Run Engine Manager API

Asyncio-Based API for Controlling RE Manager

Asyncio-based API is designed for use in applications that employ asyncio event loop. For example, HTTP server module is using asyncio-based API for communication with RE Manager. Communication is performed by instantiating ZMQCommSendAsync class and awaiting ZMQCommSendAsync.send_message() function, which accepts the method name and parameters as arguments and returns the response message.

ZMQCommSendAsync

API for communication with RE Manager via ZMQ.

ZMQCommSendAsync.send_message

Send message to ZMQ server and wait for the response.

ZMQCommSendAsync.close

Close ZMQ socket.

Thread-Based API for Controlling RE Manager

Thread-based API is designed for use in applications that are not based on asyncio event loop, such as PyQT applications or simple python scripts. Communication with RE Manager is performed by instantiating ZMQCommSendThreads class and calling ZMQCommSendThreads.send_message() function, which accepts the method name and parameters. The send_message() function may be called in blocking and non-blocking mode. In blocking mode, the function returns the response message. In non-blocking mode the function exits once ZMQ request is send to RE Manager and passes the response message to user-defined callback function once the response is received. The reference to the callback function is passed to send_message() function as one of the arguments.

ZMQCommSendThreads

Thread-based API for communication with RE Manager via ZMQ.

ZMQCommSendThreads.send_message

Send message to ZMQ server and wait for the response.

ZMQCommSendThreads.close

Close ZMQ socket.

Supported Methods For ZMQ Communication API

Brief Reference

The following reference describes the methods used for controlling RE Manager. For each method, the documentation contains the description of outgoing and returned parameters. Method name and the dictionary of method parameters are passed as arguments to send_message() API function.

The result of the request processing is returned as a dictionary of the returned parameters by send_message() function or passed as an argument of the callback. In case of communication error, send_message function may raise the exception (default) or return the error message. If reference to a callback function is passed to send_message (in thread-based API), the communication error message will be passed as msg_err parameter of the callback function

Get status and configuration information from RE Manager:

Existing and allowed plans and devices:

History of executed plans:

Open and close RE Worker environment:

Operations with the plan queue:

Start and stop execution of the plan queue:

Send commands to Bluesky Run Engine:

Monitor the list of active runs:

Run tasks in RE Worker namespace:

Lock/unlock RE Manager:

Management of IPython kernel (IPython mode):

Stopping RE Manager (mostly used in testing):

Detailed Reference

‘ping’

Method

‘ping’

Description

Causes RE Manager to send some predefined response. Currently ‘ping’ request is equivalent to ‘status’ request. The functionality may be changed in the future. Use ‘status’ method to request status information from RE Manager.

The request always succeeds.

Parameters

Returns

See ‘status’ method for information on returned data.

Execution

Immediate: no follow-up requests are required.

‘status’

Method

‘status’

Description

Returns status of RE Manager.

The request always succeeds.

Parameters

Returns

msg: str

application name and version, e.g. ‘RE Manager v0.0.10’

items_in_queue: int

the number of items in the plan queue

items_in_history: int

the number of items in the plan history

running_item_uid: str or None

item UID of the currently running plan or None if no plan is currently running. Monitor plan_queue_uid to detect updates of the queue and running item data.

plan_queue_uid: str

plan queue UID, which is updated each time the contents of the queue is changed. Monitor this parameter to determine when the queue or running item data was updated at the server (see queue_get API).

plan_history_uid: str

plan history UID, which is updated each time the contents of the history is changed. Monitor this parameter to determine when the history data should be downloaded.

task_results_uid: str

UID of the dictionary of task results. UID is updated each time results of a new completed tasks are added to the dictionary. Check the status of the pending tasks (see task_result API) once UID is changed.

plans_allowed_uid: str

UID for the list of allowed plans. UID is updated each time the contents of the list is changed. Monitor the UID to detect changes in the list of allowed plans and download the list from the server only when it is updated.

devices_allowed_uid: str

UID for the list of allowed devices. Similar to plans_allowed_uid.

plans_existing_uid: str

UID for the list of existing plans in RE Worker namespace. Similar to plans_allowed_uid.

devices_existing_uid: str

UID for the list of existing devices in RE Worker namespace. Similar to plans_allowed_uid.

‘run_list_uid’ - UID of the list of the active runs. Monitor this UID and

load the updated list of active runs once the UID is changed.

manager_state: str

state of RE Manager. Supported states:

  • ‘initializing’ - RE Manager is initializing (the RE Manager is starting or restarting)

  • ‘idle’ - RE Manager is idle and ready to execute requests. Many requests will fail if RE Manager is not in the idle state.

  • ‘paused’ - a plan was paused and Run Engine is in the paused state. The plan needs to be resumed, stopped, halted or aborted.

