From 4c06351cb356bbaf8d9a2fb2bc6e5b255efe6ce2 Mon Sep 17 00:00:00 2001
From: msramalho <19508417+msramalho@users.noreply.github.com>
Date: Thu, 20 Feb 2025 15:14:13 +0000
Subject: [PATCH] api token archive can set the author_id explicitly

---
 app/tests/web/endpoints/test_url.py | 11 ++++++++++-
 app/web/config.py                   |  2 +-
 app/web/endpoints/url.py            |  3 ++-
 3 files changed, 13 insertions(+), 3 deletions(-)

diff --git a/app/tests/web/endpoints/test_url.py b/app/tests/web/endpoints/test_url.py
index 4a6f342..1b6ee85 100644
--- a/app/tests/web/endpoints/test_url.py
+++ b/app/tests/web/endpoints/test_url.py
@@ -2,6 +2,7 @@ import json
 from unittest.mock import MagicMock, patch
 
 from app.shared.schemas import ArchiveCreate, TaskResult
+from app.web.config import ALLOW_ANY_EMAIL
 
 
 def test_archive_url_unauthenticated(client, test_no_auth):
@@ -107,13 +108,21 @@ def test_archive_url_with_api_token(m_celery, client_with_token):
     m_signature = MagicMock()
     m_signature.apply_async.return_value = TaskResult(id="123-456-789", status="PENDING", result="")
     m_celery.signature.return_value = m_signature
-    response = client_with_token.post("/url/archive", json={"url": "https://example.com"})
+    response = client_with_token.post("/url/archive", json={"url": "https://example.com", "author_id": "someone@example.com"})
     assert response.status_code == 201
     assert response.json() == {'id': '123-456-789'}
     m_celery.signature.assert_called_once()
     m_signature.apply_async.assert_called_once()
     called_val = m_celery.signature.call_args
     assert called_val[0][0] == "create_archive_task"
+    assert json.loads(called_val[1]['args'][0]) ==  {"id": None, "url": "https://example.com", "result": None, "public": False, "author_id": "someone@example.com", "group_id": "default", "tags": None, "sheet_id": None, "store_until": None, "urls": None}
+
+    # missing id should use ALLOW_ANY_EMAIL
+    response = client_with_token.post("/url/archive", json={"url": "https://example.com", "author_id": None})
+    assert response.status_code == 201
+    called_val = m_celery.signature.call_args
+    assert called_val[0][0] == "create_archive_task"
+    assert json.loads(called_val[1]['args'][0]) ==  {"id": None, "url": "https://example.com", "result": None, "public": False, "author_id": ALLOW_ANY_EMAIL, "group_id": "default", "tags": None, "sheet_id": None, "store_until": None, "urls": None}
 
 
 def test_search_by_url_unauthenticated(client, test_no_auth):
diff --git a/app/web/config.py b/app/web/config.py
index 5b2a80e..29a6806 100644
--- a/app/web/config.py
+++ b/app/web/config.py
@@ -1,4 +1,4 @@
-VERSION = "0.9.1"
+VERSION = "0.9.2"
 
 API_DESCRIPTION = """
 #### API for the Auto-Archiver project, a tool to archive web pages and Google Sheets.
diff --git a/app/web/endpoints/url.py b/app/web/endpoints/url.py
index c237893..a7ac4b4 100644
--- a/app/web/endpoints/url.py
+++ b/app/web/endpoints/url.py
@@ -27,7 +27,6 @@ def archive_url(
     email=Depends(get_token_or_user_auth),
     db: Session = Depends(get_db_dependency)
 ) -> schemas.Task:
-    archive.author_id = email
     logger.info(f"new {archive.public=} task for {email=} and {archive.group_id=}: {archive.url}")
 
     parsed_url = urlparse(archive.url)
@@ -36,6 +35,7 @@ def archive_url(
 
     archive_create = schemas.ArchiveCreate(**archive.model_dump())
     if email != ALLOW_ANY_EMAIL:
+        archive_create.author_id = email
         user = UserState(db, email)
         if archive.group_id and not user.in_group(archive.group_id):
             raise HTTPException(status_code=403, detail="User does not have access to this group.")
@@ -45,6 +45,7 @@ def archive_url(
             raise HTTPException(status_code=429, detail="User has reached their monthly MB quota.")
         group_queue = user.priority_group(archive_create.group_id)
     else:
+        archive_create.author_id = archive.author_id or email
         group_queue = convert_priority_to_queue_dict("high")