diff --git a/swh/deposit/api/common.py b/swh/deposit/api/common.py index 733996c5..dc49772c 100644 --- a/swh/deposit/api/common.py +++ b/swh/deposit/api/common.py @@ -1,826 +1,828 @@ # Copyright (C) 2017 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import hashlib from abc import ABCMeta, abstractmethod from django.core.urlresolvers import reverse from django.http import HttpResponse from django.shortcuts import render from django.utils import timezone from rest_framework import status from rest_framework.views import APIView from rest_framework.permissions import IsAuthenticated, AllowAny from ..config import SWHDefaultConfig, EDIT_SE_IRI, EM_IRI, CONT_FILE_IRI from ..config import ARCHIVE_KEY, METADATA_KEY from ..models import Deposit, DepositRequest, DepositCollection from ..models import DepositRequestType, DepositClient from ..parsers import parse_xml from ..errors import MAX_UPLOAD_SIZE_EXCEEDED, BAD_REQUEST, ERROR_CONTENT from ..errors import CHECKSUM_MISMATCH, make_error_dict, MEDIATION_NOT_ALLOWED from ..errors import make_error_response_from_dict, FORBIDDEN from ..errors import NOT_FOUND, make_error_response, METHOD_NOT_ALLOWED ACCEPT_PACKAGINGS = ['http://purl.org/net/sword/package/SimpleZip'] ACCEPT_CONTENT_TYPES = ['application/zip'] class SWHAPIView(APIView): """Mixin intended as a based API view to enforce the basic authentication check """ permission_classes = (IsAuthenticated, ) class SWHPrivateAPIView(SWHAPIView): """Mixin intended as private api (so no authentication) based API view (for the private ones). """ permission_classes = (AllowAny, ) class SWHBaseDeposit(SWHDefaultConfig, SWHAPIView, metaclass=ABCMeta): """Base deposit request class sharing multiple common behaviors. """ def __init__(self): super().__init__() deposit_request_types = DepositRequestType.objects.all() self.deposit_request_types = { type.name: type for type in deposit_request_types } def _read_headers(self, req): """Read and unify the necessary headers from the request (those are not stored in the same location or not properly formatted). Args: req (Request): Input request Returns: Dictionary with the following keys (some associated values may be None): - content-type - content-length - in-progress - content-disposition - packaging - slug - on-behalf-of """ meta = req._request.META content_type = req.content_type content_length = meta.get('CONTENT_LENGTH') if content_length and isinstance(content_length, str): content_length = int(content_length) # final deposit if not provided in_progress = meta.get('HTTP_IN_PROGRESS', False) content_disposition = meta.get('HTTP_CONTENT_DISPOSITION') if isinstance(in_progress, str): in_progress = in_progress.lower() == 'true' content_md5sum = meta.get('HTTP_CONTENT_MD5') if content_md5sum: content_md5sum = bytes.fromhex(content_md5sum) packaging = meta.get('HTTP_PACKAGING') slug = meta.get('HTTP_SLUG') on_behalf_of = meta.get('HTTP_ON_BEHALF_OF') metadata_relevant = meta.get('HTTP_METADATA_RELEVANT') return { 'content-type': content_type, 'content-length': content_length, 'in-progress': in_progress, 'content-disposition': content_disposition, 'content-md5sum': content_md5sum, 'packaging': packaging, 'slug': slug, 'on-behalf-of': on_behalf_of, 'metadata-relevant': metadata_relevant, } def _compute_md5(self, filehandler): """Compute uploaded file's md5 sum. Args: filehandler (InMemoryUploadedFile): the file to compute the md5 hash Returns: the md5 checksum (str) """ h = hashlib.md5() for chunk in filehandler: h.update(chunk) return h.digest() def _deposit_put(self, deposit_id=None, in_progress=False, external_id=None): """Save/Update a deposit in db. Args: deposit_id (int): deposit identifier in_progress (dict): The deposit's status external_id (str): The external identifier to associate to the deposit Returns: The Deposit instance saved or updated. """ if in_progress is False: complete_date = timezone.now() status_type = 'ready' else: complete_date = None status_type = 'partial' if not deposit_id: deposit = Deposit(collection=self._collection, external_id=external_id, complete_date=complete_date, status=status_type, client=self._client) else: deposit = Deposit.objects.get(pk=deposit_id) # update metadata deposit.complete_date = complete_date deposit.status = status_type deposit.save() return deposit def _deposit_request_put(self, deposit, deposit_request_data, replace_metadata=False, replace_archives=False): """Save a deposit request with metadata attached to a deposit. Args: deposit (Deposit): The deposit concerned by the request deposit_request_data (dict): The dictionary with at most 2 deposit request types (archive, metadata) to associate to the deposit replace_metadata (bool): Flag defining if we add or update existing metadata to the deposit replace_archives (bool): Flag defining if we add or update archives to existing deposit Returns: None """ if replace_metadata: DepositRequest.objects.filter( deposit=deposit, type=self.deposit_request_types[METADATA_KEY]).delete() if replace_archives: DepositRequest.objects.filter( deposit=deposit, type=self.deposit_request_types[ARCHIVE_KEY]).delete() deposit_request = None archive_file = deposit_request_data.get(ARCHIVE_KEY) if archive_file: deposit_request = DepositRequest( type=self.deposit_request_types[ARCHIVE_KEY], deposit=deposit, archive=archive_file) deposit_request.save() metadata = deposit_request_data.get(METADATA_KEY) if metadata: deposit_request = DepositRequest( type=self.deposit_request_types[METADATA_KEY], deposit=deposit, metadata=metadata) deposit_request.save() assert deposit_request is not None def _delete_archives(self, collection_name, deposit_id): """Delete archives reference from the deposit id. """ try: deposit = Deposit.objects.get(pk=deposit_id) except Deposit.DoesNotExist: return make_error_dict( NOT_FOUND, 'The deposit %s does not exist' % deposit_id) DepositRequest.objects.filter( deposit=deposit, type=self.deposit_request_types[ARCHIVE_KEY]).delete() return {} def _delete_deposit(self, collection_name, deposit_id): """Delete deposit reference. Args: collection_name (str): Client's name deposit_id (id): The deposit to delete Returns Empty dict when ok. Dict with error key to describe the failure. """ try: deposit = Deposit.objects.get(pk=deposit_id) except Deposit.DoesNotExist: return make_error_dict( NOT_FOUND, 'The deposit %s does not exist' % deposit_id) if deposit.collection.name != collection_name: summary = 'Cannot delete a deposit from another collection' description = "Deposit %s does not belong to the collection %s" % ( deposit_id, collection_name) return make_error_dict( BAD_REQUEST, summary=summary, verbose_description=description) DepositRequest.objects.filter(deposit=deposit).delete() deposit.delete() return {} def _check_preconditions_on(self, filehandler, md5sum, content_length=None): """Check preconditions on provided file are respected. That is the length and/or the md5sum hash match the file's content. Args: filehandler (InMemoryUploadedFile): The file to check md5sum (hex str): md5 hash expected from the file's content content_length (int): the expected length if provided. Returns: Either none if no error or a dictionary with a key error detailing the problem. """ if content_length: if content_length > self.config['max_upload_size']: return make_error_dict( MAX_UPLOAD_SIZE_EXCEEDED, 'Upload size limit exceeded (max %s bytes).' % self.config['max_upload_size'], 'Please consider sending the archive in ' 'multiple steps.') length = filehandler.size if length != content_length: return make_error_dict(status.HTTP_412_PRECONDITION_FAILED, 'Wrong length') if md5sum: _md5sum = self._compute_md5(filehandler) if _md5sum != md5sum: return make_error_dict( CHECKSUM_MISMATCH, 'Wrong md5 hash', 'The checksum sent %s and the actual checksum ' '%s does not match.' % (md5sum, _md5sum)) return None def _binary_upload(self, req, headers, collection_name, deposit_id=None, replace_metadata=False, replace_archives=False): """Binary upload routine. Other than such a request, a 415 response is returned. Args: req (Request): the request holding information to parse and inject in db headers (dict): request headers formatted collection_name (str): the associated client deposit_id (id): deposit identifier if provided replace_metadata (bool): 'Update or add' request to existing deposit. If False (default), this