def load_data(): """Function to load the input data from blob storage. """ account_url = "https://hecdf.blob.core.windows.net" facts_blob_service = ContainerClient(account_url=account_url, container_name=facts_container, credential=facts_sas_token) print('````````````````````````````````````') print(' Begin loding data...') print('````````````````````````````````````') for blob in list(facts_blob_service.list_blobs()): file_name = blob.name print(file_name) download_stream = facts_blob_service.get_blob_client( file_name).download_blob() Path(f'./data/raw_in/{file_name}').parent.mkdir(parents=True, exist_ok=True) with open(f"./data/raw_in/{file_name}", "wb") as data: data.write(download_stream.readall()) print('````````````````````````````````````') print(' Finished loading data!') print('````````````````````````````````````') return 0
def get_preprocessed_file_from_azure(): service = ContainerClient( account_url=cfg.azure_details['account_url'], container_name=cfg.azure_details['storage_preprocessed_transcripts'], credential=cfg.azure_details['azure_storage_account_key']) blob_list = service.list_blobs() blob_name = '' for blob in blob_list: blob_name = blob.name # Create the BlobServiceClient object which will be used to create a container client blob_service_client = BlobServiceClient.from_connection_string( cfg.azure_details['account_connection_string']) local_file_name = blob_name full_path_to_file = os.path.join(local_path, local_file_name) # Create a blob client using the local file name as the name for the blob container_name = "preprocessed-transcripts" blob_client = blob_service_client.get_blob_client( container=cfg.azure_details['storage_preprocessed_transcripts'], blob=local_file_name) download_file_path = os.path.join(local_path, local_file_name) with open(download_file_path, "wb") as download_file: download_file.write(blob_client.download_blob().readall()) return local_file_name
def home_page(request): """ Access new orders from Azure 'new-orders' container, splits the order number and submission date from order's name and lists them to home page. """ container_client = ContainerClient(storage_url, 'new-orders', credential=None) new_orders = container_client.list_blobs() order_list = [] i = 1 for order in new_orders: order_no = order['name'].split('.')[0] creation_date = order['name'].split('.')[1] order_no_and_date_tuple = [order_no, creation_date] order_list.append(order_no_and_date_tuple) i += 1 order_list.sort(reverse=True) n = 1 for order in order_list: order.insert(0, n) n += 1 paginator = Paginator(order_list, 12) page = request.GET.get('page', 1) orders = paginator.page(page) context = { 'new_orders': orders, 'page_obj': orders, 'redirect_url': 'home_page', } # print(order_list) return JsonResponse(order_list, safe=False)
def completed_orders(request): """ Access shippped orders from Azure 'completed-orders' container, splits the order number and submission date from order's name and lists them """ container_client = ContainerClient(storage_url, 'completed-orders', credential=None) updated_order = container_client.list_blobs() order_list = [] i = 1 for order in updated_order: order_no = order['name'].split('.')[0] creation_date = order['name'].split('.')[1] order_no_and_date_tuple = [order_no, creation_date] order_list.append(order_no_and_date_tuple) i += 1 order_list.sort(reverse=True) n = 1 for order in order_list: order.insert(0, n) n += 1 page = request.GET.get('page', 1) paginator = Paginator(order_list, 6) orders = paginator.page(page) context = { 'orders': orders, 'container': 'completed-orders', 'redirect_url': 'completed_orders', } return JsonResponse(order_list, safe=False)
def list_blobs(): print("\nListing blobs...") container_client = ContainerClient() # List the blobs in the container blob_list = container_client.list_blobs() for blob in blob_list: print("\t" + blob.name)
def get_all_blobs_by_blob_container_name(self, storage_account_name, account_key, container_name): container_client = ContainerClient( account_url=f'{storage_account_name}.blob.core.windows.net', container_name=container_name, credential=account_key) blobs = container_client.list_blobs() return blobs
def combine_azure(self): from azure.storage.blob import ContainerClient, ContentSettings feed_uri = self.settings.get("FEED_URI") feed_prefix = self.settings.get("CITY_SCRAPERS_DIFF_FEED_PREFIX", "%Y/%m/%d") account_name, account_key = feed_uri[8::].split("@")[0].split(":") container = feed_uri.split("@")[1].split("/")[0] container_client = ContainerClient( "{}.blob.core.windows.net".format(account_name), container, credential=account_key, ) max_days_previous = 3 days_previous = 0 prefix_blobs = [] while days_previous <= max_days_previous: prefix_blobs = [ blob for blob in container_client.list_blobs( name_starts_with=( datetime.now() - timedelta(days=days_previous) ).strftime(feed_prefix) ) ] if len(prefix_blobs) > 0: break days_previous += 1 spider_blob_names = self.get_spider_paths([blob.name for