  • ‘creating_environment’ - RE Worker environment is in the process of being created.

  • ‘starting_queue’ - preparing to execute the queue.

  • ‘executing_queue’ - queue is being executed.

  • ‘executing_task’ - foreground task (function or script) is being executed.

  • ‘closing_environment’ - RE Worker environment is in the process of being closed (safe).

  • ‘destroying_environment’ - RE Worker environment is in the process of being destroyed (emergency).

re_state: str or None

current state of Bluesky Run Engine (see Blue Sky documentation) or None if RE Worker environment does not exist (or closed).

worker_environment_state: str

current state of the worker environment. Supported states: ‘initializing’, ‘failed’, ‘idle’, ‘executing_plan’, ‘executing_task’, ‘closing’ and ‘closed’. The ‘failed’ state may be briefly reported during environment initialization in case startup failed, but the environment is not closed yet. Running background tasks does not influence the state (‘executing_task’ is not set). The environment state is different from Run Engine state (re_state), e.g. the environment is considered idle when the current plan is paused.

worker_background_tasks: int

the number of background tasks, which are currently running in the worker environment. Excessive or growing number of background task may indicated serious issue with the environment. The parameter represents the number of background tasks executed in separate threads, such as tasks started by API requests script_upload and function_execute with run_in_background=True. Foreground tasks that block execution of plans and other tasks are not counted.

plan_queue_mode: dict

the dictionary of parameters that determine queue execution mode. The key/value pairs in the dictionary represent parameter names and values. Supported parameters: ‘loop’ (boolean) - indicates if the LOOP mode is enabled, ‘ignore_failures’ (boolean) - indicates if IGNORE_FAILURES mode is enabled.

queue_stop_pending: boolean

indicates if the request to stop the queue after completion of the current plan is pending.

queue_autostart_enabled: boolean

indicates if the autostart queue mode is enabled, see ‘queue_autostart’ API.

pause_pending: boolean

indicates if the request to pause the currently running plan was send to Run Engine (see re_pause method), but the plan is not stopped yet. It may take considerable time for deferred pause to be processed.

worker_environment_exists: boolean

indicates if RE Worker environment was created and plans could be executed.

ip_kernel_state: str or None

state of worker IPython kernel: None - environment is closed ‘disabled’ - IP kernel is not started or not used, ‘starting’ - startup in progress, ‘idle’ - IP kernel is idle and ready to execute tasks, ‘busy’ - kernel is busy executing a task started by the manager or another client client.

ip_kernel_captured: boolean or None

indicates if the IPython kernel is ‘captured’ by RE Manager: None - the environment is closed, True/False - indicates if IP kernel is running execution loop started by RE Manager. The loop is expected to run when the worker is executing foreground tasks (plans, functions, scripts). The kernel can not be accessed directly by clients (e.g. using Jupyter Console) while it is ‘captured’. The parameter is always True if the environment is open, but the worker is not using IPython kernel (the execution loop is always running).

lock_info_uid: str

UID of lock_info (see lock and lock_info API). Reload lock_info using lock_info API when the UID is changed.

lock: dict

The dictionary contains information on current status of the lock:

  • environment (boolean) - indicates if the RE Worker environment is locked. See the lock API for details.

  • queue (boolean) - indicates if the queue is locked. See the lock API for details.

Execution

Immediate: no follow-up requests are required.

‘config_get’

Method

‘config_get’

Description

Returns config info for RE Manager.

The request always succeeds.

Parameters

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

config: dict

config information for RE Manager:

  • ‘ip_connect_info’ (dict) - connect info for IPython kernel running in the worker process, empty dictionary if IPython kernel is not running. The returned dictionary may be saved as JSON to create IPython kernel connection file. The connection file may be passed to Jupyter Console with ‘–existing’ parameter.

Execution

Immediate: no follow-up requests are required.

‘plans_allowed’

Method

‘plans_allowed’

Description

Returns a dictionary that contains information on the allowed plans for a given user group. Monitor ‘plans_allowed_uid’ status field and download the list from the server only when the UID is changed.

Parameters

user_group: str

the name of the user group (e.g. ‘primary’).

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

plans_allowed: dict

the dictionary that contains information on the allowed plans. Dictionary keys are plan names.

plans_allowed_uid: str or None

UID of the list of allowed plans, None if the request fails.

Execution

Immediate: no follow-up requests are required.

‘devices_allowed’

Method

‘devices_allowed’

Description

Returns a dictionary that contains information on the allowed devices for a given user group. Monitor ‘devices_allowed_uid’ status field and download the list from the server only when the UID is changed.

Parameters

user_group: str

the name of the user group (e.g. ‘primary’).

