Avoid checking mime type for documents (#155)

* avoid checking mime type in sdk

* update version and changelog

* add test

Author: magnurud

CHANGELOG.md

@@ -1,5 +1,9 @@
 # Changelog
 
+## Version 0.4.2 - 2025-08-14
+
+- Do not discard documents based on their content type
+
 ## Version 0.2.0 - 2025-07-02
 
 - Updated access token generation code to work with new IdP

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "cradl"
-version = "0.4.1"
+version = "0.4.2"
 description = "Python SDK for Cradl"
 authors = [{ name = "Cradl", email = "[email protected]" }]
 readme = "README.md"

src/cradl/client.py

@@ -641,7 +641,12 @@ def create_document(
         }
 
         document = self._make_request(requests.post, '/documents', body=dictstrip(body))
-        self._make_fileserver_request(requests.put, document['fileUrl'], content=content_bytes)
+        try:
+            self._make_fileserver_request(requests.put, document['fileUrl'], content=content_bytes)
+        except Exception as e:
+            self.delete_document(document['documentId'])
+            raise e
+
         return document
 
     def list_documents(

src/cradl/content.py

@@ -8,8 +8,6 @@
 
 def _guess_content_type(raw):
     guessed_type = filetype.guess(raw)
-    assert guessed_type, 'Could not determine content type of document. ' \
-                         'Please provide it by specifying content_type'
     return guessed_type.mime
 
 

tests/conftest.py

@@ -56,3 +56,8 @@ def content():
     Yields a random JPEG bytestring with a length 2E4
     """
     yield b'\xFF\xD8\xFF\xEE' + urandom(int(2E4))
+
+
+@pytest.fixture(scope='session')
+def pdf_content():
+    yield 'HTTP/1.0 200 OK%PDF-1.4 %½¾¼µ %%EOF'.encode()

tests/test_documents.py

@@ -6,16 +6,21 @@
 pytestmark = pytest.mark.integration
 
 
-@pytest.mark.parametrize('metadata', [util.metadata(), None])
+@pytest.mark.parametrize(('metadata', 'image'), [
+    (util.metadata(), True),
+    (None, False),
+])
 def test_create_document(
     monkeypatch,
     static_client: Client,
     content,
+    image,
     metadata,
+    pdf_content,
 ):
     consent_id = service.create_consent_id()
     post_documents_response = static_client.create_document(
-        content,
+        content if image else pdf_content,
         consent_id=consent_id,
         metadata=metadata,
     )