adds new metadata request to existing ones. Otherwise, this will replace existing metadata. replace_archives (bool): 'Update or add' request to existing deposit. If False (default), this adds new archive request to existing ones. Otherwise, this will replace existing archives. ones. Returns: In the optimal case a dict with the following keys: - deposit_id (int): Deposit identifier - deposit_date (date): Deposit date - archive: None (no archive is provided here) Otherwise, a dictionary with the key error and the associated failures, either: - 400 (bad request) if the request is not providing an external identifier - 403 (forbidden) if the length of the archive exceeds the max size configured - 412 (precondition failed) if the length or md5 hash provided mismatch the reality of the archive - 415 (unsupported media type) if a wrong media type is provided """ content_length = headers['content-length'] if not content_length: return make_error_dict( BAD_REQUEST, 'CONTENT_LENGTH header is mandatory', 'For archive deposit, the ' 'CONTENT_LENGTH header must be sent.') content_disposition = headers['content-disposition'] if not content_disposition: return make_error_dict( BAD_REQUEST, 'CONTENT_DISPOSITION header is mandatory', 'For archive deposit, the ' 'CONTENT_DISPOSITION header must be sent.') packaging = headers['packaging'] if packaging and packaging not in ACCEPT_PACKAGINGS: return make_error_dict( BAD_REQUEST, 'Only packaging %s is supported' % ACCEPT_PACKAGINGS, 'The packaging provided %s is not supported' % packaging) filehandler = req.FILES['file'] precondition_status_response = self._check_preconditions_on( filehandler, headers['content-md5sum'], content_length) if precondition_status_response: return precondition_status_response external_id = headers['slug'] # actual storage of data archive_metadata = filehandler deposit = self._deposit_put(deposit_id=deposit_id, in_progress=headers['in-progress'], external_id=external_id) self._deposit_request_put( deposit, {ARCHIVE_KEY: archive_metadata}, replace_metadata=replace_metadata, replace_archives=replace_archives) return { 'deposit_id': deposit.id, 'deposit_date': deposit.reception_date, 'archive': filehandler.name, } def _multipart_upload(self, req, headers, collection_name, deposit_id=None, replace_metadata=False, replace_archives=False): """Multipart upload supported with exactly: - 1 archive (zip) - 1 atom entry Other than such a request, a 415 response is returned. Args: req (Request): the request holding information to parse and inject in db headers (dict): request headers formatted collection_name (str): the associated client deposit_id (id): deposit identifier if provided replace_metadata (bool): 'Update or add' request to existing deposit. If False (default), this adds new metadata request to existing ones. Otherwise, this will replace existing metadata. replace_archives (bool): 'Update or add' request to existing deposit. If False (default), this adds new archive request to existing ones. Otherwise, this will replace existing archives. ones. Returns: In the optimal case a dict with the following keys: - deposit_id (int): Deposit identifier - deposit_date (date): Deposit date - archive: None (no archive is provided here) Otherwise, a dictionary with the key error and the associated failures, either: - 400 (bad request) if the request is not providing an external identifier - 412 (precondition failed) if the potentially md5 hash provided mismatch the reality of the archive - 415 (unsupported media type) if a wrong media type is provided """ external_id = headers['slug'] content_types_present = set() data = { 'application/zip': None, # expected archive 'application/atom+xml': None, } for key, value in req.FILES.items(): fh = value if fh.content_type in content_types_present: return make_error_dict( ERROR_CONTENT, 'Only 1 application/zip archive and 1 ' 'atom+xml entry is supported (as per sword2.0 ' 'specification)', 'You provided more than 1 application/zip ' 'or more than 1 application/atom+xml content-disposition ' 'header in the multipart deposit') content_types_present.add(fh.content_type) data[fh.content_type] = fh if len(content_types_present) != 2: return make_error_dict( ERROR_CONTENT, 'You must provide both 1 application/zip ' 'and 1 atom+xml entry for multipart deposit', 'You need to provide only 1 application/zip ' 'and 1 application/atom+xml content-disposition header ' 'in the multipart deposit') filehandler = data['application/zip'] precondition_status_response = self._check_preconditions_on( filehandler, headers['content-md5sum']) if precondition_status_response: return precondition_status_response # actual storage of data atom_metadata = parse_xml(data['application/atom+xml']) deposit = self._deposit_put(deposit_id=deposit_id, in_progress=headers['in-progress'], external_id=external_id) deposit_request_data = { ARCHIVE_KEY: filehandler, METADATA_KEY: atom_metadata, } self._deposit_request_put( deposit, deposit_request_data, replace_metadata, replace_archives) return { 'deposit_id': deposit.id, 'deposit_date': deposit.reception_date, 'archive': filehandler.name, } def _atom_entry(self, req, headers, collection_name, deposit_id=None, replace_metadata=False, replace_archives=False): """Atom entry deposit. Args: req (Request): the request holding information to parse and inject in db headers (dict): request headers formatted collection_name (str): the associated client deposit_id (id): deposit identifier if provided replace_metadata (bool): 'Update or add' request to existing deposit. If False (default), this adds new metadata request to existing ones. Otherwise, this will replace existing metadata. replace_archives (bool): 'Update or add' request to existing deposit. If False (default), this adds new archive request to existing ones. Otherwise, this will replace existing archives. ones. Returns: In the optimal case a dict with the following keys: - deposit_id: deposit id associated to the deposit - deposit_date: date of the deposit - archive: None (no archive is provided here) Otherwise, a dictionary with the key error and the associated failures, either: - 400 (bad request) if the request is not providing an external identifier - 400 (bad request) if the request's body is empty - 415 (unsupported media type) if a wrong media type is provided """ if not req.data: return make_error_dict( BAD_REQUEST, 'Empty body request is not supported', 'Atom entry deposit is supposed to send for metadata. ' 'If the body is empty, there is no metadata.') external_id = req.data.get( '{http://www.w3.org/2005/Atom}external_identifier', headers['slug']) deposit = self._deposit_put(deposit_id=deposit_id, in_progress=headers['in-progress'], external_id=external_id) self._deposit_request_put( deposit, {METADATA_KEY: req.data}, replace_metadata, replace_archives) return { 'deposit_id': deposit.id, 'deposit_date': deposit.reception_date, 'archive': None, } def _empty_post(self, req, headers, collection_name, deposit_id): """Empty post to finalize an empty deposit. Args: req (Request): the request holding information to parse and inject in db headers (dict): request headers formatted collection_name (str): the associated client deposit_id (id): deposit identifier Returns: Dictionary of result with the deposit's id, the date it was completed and no archive. """ deposit = Deposit.objects.get(pk=deposit_id) deposit.complete_date = timezone.now() deposit.status = 'ready' deposit.save() return { 'deposit_id': deposit_id, 'deposit_date': deposit.complete_date, 'archive': None, } def _make_iris(self, collection_name, deposit_id): """Define the IRI endpoints Args: collection_name (str): client/collection's name deposit_id (id): Deposit identifier Returns: Dictionary of keys with the iris' urls. """ return { EM_IRI: reverse( EM_IRI, args=[collection_name, deposit_id]), EDIT_SE_IRI: reverse( EDIT_SE_IRI, args=[collection_name, deposit_id]), CONT_FILE_IRI: reverse( CONT_FILE_IRI, args=[collection_name, deposit_id]), } - def additional_checks(self, req, collection_name, deposit_id=None): - """Permit the child class to enrich with additional checks. + def additional_checks(self, req, headers, collection_name, + deposit_id=None): + """Permit the child class to enrich additional checks. Returns: dict with 'error' detailing the problem. """ return {} def checks(self, req, collection_name, deposit_id=None): try: self._collection = DepositCollection.objects.get( name=collection_name) except DepositCollection.DoesNotExist: return make_error_dict( NOT_FOUND, 'Unknown collection name %s' % collection_name) username = req.user.username if username: # unauthenticated request can have the username empty try: self._client = DepositClient.objects.get(username=username) except DepositClient.DoesNotExist: return make_error_dict(NOT_FOUND, 'Unknown client name %s' % username) if self._collection.id not in self._client.collections: return make_error_dict( FORBIDDEN, 'Client %s cannot access collection %s' % ( username, collection_name)) if deposit_id: try: deposit = Deposit.objects.get(pk=deposit_id) except Deposit.DoesNotExist: return make_error_dict( NOT_FOUND, 'Deposit with id %s does not exist' % deposit_id) checks = self.restrict_access(req, deposit) if checks: return checks headers = self._read_headers(req) if headers['on-behalf-of']: return make_error_dict(MEDIATION_NOT_ALLOWED, 'Mediation is not supported.') - checks = self.additional_checks(req, collection_name, deposit_id) + checks = self.additional_checks(req, headers, + collection_name, deposit_id) if 'error' in checks: return checks return {'headers': headers} def restrict_access(self, req, deposit=None): if deposit: if req.method != 'GET' and deposit.status != 'partial': summary = "You can only act on deposit with status 'partial'" description = "This deposit has status '%s'" % deposit.status return make_error_dict( BAD_REQUEST, summary=summary, verbose_description=description) def get(self, req, *args, **kwargs): return make_error_response(req, METHOD_NOT_ALLOWED) def post(self, req, *args, **kwargs): return make_error_response(req, METHOD_NOT_ALLOWED) def put(self, req, *args, **kwargs): return make_error_response(req, METHOD_NOT_ALLOWED) def delete(self, req, *args, **kwargs): return make_error_response(req, METHOD_NOT_ALLOWED) class SWHGetDepositAPI(SWHBaseDeposit, metaclass=ABCMeta): """Mixin for class to support GET method. """ def get(self, req, collection_name, deposit_id, format=None): """Endpoint to create/add resources to deposit. Returns: 200 response when no error during routine occurred 400 if the deposit does not belong to the collection 404 if the deposit or the collection does not exist """ checks = self.checks(req, collection_name, deposit_id) if 'error' in checks: return make_error_response_from_dict(req, checks['error']) status, content, content_type = self.process_get( req, collection_name, deposit_id) return HttpResponse(content, status=status, content_type=content_type) @abstractmethod def process_get(self, req, collection_name, deposit_id): """Routine to deal with the deposit's get processing. Returns: Tuple status, stream of content, content-type """ pass class SWHPostDepositAPI(SWHBaseDeposit, metaclass=ABCMeta): """Mixin for class to support DELETE method. """ def post(self, req, collection_name, deposit_id=None, format=None): """Endpoint to create/add resources to deposit. Returns: 204 response when no error during routine occurred. 400 if the deposit does not belong to the collection 404 if the deposit or the collection does not exist """ checks = self.checks(req, collection_name, deposit_id) if 'error' in checks: return make_error_response_from_dict(req, checks['error']) headers = checks['headers'] _status, _iri_key, data = self.process_post( req, headers, collection_name, deposit_id) error = data.get('error') if error: return make_error_response_from_dict(req, error) data['packagings'] = ACCEPT_PACKAGINGS iris = self._make_iris(collection_name, data['deposit_id']) data.update(iris) response = render(req, 'deposit/deposit_receipt.xml', context=data, content_type='application/xml', status=_status) response._headers['location'] = 'Location', data[_iri_key] return response @abstractmethod def process_post(self, req, headers, collection_name, deposit_id=None): """Routine to deal with the deposit's processing. Returns Tuple of: - response status code (200, 201, etc...) - key iri (EM_IRI, EDIT_SE_IRI, etc...) - dictionary of the processing result """ pass class SWHPutDepositAPI(SWHBaseDeposit, metaclass=ABCMeta): """Mixin for class to support PUT method. """ def put(self, req, collection_name, deposit_id, format=None): """Endpoint to update deposit resources. Returns: 204 response when no error during routine occurred. 400 if the deposit does not belong to the collection 404 if the deposit or the collection does not exist """ checks = self.checks(req, collection_name, deposit_id) if 'error' in checks: return make_error_response_from_dict(req, checks['error']) headers = checks['headers'] data = self.process_put(req, headers, collection_name, deposit_id) error = data.get('error') if error: return make_error_response_from_dict(req, error) return HttpResponse(status=status.HTTP_204_NO_CONTENT) @abstractmethod def process_put(self, req, headers, collection_name, deposit_id): """Routine to deal with updating a deposit in some way. Returns dictionary of the processing result """ pass class SWHDeleteDepositAPI(SWHBaseDeposit, metaclass=ABCMeta): """Mixin for class to support DELETE method. """ def delete(self, req, collection_name, deposit_id): """Endpoint to delete some deposit's resources (archives, deposit). Returns: 204 response when no error during routine occurred. 400 if the deposit does not belong to the collection 404 if the deposit or the collection does not exist """ checks = self.checks(req, collection_name, deposit_id) if 'error' in checks: return make_error_response_from_dict(req, checks['error']) data = self.process_delete(req, collection_name, deposit_id) error = data.get('error') if error: return make_error_response_from_dict(req, error) return HttpResponse(status=status.HTTP_204_NO_CONTENT) @abstractmethod def process_delete(self, req, collection_name, deposit_id): """Routine to delete a resource. This is mostly not allowed except for the EM_IRI (cf. .api.deposit_update.SWHUpdateArchiveDeposit) """ pass diff --git a/swh/deposit/api/deposit.py b/swh/deposit/api/deposit.py index a307119d..77f73b9c 100644 --- a/swh/deposit/api/deposit.py +++ b/swh/deposit/api/deposit.py @@ -1,80 +1,91 @@ # Copyright (C) 2017 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from rest_framework import status from .common import SWHPostDepositAPI from ..config import EDIT_SE_IRI +from ..errors import make_error_dict, BAD_REQUEST from ..parsers import SWHFileUploadParser, SWHAtomEntryParser from ..parsers import SWHMultiPartParser class SWHDeposit(SWHPostDepositAPI): """Deposit request class defining api endpoints for sword deposit. What's known as 'Col IRI' in the sword specification. HTTP verbs supported: POST """ parser_classes = (SWHMultiPartParser, SWHFileUploadParser, SWHAtomEntryParser) + def additional_checks(self, req, headers, collection_name, + deposit_id=None): + slug = headers['slug'] + if not slug: + msg = 'Missing SLUG header in request' + verbose_description = 'Provide in the SLUG header one identifier, for example the url pointing to the resource you are depositing.' # noqa + return make_error_dict(BAD_REQUEST, msg, verbose_description) + + return {} + def process_post(self, req, headers, collection_name, deposit_id=None): """Create a first deposit as: - archive deposit (1 zip) - multipart (1 zip + 1 atom entry) - atom entry Args: req (Request): the request holding the information to parse and inject in db collection_name (str): the associated client Returns: An http response (HttpResponse) according to the situation. If everything is ok, a 201 response (created) with a deposit receipt. Otherwise, depending on the upload, the following errors can be returned: - archive deposit: - 400 (bad request) if the request is not providing an external identifier - 403 (forbidden) if the length of the archive exceeds the max size configured - 412 (precondition failed) if the length or hash provided mismatch the reality of the archive. - 415 (unsupported media type) if a wrong media type is provided - multipart deposit: - 400 (bad request) if the request is not providing an external identifier - 412 (precondition failed) if the potentially md5 hash provided mismatch the reality of the archive - 415 (unsupported media type) if a wrong media type is provided - Atom entry deposit: - 400 (bad request) if the request is not