blob in prefix_blobs]) meetings = [] for blob_name in spider_blob_names: feed_blob = container_client.get_blob_client(blob_name) feed_text = feed_blob.download_blob().content_as_text() meetings.extend( [json.loads(line) for line in feed_text.split("\n") if line] ) meetings = sorted(meetings, key=itemgetter(self.start_key)) yesterday_iso = (datetime.now() - timedelta(days=1)).isoformat()[:19] upcoming = [ meeting for meeting in meetings if meeting[self.start_key][:19] > yesterday_iso ] container_client.upload_blob( "latest.json", "\n".join([json.dumps(meeting) for meeting in meetings]), content_settings=ContentSettings(cache_control="no-cache"), overwrite=True, ) container_client.upload_blob( "upcoming.json", "\n".join([json.dumps(meeting) for meeting in upcoming]), content_settings=ContentSettings(cache_control="no-cache"), overwrite=True, )
def find_blobs(account_url: str, container_name: str, credential: str, prefix: str, suffix: str): blob_list = [] container = ContainerClient(account_url=account_url, container_name=container_name, credential=credential) for blob_record in container.list_blobs(name_starts_with=prefix): blob_name = blob_record['name'] if blob_name.endswith(suffix): blob_list.append(blob_name) return blob_list
def has_blob(container_client: ContainerClient, blob_name: str) -> bool: """ Return True if the blob file exists in the container. """ blobs = list(container_client.list_blobs()) logging.info('container_client.list_blobs') logging.info(blobs) filtered_blobs = list(filter(lambda b: b['name'] == blob_name, blobs)) logging.info('filtered_blobs') logging.info(filtered_blobs) return True if len(filtered_blobs) > 0 else False
class DeleteFilesAzure2AzureOperator(BaseOperator): @apply_defaults def __init__(self, account_name, account_key, container, *args, **kwargs): super(DeleteFilesAzure2AzureOperator, self).__init__(*args, **kwargs) self.client = ContainerClient( account_url=f"https://{account_name}.blob.core.windows.net/", credential=account_key, container_name=container) def execute(self, context): for file in self.client.list_blobs(): bob_cilent = self.client.get_blob_client(file) bob_cilent.delete_blob()
def listblobsapi(container, accountname, accountkey=None, nmax=None): block_blob_service = ContainerClient(accountname, container, credential=accountkey) logger.info("listing blobs in '{}/{}'".format(accountname, container)) if nmax is None: nmax = sys.maxsize blobs = block_blob_service.list_blobs() # TODO use namedtuple blob_list = [{ "name": blob.name, "date": blob.creation_time } for i, blob in enumerate(blobs) if i < nmax] return blob_list
class AzureDiffPipeline(DiffPipeline): """Azure Blob Storage backend for comparing previously scraped JSCalendar outputs""" def __init__(self, crawler, output_format): from azure.storage.blob import ContainerClient feed_uri = crawler.settings.get("FEED_URI") account_name, account_key = feed_uri[8::].split("@")[0].split(":") self.spider = crawler.spider self.container = feed_uri.split("@")[1].split("/")[0] self.container_client = ContainerClient( "{}.blob.core.windows.net".format(account_name), self.container, credential=account_key, ) self.feed_prefix = crawler.settings.get( "CITY_SCRAPERS_DIFF_FEED_PREFIX", "%Y/%m/%d" ) super().__init__(crawler, output_format) def load_previous_results(self): max_days_previous = 3 days_previous = 0 tz = timezone(self.spider.timezone) while days_previous <= max_days_previous: matching_blobs = self.container_client.list_blobs( name_starts_with=( tz.localize(datetime.now()) - timedelta(days=days_previous) ).strftime(self.feed_prefix) ) spider_blobs = [ blob for blob in matching_blobs if "{}.".format(self.spider.name) in blob.name ] if len(spider_blobs) > 0: break days_previous += 1 if len(spider_blobs) == 0: return [] blob = sorted(spider_blobs, key=attrgetter("name"))[-1] feed_blob = self.container_client.get_blob_client(blob.name) feed_text = feed_blob.download_blob().content_as_text() return [json.loads(line) for line in feed_text.split("\n") if line.strip()]
def _azure_get_configs(layer: "Layer") -> List[str]: providers = layer.gen_providers(0) credentials = Azure.get_credentials() storage_account_name = providers["terraform"]["backend"]["azurerm"][ "storage_account_name" ] container_name = providers["terraform"]["backend"]["azurerm"]["container_name"] storage_client = ContainerClient( account_url=f"https://{storage_account_name}.blob.core.windows.net", container_name=container_name, credential=credentials, ) prefix = "opta_config/" blobs = storage_client.list_blobs(name_starts_with=prefix) configs = [blob.name[len(prefix) :] for blob in blobs] if layer.name in configs: configs.remove(layer.name) return configs
