/tests/unit/core_util.py b/tests/unit/core_util.py
@@ -0,0 +1,21 @@
+from datetime import datetime, timedelta
+from pathlib import Path
+from zoneinfo import ZoneInfo
+
+from nmdc_runtime.api.core.util import sha256hash_from_file
+
+TEST_FILES_DIR = Path(__file__).parent.parent.joinpath("files")
+
+
+def test_sha256hash_from_file_is_timestamp_dependent():
+    file_path = str(TEST_FILES_DIR.joinpath("test_changesheet_update_one_ph.tsv"))
+    ts_1 = datetime.now(tz=ZoneInfo("America/Los_Angeles"))
+    ts_2 = ts_1 + timedelta(minutes=1)
+    hashes = []
+    for ts in (ts_1, ts_2):
+        hashes.append(
+            sha256hash_from_file(
+                file_path=file_path, timestamp=ts.isoformat(timespec="minutes")
+            )
+        )
+    assert hashes[0] != hashes[1]