providing an external identifier - 400 (bad request) if the request's body is empty - 415 (unsupported media type) if a wrong media type is provided """ assert deposit_id is None if req.content_type == 'application/zip': data = self._binary_upload(req, headers, collection_name) elif req.content_type.startswith('multipart/'): data = self._multipart_upload(req, headers, collection_name) else: data = self._atom_entry(req, headers, collection_name) return status.HTTP_201_CREATED, EDIT_SE_IRI, data diff --git a/swh/deposit/api/private/deposit_update_status.py b/swh/deposit/api/private/deposit_update_status.py index 7fa9ada1..9f432ed0 100644 --- a/swh/deposit/api/private/deposit_update_status.py +++ b/swh/deposit/api/private/deposit_update_status.py @@ -1,70 +1,71 @@ # Copyright (C) 2017 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from rest_framework.parsers import JSONParser from ..common import SWHPutDepositAPI, SWHPrivateAPIView from ...errors import make_error_dict, BAD_REQUEST from ...models import Deposit, DEPOSIT_STATUS_DETAIL, format_swh_id class SWHUpdateStatusDeposit(SWHPutDepositAPI, SWHPrivateAPIView): """Deposit request class to update the deposit's status. HTTP verbs supported: PUT """ parser_classes = (JSONParser, ) - def additional_checks(self, req, collection_name, deposit_id=None): + def additional_checks(self, req, headers, collection_name, + deposit_id=None): """Enrich existing checks to the default ones. New checks: - Ensure the status is provided - Ensure it exists """ data = req.data status = data.get('status') if not status: msg = 'The status key is mandatory with possible values %s' % list( DEPOSIT_STATUS_DETAIL.keys()) return make_error_dict(BAD_REQUEST, msg) if status not in DEPOSIT_STATUS_DETAIL: msg = 'Possible status in %s' % list(DEPOSIT_STATUS_DETAIL.keys()) return make_error_dict(BAD_REQUEST, msg) if status == 'success': swh_id = data.get('revision_id') if not swh_id: msg = 'Updating status to %s requires a revision_id key' % ( status, ) return make_error_dict(BAD_REQUEST, msg) return {} def restrict_access(self, req, deposit=None): """Remove restriction modification to 'partial' deposit. Update is possible regardless of the existing status. """ return None def process_put(self, req, headers, collection_name, deposit_id): """Update the deposit's status Returns: 204 No content """ deposit = Deposit.objects.get(pk=deposit_id) deposit.status = req.data['status'] # checks already done before swh_id = req.data.get('revision_id') if swh_id: deposit.swh_id = format_swh_id(collection_name, swh_id) deposit.save() return {} diff --git a/swh/deposit/tests/api/test_deposit_atom.py b/swh/deposit/tests/api/test_deposit_atom.py index c82e38b6..2525d37c 100644 --- a/swh/deposit/tests/api/test_deposit_atom.py +++ b/swh/deposit/tests/api/test_deposit_atom.py @@ -1,258 +1,273 @@ # Copyright (C) 2017 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from django.core.urlresolvers import reverse from io import BytesIO from rest_framework import status from rest_framework.test import APITestCase from swh.deposit.config import COL_IRI from swh.deposit.models import Deposit, DepositRequest from swh.deposit.parsers import parse_xml from ..common import BasicTestCase, WithAuthTestCase class DepositAtomEntryTestCase(APITestCase, WithAuthTestCase, BasicTestCase): """Try and post atom entry deposit. """ def setUp(self): super().setUp() self.atom_entry_data0 = b""" Awesome Compiler hal urn:uuid:1225c695-cfb8-4ebb-aaaa-80da344efa6a %s 2017-10-07T15:17:08Z some awesome author something awesome-compiler This is an awesome compiler destined to awesomely compile stuff and other stuff compiler,programming,language 2005-10-07T17:17:08Z 2005-10-07T17:17:08Z release note related link Awesome https://hoster.org/awesome-compiler GNU/Linux 0.0.1 running all """ self.atom_entry_data1 = b""" hal urn:uuid:2225c695-cfb8-4ebb-aaaa-80da344efa6a 2017-10-07T15:17:08Z some awesome author something awesome-compiler This is an awesome compiler destined to awesomely compile stuff and other stuff compiler,programming,language 2005-10-07T17:17:08Z 2005-10-07T17:17:08Z release note related link Awesome https://hoster.org/awesome-compiler GNU/Linux 0.0.1 running all """ self.atom_entry_data2 = b""" %s """ self.atom_entry_data_empty_body = b""" """ self.atom_entry_data3 = b""" something """ self.atom_entry_tei = b"""HAL TEI export of hal-01587083CCSDDistributed under a Creative Commons Attribution 4.0 International License

HAL API platform

questionnaire software metadataMoraneGruenpeter7de56c632362954fa84172cad80afe4einria.fr1556733MoraneGruenpeterf85a43a5fb4a2e0778a77e017f28c8fdgmail.com2017-09-29 11:21:322017-10-03 17:20:132017-10-03 17:20:132017-09-292017-09-29contributorMoraneGruenpeterf85a43a5fb4a2e0778a77e017f28c8fdgmail.comCCSDhal-01587083https://hal.inria.fr/hal-01587083gruenpeter:hal-0158708320172017questionnaire software metadataMoraneGruenpeter7de56c632362954fa84172cad80afe4einria.fr1556733EnglishComputer Science [cs]SoftwareIRILLInitiative pour la Recherche et l'Innovation sur le Logiciel Libre
https://www.irill.org/Universite Pierre et Marie Curie - Paris 6UPMC
4 place Jussieu - 75005 Paris
http://www.upmc.fr/Institut National de Recherche en Informatique et en AutomatiqueInria
Domaine de VoluceauRocquencourt - BP 10578153 Le Chesnay Cedex
http://www.inria.fr/en/Universite Paris Diderot - Paris 7UPD7
5 rue Thomas-Mann - 75205 Paris cedex 13
http://www.univ-paris-diderot.fr""" # noqa self.atom_entry_data_badly_formatted = b""" """ def test_post_deposit_atom_empty_body_request(self): response = self.client.post( reverse(COL_IRI, args=[self.collection.name]), content_type='application/atom+xml;type=entry', data=self.atom_entry_data_empty_body) self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST) def test_post_deposit_atom_badly_formatted_is_a_bad_request(self): response = self.client.post( reverse(COL_IRI, args=[self.collection.name]), content_type='application/atom+xml;type=entry', data=self.atom_entry_data_badly_formatted) self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST) + def test_post_deposit_atom_without_slug_header_is_bad_request(self): + url = reverse(COL_IRI, args=[self.collection.name]) + + # when + response = self.client.post( + url, + content_type='application/atom+xml;type=entry', + data=self.atom_entry_data0, + # + headers + HTTP_IN_PROGRESS='false') + + self.assertIn(b'Missing SLUG header', response.content) + self.assertEqual(response.status_code, + status.HTTP_400_BAD_REQUEST) + def test_post_deposit_atom_unknown_collection(self): response = self.client.post( reverse(COL_IRI, args=['unknown-one']), content_type='application/atom+xml;type=entry', data=self.atom_entry_data3, HTTP_SLUG='something') self.assertEqual(response.status_code, status.HTTP_404_NOT_FOUND) def test_post_deposit_atom_entry_initial(self): """One deposit upload as atom entry """ # given external_id = 'urn:uuid:1225c695-cfb8-4ebb-aaaa-80da344efa6a' with self.assertRaises(Deposit.DoesNotExist): Deposit.objects.get(external_id=external_id) atom_entry_data = self.atom_entry_data0 % external_id.encode('utf-8') # when response = self.client.post( reverse(COL_IRI, args=[self.collection.name]), content_type='application/atom+xml;type=entry', data=atom_entry_data, + HTTP_SLUG='external-id', HTTP_IN_PROGRESS='false') # then self.assertEqual(response.status_code, status.HTTP_201_CREATED) response_content = parse_xml(BytesIO(response.content)) deposit_id = response_content[ '{http://www.w3.org/2005/Atom}deposit_id'] deposit = Deposit.objects.get(pk=deposit_id) self.assertEqual(deposit.collection, self.collection) self.assertEqual(deposit.external_id, external_id) self.assertEqual(deposit.status, 'ready') self.assertEqual(deposit.client, self.user) # one associated request to a deposit deposit_request = DepositRequest.objects.get(deposit=deposit) self.assertIsNotNone(deposit_request.metadata) self.assertFalse(bool(deposit_request.archive)) def test_post_deposit_atom_entry_tei(self): """One deposit upload as atom entry """ # given external_id = 'urn:uuid:1225c695-cfb8-4ebb-aaaa-80da344efa6a' with self.assertRaises(Deposit.DoesNotExist): Deposit.objects.get(external_id=external_id) atom_entry_data = self.atom_entry_tei # when response = self.client.post( reverse(COL_IRI, args=[self.collection.name]), content_type='application/atom+xml;type=entry', data=atom_entry_data, HTTP_SLUG=external_id, HTTP_IN_PROGRESS='false') # then