def deleted_orders(request): """ Access all deleted orders from 'deleted-orders' container, splits the order number and submission date from order's name and lists them :param request: :return: """ container_client = ContainerClient(storage_url, 'deleted-orders', credential=None) deleted_order_list = container_client.list_blobs() order_list = [] i = 1 for order in deleted_order_list: order_no = order['name'].split('.')[0] creation_date = order['name'].split('.')[1] order_id_and_date_tuple = [order_no, creation_date] order_list.append(order_id_and_date_tuple) i += 1 order_list.sort(reverse=True) n = 1 for order in order_list: order.insert(0, n) n += 1 page = request.GET.get('page', 1) paginator = Paginator(order_list, 4) orders = paginator.page(page) # messages.info(request, 'yes man yes') context = { 'deleted_orders': orders, 'redirect_url': 'deleted_orders', } print(request.COOKIES) response = render(request, 'deleted_orders.html', context) # response.delete_cookie('cookie_name1') return response
def get_blob_manifest(container_client: ContainerClient) -> (list): '''Returns list of filenames.''' return [y.name for y in [x for x in container_client.list_blobs()]]
def list_blobs( container_client: ContainerClient, name_starts_with: Optional[str] = None ) -> List[str]: """Return a list of blobs.""" return [blob["name"] for blob in container_client.list_blobs(name_starts_with)]
def main(myblob: func.InputStream): try: logging.info('Python blob trigger function processed a request.') account_name = "###" # confidential account_key = "###" # confidential top_level_container_name = "###" # confidential blob_service = ContainerClient(account_url=account_name, container_name=top_level_container_name, credential=account_key) # Make connection with Azure SQL database server = 'datatrust-ff.database.windows.net' database = 'DataTrust' username = '******' password = '******' cnxn = pyodbc.connect('DRIVER={ODBC Driver 17 for SQL Server};SERVER='+server+';PORT=1433;DATABASE='+database+';UID='+username+';PWD='+ password) cursor = cnxn.cursor() cursor.execute("SELECT @@version;") row = cursor.fetchone() while row: print(row[0]) row = cursor.fetchone() # Download xlsx-files from Azure blob storage logging.info("\nList blobs in the container") generator = blob_service.list_blobs() for blob in generator: if blob.name.endswith('.xlsx'): logging.info("\t Blob name: " + blob.name) file_name = re.sub('.*/', '', blob.name) xlsx_file = open(file_name, 'wb') b = blob_service.download_blob(blob) b.readinto(xlsx_file) #xlsx_file.write(b) xlsx_file = open(file_name, 'rb') data = xlsx_file.read() data = pd.read_excel(data) headers = list(data.columns.values) list_values = data.values.tolist() blob_name = blob.name tableName = re.sub('.xlsx', '', blob_name) tableName = re.sub('^.*/', '', tableName) tableName = re.sub(' ', '_', tableName) # If table exists: remove and rewrite try: cursor.execute("DROP TABLE dbo." + tableName) except: print('Table does not exist yet') # Create new table query_string = 'CREATE TABLE dbo.' + tableName + ' (' # Add columns to table columns = '' for i in range(len(headers)): headers[i] = re.sub('[ /-]', '_', str(headers[i])) headers[i] = re.sub("[\(\)€'\.,]", '', str(headers[i])) columns += headers[i] + ', ' if i == len(headers) - 1: query_string += '\n' + headers[i] + ' VARCHAR(1000)' else: query_string += '\n' + headers[i] + ' VARCHAR(1000),' query_string += '\n);' query_string = re.sub('[/-]', '', query_string) cursor.execute(query_string) # Add rows to table query_string = "INSERT INTO dbo." + tableName + "(" + columns[:-2] +") VALUES " for row in range(len(list_values)): list_values[row] = [str(i) for i in list_values[row]] row_new = [] for item in list_values[row]: item = re.sub('[\(\)\r\n\,\'\-]', '', item) item = "'" + item + "'" row_new.append(item) row_new = ','.join(row_new) if (row + 1) % 1000 == 0 or row + 1 == len(list_values): query_string += '(' + row_new + ');' print(query_string) cursor.execute(query_string) query_string = "INSERT INTO dbo." + tableName + "(" + columns[:-2] +") VALUES " else: query_string += '(' + row_new + '),' cnxn.commit() except Exception as e: logging.exception(e)
def get_blob_list_information(container_client: ContainerClient) -> (list): '''Returns list of tuples with information about blobs..''' return [(y.name, y.blob_tier, y.size) for y in [x for x in container_client.list_blobs()]]
) account_info = blob_service_client.get_account_information() container_name = config['container_name'] container_client = ContainerClient( account_url=storage_url, container_name=container_name, credential=storage_key ) print(dir(container_client)) #print(dir(blob_client)) blob_list = container_client.list_blobs() for blob in blob_list: print("\t" + blob.name) blob_client_positions = BlobClient( account_url=storage_url, container_name=storage_name, blob_name="104人力銀行_SAP_positions.csv", credential=storage_key ) blob_client_companies = BlobClient( account_url=storage_url, container_name=storage_name, blob_name="104人力銀行_SAP_companies.csv", credential=storage_key