self.assertEqual(response.status_code, status.HTTP_201_CREATED) response_content = parse_xml(BytesIO(response.content)) deposit_id = response_content[ '{http://www.w3.org/2005/Atom}deposit_id'] deposit = Deposit.objects.get(pk=deposit_id) self.assertEqual(deposit.collection, self.collection) self.assertEqual(deposit.external_id, external_id) self.assertEqual(deposit.status, 'ready') self.assertEqual(deposit.client, self.user) # one associated request to a deposit deposit_request = DepositRequest.objects.get(deposit=deposit) self.assertIsNotNone(deposit_request.metadata) self.assertFalse(bool(deposit_request.archive)) def test_post_deposit_atom_entry_multiple_steps(self): """Test one deposit upload.""" # given external_id = 'urn:uuid:2225c695-cfb8-4ebb-aaaa-80da344efa6a' with self.assertRaises(Deposit.DoesNotExist): deposit = Deposit.objects.get(external_id=external_id) # when response = self.client.post( reverse(COL_IRI, args=[self.collection.name]), content_type='application/atom+xml;type=entry', data=self.atom_entry_data1, HTTP_IN_PROGRESS='True', HTTP_SLUG=external_id) # then self.assertEqual(response.status_code, status.HTTP_201_CREATED) response_content = parse_xml(BytesIO(response.content)) deposit_id = response_content[ '{http://www.w3.org/2005/Atom}deposit_id'] deposit = Deposit.objects.get(pk=deposit_id) self.assertEqual(deposit.collection, self.collection) self.assertEqual(deposit.external_id, external_id) self.assertEqual(deposit.status, 'partial') self.assertEqual(deposit.client, self.user) # one associated request to a deposit deposit_requests = DepositRequest.objects.filter(deposit=deposit) self.assertEqual(len(deposit_requests), 1) atom_entry_data = self.atom_entry_data2 % external_id.encode('utf-8') update_uri = response._headers['location'][1] # when updating the first deposit post response = self.client.post( update_uri, content_type='application/atom+xml;type=entry', data=atom_entry_data, - HTTP_IN_PROGRESS='False', - HTTP_SLUG=external_id) + HTTP_IN_PROGRESS='False') # then self.assertEqual(response.status_code, status.HTTP_201_CREATED) response_content = parse_xml(BytesIO(response.content)) deposit_id = response_content[ '{http://www.w3.org/2005/Atom}deposit_id'] deposit = Deposit.objects.get(pk=deposit_id) self.assertEqual(deposit.collection, self.collection) self.assertEqual(deposit.external_id, external_id) self.assertEqual(deposit.status, 'ready') self.assertEqual(deposit.client, self.user) self.assertEqual(len(Deposit.objects.all()), 1) # now 2 associated requests to a same deposit deposit_requests = DepositRequest.objects.filter(deposit=deposit) self.assertEqual(len(deposit_requests), 2) for deposit_request in deposit_requests: actual_metadata = deposit_request.metadata self.assertIsNotNone(actual_metadata) self.assertFalse(bool(deposit_request.archive)) diff --git a/swh/deposit/tests/api/test_deposit_binary.py b/swh/deposit/tests/api/test_deposit_binary.py index d948b2b0..1894c75b 100644 --- a/swh/deposit/tests/api/test_deposit_binary.py +++ b/swh/deposit/tests/api/test_deposit_binary.py @@ -1,663 +1,684 @@ # Copyright (C) 2017 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import hashlib from django.core.files.uploadedfile import InMemoryUploadedFile from django.core.urlresolvers import reverse from io import BytesIO from rest_framework import status from rest_framework.test import APITestCase from swh.deposit.config import COL_IRI, EM_IRI from swh.deposit.models import Deposit, DepositRequest from swh.deposit.parsers import parse_xml from ..common import BasicTestCase, WithAuthTestCase class DepositNoAuthCase(APITestCase, BasicTestCase): """Deposit access are protected with basic authentication. """ def test_post_will_fail_with_401(self): """Without authentication, endpoint refuses access with 401 response """ url = reverse(COL_IRI, args=[self.collection.name]) data_text = b'some content' md5sum = hashlib.md5(data_text).hexdigest() external_id = 'some-external-id-1' # when response = self.client.post( url, content_type='application/zip', # as zip data=data_text, # + headers HTTP_SLUG=external_id, HTTP_CONTENT_MD5=md5sum, HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', HTTP_IN_PROGRESS='false', HTTP_CONTENT_LENGTH=len(data_text), HTTP_CONTENT_DISPOSITION='attachment; filename=filename0') # then self.assertEqual(response.status_code, status.HTTP_401_UNAUTHORIZED) class DepositTestCase(APITestCase, WithAuthTestCase, BasicTestCase): """Try and upload one single deposit """ def setUp(self): super().setUp() self.atom_entry_data0 = b""" Awesome Compiler hal urn:uuid:1225c695-cfb8-4ebb-aaaa-80da344efa6a %s 2017-10-07T15:17:08Z some awesome author something awesome-compiler This is an awesome compiler destined to awesomely compile stuff and other stuff compiler,programming,language 2005-10-07T17:17:08Z 2005-10-07T17:17:08Z release note related link Awesome https://hoster.org/awesome-compiler GNU/Linux 0.0.1 running all """ self.atom_entry_data1 = b""" hal urn:uuid:2225c695-cfb8-4ebb-aaaa-80da344efa6a 2017-10-07T15:17:08Z some awesome author something awesome-compiler This is an awesome compiler destined to awesomely compile stuff and other stuff compiler,programming,language 2005-10-07T17:17:08Z 2005-10-07T17:17:08Z release note related link Awesome https://hoster.org/awesome-compiler GNU/Linux 0.0.1 running all """ self.atom_entry_data2 = b""" %s """ self.atom_entry_data_empty_body = b""" """ self.atom_entry_data3 = b""" something """ self.data_atom_entry_ok = b""" Title urn:uuid:1225c695-cfb8-4ebb-aaaa-80da344efa6a 2005-10-07T17:17:08Z Contributor The abstract The abstract Access Rights Alternative Title Date Available Bibliographic Citation # noqa Contributor Description Has Part Has Version Identifier Is Part Of Publisher References Rights Holder Source Title Type """ + def test_post_deposit_binary_without_slug_header_is_bad_request(self): + url = reverse(COL_IRI, args=[self.collection.name]) + data_text = b'some content' + md5sum = hashlib.md5(data_text).hexdigest() + + # when + response = self.client.post( + url, + content_type='application/zip', # as zip + data=data_text, + # + headers + HTTP_CONTENT_MD5=md5sum, + HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', + HTTP_IN_PROGRESS='false', + HTTP_CONTENT_LENGTH=len(data_text), + HTTP_CONTENT_DISPOSITION='attachment; filename=filename0') + + self.assertIn(b'Missing SLUG header', response.content) + self.assertEqual(response.status_code, + status.HTTP_400_BAD_REQUEST) + def test_post_deposit_binary_upload_final_and_status_check(self): """Binary upload should be accepted """ # given url = reverse(COL_IRI, args=[self.collection.name]) data_text = b'some content' md5sum = hashlib.md5(data_text).hexdigest() external_id = 'some-external-id-1' # when response = self.client.post( url, content_type='application/zip', # as zip data=data_text, # + headers HTTP_SLUG=external_id, HTTP_CONTENT_MD5=md5sum, HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', HTTP_IN_PROGRESS='false', HTTP_CONTENT_LENGTH=len(data_text), HTTP_CONTENT_DISPOSITION='attachment; filename=filename0') # then self.assertEqual(response.status_code, status.HTTP_201_CREATED) response_content = parse_xml(BytesIO(response.content)) deposit_id = response_content[ '{http://www.w3.org/2005/Atom}deposit_id'] deposit = Deposit.objects.get(pk=deposit_id) self.assertEqual(deposit.status, 'ready') self.assertEqual(deposit.external_id, external_id) self.assertEqual(deposit.collection, self.collection) self.assertEqual(deposit.client, self.user) self.assertIsNone(deposit.swh_id) deposit_request = DepositRequest.objects.get(deposit=deposit) self.assertEquals(deposit_request.deposit, deposit) self.assertRegex(deposit_request.archive.name, 'filename0') response_content = parse_xml(BytesIO(response.content)) self.assertEqual( response_content['{http://www.w3.org/2005/Atom}deposit_archive'], 'filename0') self.assertEqual( response_content['{http://www.w3.org/2005/Atom}deposit_id'], deposit.id) edit_se_iri = reverse('edit_se_iri', args=[self.collection.name, deposit.id]) self.assertEqual(response._headers['location'], ('Location', edit_se_iri)) def test_post_deposit_binary_upload_only_supports_zip(self): """Binary upload only supports application/zip (for now)... """ # given url = reverse(COL_IRI, args=[self.collection.name]) data_text = b'some content' md5sum = hashlib.md5(data_text).hexdigest() external_id = 'some-external-id-1' # when response = self.client.post( url, content_type='application/octet-stream', data=data_text, # + headers HTTP_SLUG=external_id, HTTP_CONTENT_MD5=md5sum, HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', HTTP_IN_PROGRESS='false', HTTP_CONTENT_LENGTH=len(data_text), HTTP_CONTENT_DISPOSITION='attachment; filename=filename0') # then self.assertEqual(response.status_code, status.HTTP_415_UNSUPPORTED_MEDIA_TYPE) with self.assertRaises(Deposit.DoesNotExist): Deposit.objects.get(external_id=external_id) def test_post_deposit_binary_fails_if_unsupported_packaging_header( self): """Binary upload must have content_disposition header provided... """ # given url = reverse(COL_IRI, args=[self.collection.name]) data_text = b'some content' md5sum = hashlib.md5(data_text).hexdigest() external_id = 'some-external-id' # when response = self.client.post( url, content_type='application/zip', data=data_text, # + headers HTTP_SLUG=external_id, HTTP_CONTENT_MD5=md5sum, HTTP_PACKAGING='something-unsupported', HTTP_CONTENT_LENGTH=len(data_text), HTTP_CONTENT_DISPOSITION='attachment; filename=filename0') # then self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST) with self.assertRaises(Deposit.DoesNotExist): Deposit.objects.get(external_id=external_id) def test_post_deposit_binary_upload_fail_if_no_content_disposition_header( self): """Binary upload must have content_disposition header provided... """ # given url = reverse(COL_IRI, args=[self.collection.name]) data_text = b'some content' md5sum = hashlib.md5(data_text).hexdigest() external_id = 'some-external-id' # when response = self.client.post( url, content_type='application/zip', data=data_text, # + headers HTTP_SLUG=external_id, HTTP_CONTENT_MD5=md5sum, HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', HTTP_IN_PROGRESS='false', HTTP_CONTENT_LENGTH=len(data_text)) # then self.assertEqual(response.status_code, status.HTTP_400_BAD_REQUEST) with self.assertRaises(Deposit.DoesNotExist): Deposit.objects.get(external_id=external_id) def test_post_deposit_mediation_not_supported(self): """Binary upload only supports application/zip (for now)... """ # given url = reverse(COL_IRI, args=[self.collection.name]) data_text = b'some content' md5sum = hashlib.md5(data_text).hexdigest() external_id = 'some-external-id-1' # when response = self.client.post( url, content_type='application/zip', data=data_text, # + headers HTTP_SLUG=external_id, HTTP_CONTENT_MD5=md5sum, HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', HTTP_IN_PROGRESS='false', HTTP_CONTENT_LENGTH=len(data_text), HTTP_ON_BEHALF_OF='someone', HTTP_CONTENT_DISPOSITION='attachment; filename=filename0') # then self.assertEqual(response.status_code, status.HTTP_412_PRECONDITION_FAILED) with self.assertRaises(Deposit.DoesNotExist): Deposit.objects.get(external_id=external_id) # FIXME: Test this scenario (need a way to override the default # size limit in test scenario) # def test_post_deposit_binary_upload_fail_if_upload_size_limit_exceeded( # self): # """Binary upload must not exceed the limit set up... # """ # # given # url = reverse(COL_IRI, args=[self.collection.name]) # data_text = b'some content' # md5sum = hashlib.md5(data_text).hexdigest() # external_id = 'some-external-id' # # when # response = self.client.post( # url, # content_type='application/zip', # data=data_text, # # + headers # HTTP_SLUG=external_id, # HTTP_CONTENT_MD5=md5sum, # HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', # HTTP_IN_PROGRESS='false', # CONTENT_LENGTH=len(data_text), # HTTP_CONTENT_DISPOSITION='attachment; filename=filename0') # # then # self.assertEqual(response.status_code, # status.HTTP_403_FORBIDDEN) # with self.assertRaises(Deposit.DoesNotExist): # Deposit.objects.get(external_id=external_id) def test_post_deposit_2_post_2_different_deposits(self): """Making 2 post requests result in 2 different deposit """ url = reverse(COL_IRI, args=[self.collection.name]) data_text = b'some content' md5sum = hashlib.md5(data_text).hexdigest() # when response = self.client.post( url, content_type='application/zip', # as zip data=data_text, # + headers HTTP_SLUG='some-external-id-1', HTTP_CONTENT_MD5=md5sum, HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', HTTP_IN_PROGRESS='false', HTTP_CONTENT_LENGTH=len(data_text), HTTP_CONTENT_DISPOSITION='attachment; filename=filename0') # then self.assertEqual(response.status_code, status.HTTP_201_CREATED) response_content = parse_xml(BytesIO(response.content)) deposit_id = response_content[ '{http://www.w3.org/2005/Atom}deposit_id'] deposit = Deposit.objects.get(pk=deposit_id) deposits = Deposit.objects.all() self.assertEqual(len(deposits), 1) self.assertEqual(deposits[0], deposit) # second post response = self.client.post( url, content_type='application/zip', # as zip data=data_text, # + headers HTTP_SLUG='another-external-id', HTTP_CONTENT_MD5=md5sum, HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', HTTP_IN_PROGRESS='false', HTTP_CONTENT_LENGTH=len(data_text), HTTP_CONTENT_DISPOSITION='attachment; filename=filename1') self.assertEqual(response.status_code, status.HTTP_201_CREATED) response_content = parse_xml(BytesIO(response.content)) deposit_id2 = response_content[ '{http://www.w3.org/2005/Atom}deposit_id'] deposit2 = Deposit.objects.get(pk=deposit_id2) self.assertNotEqual(deposit, deposit2) deposits = Deposit.objects.all().order_by('id') self.assertEqual(len(deposits), 2) self.assertEqual(list(deposits), [deposit, deposit2]) def test_post_deposit_binary_and_post_to_add_another_archive(self): """One post to post a binary deposit, One other post to update the first deposit. """ # given url = reverse(COL_IRI, args=[self.collection.name]) external_id = 'some-external-id-1' # 1st archive to upload data_text0 = b'some other content' md5sum0 = hashlib.md5(data_text0).hexdigest() # when response = self.client.post( url, content_type='application/zip', # as zip data=data_text0, # + headers HTTP_SLUG=external_id, HTTP_CONTENT_MD5=md5sum0, HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', HTTP_IN_PROGRESS='true', HTTP_CONTENT_LENGTH=len(data_text0), HTTP_CONTENT_DISPOSITION='attachment; filename=filename0') # then self.assertEqual(response.status_code, status.HTTP_201_CREATED) response_content = parse_xml(BytesIO(response.content)) deposit_id = response_content[ '{http://www.w3.org/2005/Atom}deposit_id'] deposit = Deposit.objects.get(pk=deposit_id) self.assertEqual(deposit.status, 'partial') self.assertEqual(deposit.external_id, external_id) self.assertEqual(deposit.collection, self.collection) self.assertEqual(deposit.client, self.user) self.assertIsNone(deposit.swh_id) deposit_request = DepositRequest.objects.get(deposit=deposit) self.assertEquals(deposit_request.deposit, deposit) self.assertEquals(deposit_request.type.name, 'archive') self.assertRegex(deposit_request.archive.name, 'filename0') # 2nd archive to upload data_text = b'second archive uploaded' md5sum1 = hashlib.md5(data_text).hexdigest() # uri to update the content update_uri = reverse(EM_IRI, args=[self.collection.name, deposit_id]) # adding another archive for the deposit response = self.client.post( update_uri, content_type='application/zip', # as zip data=data_text, # + headers HTTP_SLUG=external_id, HTTP_CONTENT_MD5=md5sum1, HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', HTTP_IN_PROGRESS='false', HTTP_CONTENT_LENGTH=len(data_text), HTTP_CONTENT_DISPOSITION='attachment; filename=filename1') deposit = Deposit.objects.get(pk=deposit_id) self.assertEqual(deposit.status, 'ready') self.assertEqual(deposit.external_id, external_id) self.assertEqual(deposit.collection, self.collection) self.assertEqual(deposit.client, self.user) self.assertIsNone(deposit.swh_id) deposit_requests = list(DepositRequest.objects.filter(deposit=deposit). order_by('id')) # 2 deposit requests for the same deposit self.assertEquals(len(deposit_requests), 2) self.assertEquals(deposit_requests[0].deposit, deposit) self.assertEquals(deposit_requests[0].type.name, 'archive') self.assertRegex(deposit_requests[0].archive.name, 'filename0') self.assertEquals(deposit_requests[1].deposit, deposit) self.assertEquals(deposit_requests[1].type.name, 