def check_subscription(tenant_id, tenant_name, sub_id, sub_name, creds): print("\n\t[*] Checking subscription {}:".format(sub_name), flush=True) storage_client = StorageManagementClient(creds, sub_id) # Obtain the management object for resources resource_client = ResourceManagementClient(creds, sub_id) # Retrieve the list of resource groups group_list = resource_client.resource_groups.list() resource_groups = [group.name for group in list(group_list)] print("\t\t[+] Found {} resource groups".format(len(resource_groups)), flush=True) group_to_names_dict = {group: dict() for group in resource_groups} accounts_counter = 0 for group in resource_groups: for item in storage_client.storage_accounts.list_by_resource_group( group): accounts_counter += 1 group_to_names_dict[group][item.name] = '' print("\t\t[+] Found {} storage accounts".format(accounts_counter), flush=True) for group in resource_groups: for account in group_to_names_dict[group].keys(): try: storage_keys = storage_client.storage_accounts.list_keys( group, account) storage_keys = {v.key_name: v.value for v in storage_keys.keys} group_to_names_dict[group][account] = storage_keys['key1'] except azure.core.exceptions.HttpResponseError: print( "\t\t[-] User do not have permissions to retrieve storage accounts keys in the given" " subscription", flush=True) print("\t\t Can not scan storage accounts", flush=True) return output_list = list() for group in resource_groups: for account in group_to_names_dict[group].keys(): key = group_to_names_dict[group][account] public_containers = check_storage_account(account, key) for cont in public_containers: access_level = cont.public_access container_client = ContainerClient( ENDPOINT_URL.format(account), cont.name, credential=key) files = [f.name for f in container_client.list_blobs()] ext_dict = count_files_extensions(files, EXTENSIONS) row = [ tenant_id, tenant_name, sub_id, sub_name, group, account, cont.name, access_level, CONTAINER_URL.format(account, cont.name), len(files) ] for ext in ext_dict.keys(): row.append(ext_dict[ext]) output_list.append(row) print("\t\t[+] Scanned all storage accounts successfully", flush=True) if len(output_list) > 0: print("\t\t[+] Found {} PUBLIC containers".format(len(output_list)), flush=True) else: print("\t\t[+] No PUBLIC containers found") header = [ "Tenant ID", "Tenant Name", "Subscription ID", "Subscription Name", "Resource Group", "Storage Account", "Container", "Public Access Level", "URL", "Total Files" ] for ext in EXTENSIONS: header.append(ext) header.append("others") write_csv('public-containers-{}.csv'.format(date.today()), header, output_list)
class TartanAir(object): def on_start(self): """This function is called once the Block is started """ account_url = 'https://tartanair.blob.core.windows.net/' container_name = 'tartanair-release1' self.container_client = ContainerClient(account_url=account_url, container_name=container_name, credential=None) self.envlist = [ 'abandonedfactory/', 'abandonedfactory_night/', 'amusement/', 'carwelding/', 'endofworld/', 'gascola/', 'hospital/', 'japanesealley/', 'neighborhood/', 'ocean/', 'office/', 'office2/', 'oldtown/', 'seasidetown/', 'seasonsforest/', 'seasonsforest_winter/', 'soulcity/', 'westerndesert/' ] self.diff_level = ["Easy", "Hard"][int(self.get_property("diff_level"))] self.env_ind = self.get_property("env_ind") self.trajlist = self.get_trajectory_list(self.envlist[self.env_ind], easy_hard=self.diff_level) self.trajs_len = len(self.trajlist) self.traj_id = self.get_property("traj_id") self.alert( "Selected Environment: {}".format(self.envlist[self.env_ind]), "INFO") self.alert("Difficulty Level: {}".format(self.diff_level), "INFO") self.alert( "Number of available trajectories: {}".format(self.trajs_len), "INFO") if (self.traj_id >= self.trajs_len): self.alert( "Trajectory id out of range[0, {}]".format(self.trajs_len - 1), "ERROR") self.frequency = self.get_property("fps") self.traj_dir = self.trajlist[self.traj_id] # Load Images List self.left_img_list = self.get_image_list(self.traj_dir, left_right='left') print('Find {} left images in {}'.format(len(self.left_img_list), self.traj_dir)) self.right_img_list = self.get_image_list(self.traj_dir, left_right='right') self.left_depth_list = self.get_depth_list(self.traj_dir, left_right='left') self.right_depth_list = self.get_depth_list(self.traj_dir, left_right='right') self.left_seg_list = self.get_seg_list(self.traj_dir, left_right='left') self.right_seg_list = self.get_seg_list(self.traj_dir, left_right='left') self.flow_list = self.get_flow_list(self.traj_dir) self.flow_mask_list = self.get_flow_mask_list(self.traj_dir) self.left_pose_file = self.get_posefile(self.traj_dir, left_right='left') self.right_pose_file = self.get_posefile(self.traj_dir, left_right='right') # Load poses bc = self.container_client.get_blob_client(blob=self.left_pose_file) data = bc.download_blob() text_file = open("OutputL.txt", "w") text_file.write(data.content_as_text()) text_file.close() self.pose_l = np.loadtxt("OutputL.txt") bc = self.container_client.get_blob_client(blob=self.right_pose_file) data = bc.download_blob() text_file = open("OutputR.txt", "w") text_file.write(data.content_as_text()) text_file.close() self.pose_r = np.loadtxt("OutputR.txt") def run(self): ltime = time.time() idx = 0 while True: if (time.time() - ltime >= 1 / self.frequency): if (idx == len(self.left_img_list)): idx = 0 # RGB Images left_img = self.read_image_file(self.left_img_list[idx]) right_img = self.read_image_file(self.right_img_list[idx]) header = Header() set_timestamp(header, time.time()) header.frame_id = "left_img" left_msg = from_ndarray(left_img, header) self.publish("left_img", left_msg) header.frame_id = "right_img" right_msg = from_ndarray(right_img, header) self.publish("right_img", right_msg) # Depth Images left_depth = self.read_numpy_file(self.left_depth_list[idx]) left_depth_vis = depth2vis(left_depth) header.frame_id = "left_depth" left_msg = from_ndarray(left_depth_vis, header) self.publish("left_depth", left_msg) right_depth = self.read_numpy_file(self.right_depth_list[idx]) right_depth_vis = depth2vis(right_depth) header.frame_id = "right_depth" right_msg = from_ndarray(right_depth_vis, header) self.publish("right_depth", right_msg) # Semantic Segmentation left_seg = self.read_numpy_file(self.left_seg_list[idx]) left_seg_vis = seg2vis(left_seg) header.frame_id = "left_segmentation" left_msg = from_ndarray(left_seg_vis, header) self.publish("left_segmentation", left_msg) right_seg = self.read_numpy_file(self.right_seg_list[idx]) right_seg_vis = seg2vis(right_seg) header.frame_id = "right_segmentation" right_msg = from_ndarray(right_seg_vis, header) self.publish("right_segmentation", right_msg) # Left Camera Pose pose_stamped = PoseStamped() pose_stamped.header = header pose_stamped.header.frame_id = "left_camera" pose = Pose() pose.position.x = self.pose_l[idx][0] pose.position.y = self.pose_l[idx][1] pose.position.z = self.pose_l[idx][2] pose.orientation.x = self.pose_l[idx][3] pose.orientation.y = self.pose_l[idx][4] pose.orientation.z = self.pose_l[idx][5] pose.orientation.w = self.pose_l[idx][6] pose_stamped.pose = pose self.publish("left_pose", pose_stamped) # Right Camera Pose pose_stamped = PoseStamped() pose_stamped.header = header pose_stamped.header.frame_id = "right_camera" pose = Pose() pose.position.x = self.pose_r[idx][0] pose.position.y = self.pose_r[idx][1] pose.position.z = self.pose_r[idx][2] pose.orientation.x = self.pose_r[idx][3] pose.orientation.y = self.pose_r[idx][4] pose.orientation.z = self.pose_r[idx][5] pose.orientation.w = self.pose_r[idx][6] pose_stamped.pose = pose self.publish("right_pose", pose_stamped) if (idx > 0): flow = self.read_numpy_file(self.flow_list[idx - 1]) flow_vis = flow2vis(flow) header.frame_id = "optical_flow" left_msg = from_ndarray(flow_vis, header) self.publish("optical_flow", left_msg) flow_mask = self.read_numpy_file(self.flow_mask_list[idx - 1]) flow_vis_w_mask = flow2vis(flow, mask=flow_mask) header.frame_id = "optical_flow_mask" right_msg = from_ndarray(flow_vis_w_mask, header) self.publish("optical_flow_mask", right_msg) ltime = time.time() idx += 1 def on_properties_changed(self, affected_properties): self.on_start() def get_environment_list(self): ''' List all the environments shown in the root directory ''' env_gen = self.container_client.walk_blobs() envlist = [] for env in env_gen: envlist.append(env.name) return envlist def get_trajectory_list(self, envname, easy_hard='Easy'): ''' List all the trajectory folders, which is named as 'P0XX' ''' assert (easy_hard == 'Easy' or easy_hard == 'Hard') traj_gen = self.container_client.walk_blobs(name_starts_with=envname + '/' + easy_hard + '/') trajlist = [] for traj in traj_gen: trajname = traj.name trajname_split = trajname.split('/') trajname_split = [tt for tt in trajname_split if len(tt) > 0] if trajname_split[-1][0] == 'P': trajlist.append(trajname) return trajlist def _list_blobs_in_folder(self, folder_name): """ List all blobs in a virtual folder in an Azure blob container """ files = [] generator = self.container_client.list_blobs( name_starts_with=folder_name) for blob in generator: files.append(blob.name) return files def get_image_list(self, trajdir, left_right='left'): assert (left_right == 'left' or left_right == 'right') files = self._list_blobs_in_folder(trajdir + '/image_' + left_right + '/') files = [fn for fn in files if fn.endswith('.png')] return files def get_depth_list(self, trajdir, left_right='left'): assert (left_right == 'left' or left_right == 'right') files = self._list_blobs_in_folder(trajdir + '/depth_' + left_right + '/') files = [fn for fn in files if fn.endswith('.npy')] return files def get_flow_list( self, trajdir, ): files = self._list_blobs_in_folder(trajdir + '/flow/') files = [fn for fn in files if fn.endswith('flow.npy')] return files def get_flow_mask_list( self, trajdir, ): files = self._list_blobs_in_folder(trajdir + '/flow/') files = [fn for fn in files if fn.endswith('mask.npy')] return files def get_posefile(self, trajdir, left_right='left'): assert (left_right == 'left' or left_right == 'right') return trajdir + '/pose_' + left_right + '.txt' def get_seg_list(self, trajdir, left_right='left'): assert (left_right == 'left' or left_right == 'right') files = self._list_blobs_in_folder(trajdir + '/seg_' + left_right + '/') files = [fn for fn in files if fn.endswith('.npy')] return files def read_numpy_file( self, numpy_file, ): ''' return a numpy array given the file path ''' bc = self.container_client.get_blob_client(blob=numpy_file) data = bc.download_blob() ee = io.BytesIO(data.content_as_bytes()) ff = np.load(ee) return ff def read_image_file( self, image_file, ): ''' return a uint8 numpy array given the file path ''' bc = self.container_client.get_blob_client(blob=image_file) data = bc.download_blob() ee = io.BytesIO(data.content_as_bytes()) img = cv2.imdecode(np.asarray(bytearray(ee.read()), dtype=np.uint8), cv2.IMREAD_COLOR) # im_rgb = img[:, :, [2, 1, 0]] # BGR2RGB return img
class AzureCloudInterface(CloudInterface): # Azure block blob limitations # https://docs.microsoft.com/en-us/rest/api/storageservices/understanding-block-blobs--append-blobs--and-page-blobs MAX_CHUNKS_PER_FILE = 50000 # Minimum block size allowed in Azure Blob Storage is 64KB MIN_CHUNK_SIZE = 64 << 10 # Azure Blob Storage permit a maximum of 4.75TB per file # This is a hard limit, while our upload procedure can go over the specified # MAX_ARCHIVE_SIZE - so we set a maximum of 1TB per file MAX_ARCHIVE_SIZE = 1 << 40 # The size of each chunk in a single object upload when the size of the # object exceeds max_single_put_size. We default to 2MB in order to # allow the default max_concurrency of 8 to be achieved when uploading # uncompressed WAL segments of the default 16MB size. DEFAULT_MAX_BLOCK_SIZE = 2 << 20 # The maximum amount of concurrent chunks allowed in a single object upload # where the size exceeds max_single_put_size. We default to 8 based on # experiments with in-region and inter-region transfers within Azure. DEFAULT_MAX_CONCURRENCY = 8 # The largest file size which will be uploaded in a single PUT request. This # should be lower than the size of the compressed WAL segment in order to # force the Azure client to use concurrent chunk upload for archiving WAL files. DEFAULT_MAX_SINGLE_PUT_SIZE = 4 << 20 # The maximum size of the requests connection pool used by the Azure client # to upload objects. REQUESTS_POOL_MAXSIZE = 32 def __init__( self, url, jobs=2, encryption_scope=None, credential=None, tags=None, max_block_size=DEFAULT_MAX_BLOCK_SIZE, max_concurrency=DEFAULT_MAX_CONCURRENCY, max_single_put_size=DEFAULT_MAX_SINGLE_PUT_SIZE, ): """ Create a new Azure Blob Storage interface given the supplied account url :param str url: Full URL of the cloud destination/source :param int jobs: How many sub-processes to use for asynchronous uploading, defaults to 2. """ super(AzureCloudInterface, self).__init__( url=url, jobs=jobs, tags=tags, ) self.encryption_scope = encryption_scope self.credential = credential self.max_block_size = max_block_size self.max_concurrency = max_concurrency self.max_single_put_size = max_single_put_size parsed_url = urlparse(url) if parsed_url.netloc.endswith(AZURE_BLOB_STORAGE_DOMAIN): # We have an Azure Storage URI so we use the following form: # <http|https>://<account-name>.<service-name>.core.windows.net/<resource-path> # where <resource-path> is <container>/<blob>. # Note that although Azure supports an implicit root container, we require # that the container is always included. self.account_url = parsed_url.netloc try: self.bucket_name = parsed_url.path.split("/")[1] except IndexError: raise ValueError("azure blob storage URL %s is malformed" % url) path = parsed_url.path.split("/")[2:] else: # We are dealing with emulated storage so we use the following form: # http://<local-machine-address>:<port>/<account-name>/<resource-path> logging.info("Using emulated storage URL: %s " % url) if "AZURE_STORAGE_CONNECTION_STRING" not in os.environ: raise ValueError( "A connection string must be provided when using emulated storage" ) try: self.bucket_name = parsed_url.path.split("/")[2] except IndexError: raise ValueError("emulated storage URL %s is malformed" % url) path = parsed_url.path.split("/")[3:] self.path = "/".join(path) self.bucket_exists = None self._reinit_session() def _reinit_session(self): """ Create a new session """ if self.credential: # Any supplied credential takes precedence over the environment credential = self.credential elif "AZURE_STORAGE_CONNECTION_STRING" in os.environ: logging.info("Authenticating to Azure with connection string") self.container_client = ContainerClient.from_connection_string( conn_str=os.getenv("AZURE_STORAGE_CONNECTION_STRING"), container_name=self.bucket_name, ) return else: if "AZURE_STORAGE_SAS_TOKEN" in os.environ: logging.info("Authenticating to Azure with SAS token") credential = os.getenv("AZURE_STORAGE_SAS_TOKEN") elif "AZURE_STORAGE_KEY" in os.environ: logging.info("Authenticating to Azure with shared key") credential = os.getenv("AZURE_STORAGE_KEY") else: logging.info( "Authenticating to Azure with default credentials") # azure-identity is not part of azure-storage-blob so only import # it if needed try: from azure.identity import DefaultAzureCredential except ImportError: raise SystemExit( "Missing required python module: azure-identity") credential = DefaultAzureCredential() session = requests.Session() adapter = requests.adapters.HTTPAdapter( pool_maxsize=self.REQUESTS_POOL_MAXSIZE) session.mount("https://", adapter) self.container_client = ContainerClient( account_url=self.account_url, container_name=self.bucket_name, credential=credential, max_single_put_size=self.max_single_put_size, max_block_size=self.max_block_size, session=session, ) @property def _extra_upload_args(self): optional_args = {} if self.encryption_scope: optional_args["encryption_scope"] = self.encryption_scope return optional_args def test_connectivity(self): """ Test Azure connectivity by trying to access a container """ try: # We are not even interested in the existence of the bucket, # we just want to see if Azure blob service is reachable. self.bucket_exists = self._check_bucket_existence() return True except (HttpResponseError, ServiceRequestError) as exc: logging.error("Can't connect to cloud provider: %s", exc) return False def _check_bucket_existence(self): """ Chck Azure Blob Storage for the target container Although there is an `exists` function it cannot be called by container-level shared access tokens. We therefore check for existence by calling list_blobs on the container. :return: True if the container exists, False otherwise :rtype: bool """ try: self.container_client.list_blobs().next() except ResourceNotFoundError: return False except StopIteration: # The bucket is empty but it does exist pass return True def _create_bucket(self): """ Create the container in cloud storage """ # By default public access is disabled for newly created containers. # Unlike S3 there is no concept of regions for containers (this is at # the storage account level in Azure) self.container_client.create_container() def _walk_blob_tree(self, obj, ignore=None): """ Walk a blob tree in a directory manner and return a list of directories and files. :param ItemPaged[BlobProperties] obj: Iterable response of BlobProperties obtained from ContainerClient.walk_blobs :param str|None ignore: An entry to be excluded from the returned list, typically the top level prefix :return: List of objects and directories in the tree :rtype: List[str] """ if obj.name != ignore: yield obj.name if isinstance(obj, BlobPrefix): # We are a prefix and not a leaf so iterate children for child in obj: for v in self._walk_blob_tree(child): yield v def list_bucket(self, prefix="", delimiter=DEFAULT_DELIMITER): """ List bucket content in a directory manner :param str prefix: :param str delimiter: :return: List of objects and dirs right under the prefix :rtype: List[str] """ res = self.container_client.walk_blobs(name_starts_with=prefix, delimiter=delimiter) return self._walk_blob_tree(res, ignore=prefix) def download_file(self, key, dest_path, decompress=None): """ Download a file from Azure Blob Storage :param str key: The key to download :param str dest_path: Where to put the destination file :param str|None decompress: Compression scheme to use for decompression """ obj = self.container_client.download_blob(key) with open(dest_path, "wb") as dest_file: if decompress is None: obj.download_to_stream(dest_file) return blob = StreamingBlobIO(obj) decompress_to_file(blob, dest_file, decompress) def remote_open(self, key, decompressor=None): """ Open a remote Azure Blob Storage object and return a readable stream :param str key: The key identifying the object to open :param barman.clients.cloud_compression.ChunkedCompressor decompressor: A ChunkedCompressor object which will be used to decompress chunks of bytes as they are read from the stream :return: A file-like object from which the stream can be read or None if the key does not exist """ try: obj = self.container_client.download_blob(key) resp = StreamingBlobIO(obj) if decompressor: return DecompressingStreamingIO(resp, decompressor) else: return resp except ResourceNotFoundError: return None def upload_fileobj( self, fileobj, key, override_tags=None, ): """ Synchronously upload the content of a file-like object to a cloud key :param fileobj IOBase: File-like object to upload :param str key: The key to identify the uploaded object :param List[tuple] override_tags: List of tags as k,v tuples to be added to the uploaded object """ # Find length of the file so we can pass it to the Azure client fileobj.seek(0, SEEK_END) length = fileobj.tell() fileobj.seek(0) extra_args = self._extra_upload_args.copy() tags = override_tags or self.tags if tags is not None: extra_args["tags"] = dict(tags) self.container_client.upload_blob(name=key, data=fileobj, overwrite=True, length=length, max_concurrency=self.max_concurrency, **extra_args) def create_multipart_upload(self, key): """No-op method