'archive') self.assertRegex(deposit_requests[1].archive.name, 'filename1') # only 1 deposit in db deposits = Deposit.objects.all() self.assertEqual(len(deposits), 1) def test_post_deposit_then_post_or_put_is_refused_when_status_ready(self): """When a deposit is complete, updating/adding new data to it is forbidden. """ url = reverse(COL_IRI, args=[self.collection.name]) external_id = 'some-external-id-1' # 1st archive to upload data_text0 = b'some other content' md5sum0 = hashlib.md5(data_text0).hexdigest() # when response = self.client.post( url, content_type='application/zip', # as zip data=data_text0, # + headers HTTP_SLUG=external_id, HTTP_CONTENT_MD5=md5sum0, HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', HTTP_IN_PROGRESS='false', HTTP_CONTENT_LENGTH=len(data_text0), HTTP_CONTENT_DISPOSITION='attachment; filename=filename0') # then self.assertEqual(response.status_code, status.HTTP_201_CREATED) response_content = parse_xml(BytesIO(response.content)) deposit_id = response_content[ '{http://www.w3.org/2005/Atom}deposit_id'] deposit = Deposit.objects.get(pk=deposit_id) self.assertEqual(deposit.status, 'ready') self.assertEqual(deposit.external_id, external_id) self.assertEqual(deposit.collection, self.collection) self.assertEqual(deposit.client, self.user) self.assertIsNone(deposit.swh_id) deposit_request = DepositRequest.objects.get(deposit=deposit) self.assertEquals(deposit_request.deposit, deposit) self.assertRegex(deposit_request.archive.name, 'filename0') # updating/adding is forbidden # uri to update the content edit_se_iri = reverse( 'edit_se_iri', args=[self.collection.name, deposit_id]) em_iri = reverse( 'em_iri', args=[self.collection.name, deposit_id]) # Testing all update/add endpoint should fail # since the status is ready # replacing file is no longer possible since the deposit's # status is ready r = self.client.put( em_iri, content_type='application/zip', data=data_text0, HTTP_SLUG=external_id, HTTP_CONTENT_MD5=md5sum0, HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', HTTP_IN_PROGRESS='false', HTTP_CONTENT_LENGTH=len(data_text0), HTTP_CONTENT_DISPOSITION='attachment; filename=filename0') self.assertEquals(r.status_code, status.HTTP_400_BAD_REQUEST) # adding file is no longer possible since the deposit's status # is ready r = self.client.post( em_iri, content_type='application/zip', data=data_text0, HTTP_SLUG=external_id, HTTP_CONTENT_MD5=md5sum0, HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', HTTP_IN_PROGRESS='false', HTTP_CONTENT_LENGTH=len(data_text0), HTTP_CONTENT_DISPOSITION='attachment; filename=filename0') self.assertEquals(r.status_code, status.HTTP_400_BAD_REQUEST) # replacing metadata is no longer possible since the deposit's # status is ready r = self.client.put( edit_se_iri, content_type='application/atom+xml;type=entry', data=self.data_atom_entry_ok, HTTP_SLUG=external_id, HTTP_CONTENT_LENGTH=len(self.data_atom_entry_ok)) self.assertEquals(r.status_code, status.HTTP_400_BAD_REQUEST) # adding new metadata is no longer possible since the # deposit's status is ready r = self.client.post( edit_se_iri, content_type='application/atom+xml;type=entry', data=self.data_atom_entry_ok, HTTP_SLUG=external_id, HTTP_CONTENT_LENGTH=len(self.data_atom_entry_ok)) self.assertEquals(r.status_code, status.HTTP_400_BAD_REQUEST) archive_content = b'some content representing archive' archive = InMemoryUploadedFile( BytesIO(archive_content), field_name='archive0', name='archive0', content_type='application/zip', size=len(archive_content), charset=None) atom_entry = InMemoryUploadedFile( BytesIO(self.data_atom_entry_ok), field_name='atom0', name='atom0', content_type='application/atom+xml; charset="utf-8"', size=len(self.data_atom_entry_ok), charset='utf-8') # replacing multipart metadata is no longer possible since the # deposit's status is ready r = self.client.put( edit_se_iri, format='multipart', data={ 'archive': archive, 'atom_entry': atom_entry, }) self.assertEquals(r.status_code, status.HTTP_400_BAD_REQUEST) # adding new metadata is no longer possible since the # deposit's status is ready r = self.client.post( edit_se_iri, format='multipart', data={ 'archive': archive, 'atom_entry': atom_entry, }) self.assertEquals(r.status_code, status.HTTP_400_BAD_REQUEST) diff --git a/swh/deposit/tests/api/test_deposit_multipart.py b/swh/deposit/tests/api/test_deposit_multipart.py index f002f201..f00f2afb 100644 --- a/swh/deposit/tests/api/test_deposit_multipart.py +++ b/swh/deposit/tests/api/test_deposit_multipart.py @@ -1,281 +1,318 @@ # Copyright (C) 2017 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from django.core.files.uploadedfile import InMemoryUploadedFile from django.core.urlresolvers import reverse from io import BytesIO from rest_framework import status from rest_framework.test import APITestCase from swh.deposit.config import COL_IRI from swh.deposit.models import Deposit, DepositRequest from swh.deposit.parsers import parse_xml from ..common import BasicTestCase, WithAuthTestCase class DepositMultipartTestCase(APITestCase, WithAuthTestCase, BasicTestCase): """Post multipart deposit scenario """ def setUp(self): super().setUp() self.data_atom_entry_ok = b""" Title urn:uuid:1225c695-cfb8-4ebb-aaaa-80da344efa6a 2005-10-07T17:17:08Z Contributor The abstract The abstract Access Rights Alternative Title Date Available Bibliographic Citation # noqa Contributor Description Has Part Has Version Identifier Is Part Of Publisher References Rights Holder Source Title Type """ self.data_atom_entry_update_in_place = """ urn:uuid:1225c695-cfb8-4ebb-aaaa-80da344efa7b Title Type """ + def test_post_deposit_multipart_without_slug_header_is_bad_request(self): + # given + url = reverse(COL_IRI, args=[self.collection.name]) + data_atom_entry = self.data_atom_entry_ok + + archive_content = b'some content representing archive' + archive = InMemoryUploadedFile( + BytesIO(archive_content), + field_name='archive0', + name='archive0', + content_type='application/zip', + size=len(archive_content), + charset=None) + + atom_entry = InMemoryUploadedFile( + BytesIO(data_atom_entry), + field_name='atom0', + name='atom0', + content_type='application/atom+xml; charset="utf-8"', + size=len(data_atom_entry), + charset='utf-8') + + # when + response = self.client.post( + url, + format='multipart', + data={ + 'archive': archive, + 'atom_entry': atom_entry, + }, + # + headers + HTTP_IN_PROGRESS='false') + + self.assertIn(b'Missing SLUG header', response.content) + self.assertEqual(response.status_code, + status.HTTP_400_BAD_REQUEST) + def test_post_deposit_multipart(self): """one multipart deposit should be accepted """ # given url = reverse(COL_IRI, args=[self.collection.name]) # from django.core.files import uploadedfile data_atom_entry = self.data_atom_entry_ok archive_content = b'some content representing archive' archive = InMemoryUploadedFile( BytesIO(archive_content), field_name='archive0', name='archive0', content_type='application/zip', size=len(archive_content), charset=None) atom_entry = InMemoryUploadedFile( BytesIO(data_atom_entry), field_name='atom0', name='atom0', content_type='application/atom+xml; charset="utf-8"', size=len(data_atom_entry), charset='utf-8') external_id = 'external-id' # when response = self.client.post( url, format='multipart', data={ 'archive': archive, 'atom_entry': atom_entry, }, # + headers HTTP_IN_PROGRESS='false', HTTP_SLUG=external_id) # then self.assertEqual(response.status_code, status.HTTP_201_CREATED) response_content = parse_xml(BytesIO(response.content)) deposit_id = response_content[ '{http://www.w3.org/2005/Atom}deposit_id'] deposit = Deposit.objects.get(pk=deposit_id) self.assertEqual(deposit.status, 'ready') self.assertEqual(deposit.external_id, external_id) self.assertEqual(deposit.collection, self.collection) self.assertEqual(deposit.client, self.user) self.assertIsNone(deposit.swh_id) deposit_requests = DepositRequest.objects.filter(deposit=deposit) self.assertEquals(len(deposit_requests), 2) for deposit_request in deposit_requests: self.assertEquals(deposit_request.deposit, deposit) if deposit_request.type.name == 'archive': self.assertRegex(deposit_request.archive.name, 'archive0') else: self.assertEquals( deposit_request.metadata[ '{http://www.w3.org/2005/Atom}id'], 'urn:uuid:1225c695-cfb8-4ebb-aaaa-80da344efa6a') def test_post_deposit_multipart_put_to_replace_metadata(self): """One multipart deposit followed by a metadata update should be accepted """ # given url = reverse(COL_IRI, args=[self.collection.name]) data_atom_entry = self.data_atom_entry_ok archive_content = b'some content representing archive' archive = InMemoryUploadedFile( BytesIO(archive_content), field_name='archive0', name='archive0', content_type='application/zip', size=len(archive_content), charset=None) atom_entry = InMemoryUploadedFile( BytesIO(data_atom_entry), field_name='atom0', name='atom0', content_type='application/atom+xml; charset="utf-8"', size=len(data_atom_entry), charset='utf-8') external_id = 'external-id' # when response = self.client.post( url, format='multipart', data={ 'archive': archive, 'atom_entry': atom_entry, }, # + headers HTTP_IN_PROGRESS='true', HTTP_SLUG=external_id) # then self.assertEqual(response.status_code, status.HTTP_201_CREATED) response_content = parse_xml(BytesIO(response.content)) deposit_id = response_content[ '{http://www.w3.org/2005/Atom}deposit_id'] deposit = Deposit.objects.get(pk=deposit_id) self.assertEqual(deposit.status, 'partial') self.assertEqual(deposit.external_id, external_id) self.assertEqual(deposit.collection, self.collection) self.assertEqual(deposit.client, self.user) self.assertIsNone(deposit.swh_id) deposit_requests = DepositRequest.objects.filter(deposit=deposit) self.assertEquals(len(deposit_requests), 2) for deposit_request in deposit_requests: self.assertEquals(deposit_request.deposit, deposit) if deposit_request.type.name == 'archive': self.assertRegex(deposit_request.archive.name, 'archive0') else: self.assertEquals( deposit_request.metadata[ '{http://www.w3.org/2005/Atom}id'], 'urn:uuid:1225c695-cfb8-4ebb-aaaa-80da344efa6a') replace_metadata_uri = response._headers['location'][1] response = self.client.put( replace_metadata_uri, content_type='application/atom+xml;type=entry', data=self.data_atom_entry_update_in_place, HTTP_IN_PROGRESS='false') self.assertEqual(response.status_code, status.HTTP_204_NO_CONTENT) # deposit_id did not change deposit = Deposit.objects.get(pk=deposit_id) self.assertEqual(deposit.status, 'ready') self.assertEqual(deposit.external_id, external_id) self.assertEqual(deposit.collection, self.collection) self.assertEqual(deposit.client, self.user) self.assertIsNone(deposit.swh_id) deposit_requests = DepositRequest.objects.filter(deposit=deposit) self.assertEquals(len(deposit_requests), 2) for deposit_request in deposit_requests: self.assertEquals(deposit_request.deposit, deposit) if deposit_request.type.name == 'archive': self.assertRegex(deposit_request.archive.name, 'archive0') else: self.assertEquals( deposit_request.metadata[ '{http://www.w3.org/2005/Atom}id'], 'urn:uuid:1225c695-cfb8-4ebb-aaaa-80da344efa7b') # FAILURE scenarios def test_post_deposit_multipart_only_archive_and_atom_entry(self): """Multipart deposit only accepts one archive and one atom+xml""" # given url = reverse(COL_IRI, args=[self.collection.name]) # from django.core.files import uploadedfile archive_content = b'some content representing archive' archive = InMemoryUploadedFile(BytesIO(archive_content), field_name='archive0', name='archive0', content_type='application/zip', size=len(archive_content), charset=None) other_archive_content = b"some-other-content" other_archive = InMemoryUploadedFile(BytesIO(other_archive_content), field_name='atom0', name='atom0', content_type='application/zip', size=len(other_archive_content), charset='utf-8') # when response = self.client.post( url, format='multipart', data={ 'archive': archive, 'atom_entry': other_archive, }, # + headers HTTP_IN_PROGRESS='false', HTTP_SLUG='external-id') # then self.assertEqual(response.status_code, status.HTTP_415_UNSUPPORTED_MEDIA_TYPE) # when archive.seek(0) response = self.client.post( url, format='multipart', data={ 'archive': archive, }, # + headers HTTP_IN_PROGRESS='false', HTTP_SLUG='external-id') # then self.assertEqual(response.status_code, status.HTTP_415_UNSUPPORTED_MEDIA_TYPE) diff --git a/swh/deposit/tests/common.py b/swh/deposit/tests/common.py index ad3aafed..fa5faf36 100644 --- a/swh/deposit/tests/common.py +++ b/swh/deposit/tests/common.py @@ -1,165 +1,166 @@ # Copyright (C) 2017 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import base64 import hashlib import os import shutil from django.core.urlresolvers import reverse from django.test import TestCase from io import BytesIO from rest_framework import status from swh.deposit.config import COL_IRI, EM_IRI from swh.deposit.models import DepositClient, DepositCollection from swh.deposit.models import DepositRequestType from swh.deposit.parsers import parse_xml from swh.deposit.settings.testing import MEDIA_ROOT class BasicTestCase(TestCase): """Mixin intended for data setup purposes (user, collection, etc...) """ def setUp(self): """Define the test client and other test variables.""" super().setUp() # expanding diffs in tests self.maxDiff = None # basic minimum test data deposit_request_types = {} # Add deposit request types for deposit_request_type in ['archive', 'metadata']: drt = DepositRequestType(name=deposit_request_type) drt.save() deposit_request_types[deposit_request_type] = drt _name = 'hal' # set collection up _collection = DepositCollection(name=_name) _collection.save() # set user/client up _client = DepositClient.objects.create_user(username=_name, password=_name) _client.collections = [_collection.id] _client.save() self.collection = _collection self.user = _client self.username = _name self.userpass = _name self.deposit_request_types = deposit_request_types def tearDown(self): # Clean up uploaded files in temporary directory (tests have # their own media root folder) if os.path.exists(MEDIA_ROOT): for d in os.listdir(MEDIA_ROOT): shutil.rmtree(os.path.join(MEDIA_ROOT, d)) class WithAuthTestCase(TestCase): """Mixin intended for testing the api with basic authentication. """ def setUp(self): super().setUp() _token = '%s:%s' % (self.username, self.userpass) token = base64.b64encode(_token.encode('utf-8')) authorization = 'Basic %s' % token.decode('utf-8') self.client.credentials(HTTP_AUTHORIZATION=authorization) def tearDown(self): super().tearDown() self.client.credentials() class CommonCreationRoutine(TestCase): """Mixin class to share initialization routine. cf: `class`:test_deposit_update.DepositReplaceExistingDataTest `class`:test_deposit_update.DepositUpdateDepositWithNewDataTest `class`:test_deposit_update.DepositUpdateFailuresTest `class`:test_deposit_delete.DepositDeleteTest """ def setUp(self): super().setUp() self.atom_entry_data0 = b""" some-external-id """ def create_simple_deposit_partial(self): """Create a simple deposit (1 request) in `partial` state and returns its new identifier. Returns: deposit id """ response = self.client.post( reverse(COL_IRI, args=[self.username]), content_type='application/atom+xml;type=entry', data=self.atom_entry_data0, + HTTP_SLUG='external-id', HTTP_IN_PROGRESS='true') # then assert response.status_code == status.HTTP_201_CREATED response_content = parse_xml(BytesIO(response.content)) deposit_id = response_content[ '{http://www.w3.org/2005/Atom}deposit_id'] return deposit_id def _update_deposit_with_status(self, deposit_id, status_partial=False): """Add to a given deposit another archive and update its current status to `ready` (by default). Returns: deposit id """ # add an archive data_text = b'some content' md5sum = hashlib.md5(data_text).hexdigest() # when response = self.client.post( reverse(EM_IRI, args=[self.username, deposit_id]), content_type='application/zip', # as zip data=data_text, # + headers HTTP_CONTENT_MD5=md5sum, HTTP_PACKAGING='http://purl.org/net/sword/package/SimpleZip', HTTP_IN_PROGRESS=status_partial, HTTP_CONTENT_LENGTH=len(data_text), HTTP_CONTENT_DISPOSITION='attachment; filename=filename0') # then assert response.status_code == status.HTTP_201_CREATED return deposit_id def create_deposit_ready(self): """Create a complex deposit (2 requests) in status `ready`. """ deposit_id = self.create_simple_deposit_partial() deposit_id = self._update_deposit_with_status(deposit_id) return deposit_id def create_deposit_partial(self): """Create a complex deposit (2 requests) in status `partial`. """ deposit_id = self.create_simple_deposit_partial() deposit_id = self._update_deposit_with_status( deposit_id, status_partial=True) return deposit_id