because Azure has no concept of multipart uploads Instead of multipart upload, blob blocks are staged and then committed. However this does not require anything to be created up front. This method therefore does nothing. """ pass def _upload_part(self, upload_metadata, key, body, part_number): """ Upload a single block of this block blob. Uses the supplied part number to generate the block ID and returns it as the "PartNumber" in the part metadata. :param dict upload_metadata: Provider-specific metadata about the upload (not used in Azure) :param str key: The key to use in the cloud service :param object body: A stream-like object to upload :param int part_number: Part number, starting from 1 :return: The part metadata :rtype: dict[str, None|str] """ # Block IDs must be the same length for all bocks in the blob # and no greater than 64 characters. Given there is a limit of # 50000 blocks per blob we zero-pad the part_number to five # places. block_id = str(part_number).zfill(5) blob_client = self.container_client.get_blob_client(key) blob_client.stage_block(block_id, body, **self._extra_upload_args) return {"PartNumber": block_id} def _complete_multipart_upload(self, upload_metadata, key, parts): """ Finish a "multipart upload" by committing all blocks in the blob. :param dict upload_metadata: Provider-specific metadata about the upload (not used in Azure) :param str key: The key to use in the cloud service :param parts: The list of block IDs for the blocks which compose this blob """ blob_client = self.container_client.get_blob_client(key) block_list = [part["PartNumber"] for part in parts] extra_args = self._extra_upload_args.copy() if self.tags is not None: extra_args["tags"] = dict(self.tags) blob_client.commit_block_list(block_list, **extra_args) def _abort_multipart_upload(self, upload_metadata, key): """ Abort the upload of a block blob The objective of this method is to clean up any dangling resources - in this case those resources are uncommitted blocks. :param dict upload_metadata: Provider-specific metadata about the upload (not used in Azure) :param str key: The key to use in the cloud service """ # Ideally we would clean up uncommitted blocks at this point # however there is no way of doing that. # Uncommitted blocks will be discarded after 7 days or when # the blob is committed (if they're not included in the commit). # We therefore create an empty blob (thereby discarding all uploaded # blocks for that blob) and then delete it. blob_client = self.container_client.get_blob_client(key) blob_client.commit_block_list([], **self._extra_upload_args) blob_client.delete_blob() def delete_objects(self, paths): """ Delete the objects at the specified paths :param List[str] paths: """ try: # If paths is empty because the files have already been deleted then # delete_blobs will return successfully so we just call it with whatever # we were given responses = self.container_client.delete_blobs(*paths) except PartialBatchErrorException as exc: # Although the docs imply any errors will be returned in the response # object, in practice a PartialBatchErrorException is raised which contains # the response objects in its `parts` attribute. # We therefore set responses to reference the response in the exception and # treat it the same way we would a regular response. logging.warning( "PartialBatchErrorException received from Azure: %s" % exc.message) responses = exc.parts # resp is an iterator of HttpResponse objects so we check the status codes # which should all be 202 if successful errors = False for resp in responses: if resp.status_code == 404: logging.warning( "Deletion of object %s failed because it could not be found" % resp.request.url) elif resp.status_code != 202: errors = True logging.error( 'Deletion of object %s failed with error code: "%s"' % (resp.request.url, resp.status_code)) if errors: raise CloudProviderError( "Error from cloud provider while deleting objects - " "please check the Barman logs")
.config(f"fs.azure.sas.{workspace_container}.hecdf.blob.core.windows.net", workspace_sas_token) \ .getOrCreate() ## Define your blob services to access files on Azure Blob Storage from azure.storage.blob import ContainerClient testname = "koalas-tutorial/datasets/loan_preprocessed.csv" account_url = "https://hecdf.blob.core.windows.net" facts_blob_service = ContainerClient(account_url=account_url, container_name=facts_container, credential=facts_sas_token) workspace_blob_service = ContainerClient(account_url=account_url, container_name=workspace_container, credential=workspace_sas_token) # Create the parent folder blobs = list(facts_blob_service.list_blobs()) for blob in blobs: from pathlib import Path Path(f'../../data/raw/{blob.name}').parent.mkdir(parents=True, exist_ok=True) # From facts to your home directory with open(f"../../data/raw/{blob.name}", "wb") as data: download_stream = facts_blob_service.get_blob_client( blob.name).download_blob() data.write(download_stream.readall()) spark.stop()