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

devices_allowed: dict

the dictionary that contains information on the allowed devices. Dictionary keys are device names.

devices_allowed_uid: str or None

UID of the list of allowed devices, None if the request fails.

Execution

Immediate: no follow-up requests are required.

‘plans_existing’

Method

‘plans_existing’

Description

Returns a dictionary that contains information on existing plans in the RE Worker namespace. Monitor ‘plans_existing_uid’ status field and download the list from the server only when the UID is changed.

Parameters

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

plans_existing: dict

the dictionary that contains information on the existing plans. Dictionary keys are plan names.

plans_existing_uid: str or None

UID of the list of existing plans, None if the request fails.

Execution

Immediate: no follow-up requests are required.

‘devices_existing’

Method

‘devices_existing’

Description

Returns a dictionary that contains information on the existing devices in RE Worker namespace. Monitor ‘devices_existing_uid’ status field and download the list from the server only when the UID is changed.

Parameters

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

devices_existing: dict

the dictionary that contains information on the existing devices. Dictionary keys are device names.

devices_existing_uid: str or None

UID of the list of existing devices, None if the request fails

Execution

Immediate: no follow-up requests are required.

‘permissions_reload’

Method

‘permissions_reload’

Description

Reload user group permissions from the default location or the location set using command line parameters and generate lists of allowed plans and devices based on the lists of existing plans and devices. By default, the method will use current lists of existing plans and devices stored in memory. Optionally the method can reload the lists from the disk file (see restore_plans_devices parameter). The method always updates UIDs of the lists of allowed plans and devices even if the contents remain the same.

Parameters

restore_plans_devices: boolean (optional)

reload the lists of existing plans and devices from disk if True, otherwise use current lists stored in memory. Default: False.

restore_permissions: boolean (optional)

reload user group permissions from disk if True, otherwise use current permissions. Default: True.

lock_key: str (optional)

Lock key. The API fails if the queue is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

Immediate: no follow-up requests are required.

‘permissions_get’

Method

‘permissions_get’

Description

Download the dictionary of user group permissions currently used by RE Manager.

Parameters

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

user_group_permissions: dict

dictionary containing user group permissions.

Execution

Immediate: no follow-up requests are required.

‘permissions_set’

Method

‘permissions_set’

Description

Uploads the dictionary of user group permissions. If the uploaded dictionary contains a valid set of permissions different from currently used one, the new permissions are set as current and the updated lists of allowed plans and devices are generated. The method does nothing if the uploaded permissions are identical to currently used permissions. The API request fails if the uploaded dictionary does not pass validation.

Parameters

user_group_permissions: dict

dictionary, which contains user group permissions.

lock_key: str (optional)

Lock key. The API fails if the queue is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

Immediate: no follow-up requests are required.

‘history_get’

Method

‘history_get’

Description

Returns the list of items in the plan history.

The request always succeeds.

Parameters

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

items: list

list of items in the plan history, each item is represented by a dictionary of item parameters. Currently the plan history may contain only plans.

The dictionary contains a copy of the original item parameters and dictionary with results of execution (‘result’ key), which contains the following keys:

  • exit_status - exit status of the plan. The values are ‘completed’ (the plan execution is successfully completed), ‘failed’ (the plan execution failed; a plan can fail due to multiple reasons, including internal error of RE Manager; see the error message to determine the reason of failure), ‘stopped’ (the plan was paused, then stopped; the plan is considered successfully executed), ‘aborted’ and ‘halted’ (the plan was paused, then aborted or halted; the plan is considered failed), ‘unknown’ (the exit status information is lost, e.g. due to restart of RE Manager process, but plan information still needs to be placed in the history; this is very unlikely to happen in practice). IPython mode: if a plan is started by the manager, then paused and resumed by a client directly connected to IPython kernel (e.g. Jupyter Console), the exit status can be unknown (all runs are successfully completed or the plan opened no runs), aborted (at least one of the runs is aborted or halted) or failed (at least one of the runs failed);

  • run_uids - list of UIDs (str) of runs executed by the plan;

  • scan_ids - list of scan IDs (int) of runs executed by the plan;

  • time_start and time_stop - time of start and completion of the plan (not runs), floating point number returned by time.time().

  • msg - error message if the plan failed, empty string otherwise.

  • traceback - full traceback if the plan failed, empty string otherwise.

plan_history_uid: str

current plan history UID.

Execution

Immediate: no follow-up requests are required.

‘history_clear’

Method

‘history_clear’

Description

Clear the contents of the plan history.

The request always succeeds.

Parameters

lock_key: str (optional)

Lock key. The API fails if the queue is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

Immediate: no follow-up requests are required.

‘environment_open’

Method

‘environment_open’

Description

Initiate the creation of a new RE Worker environment. The request will succeed only if RE Manager is in ‘idle’ state and the environment does not exist. RE Worker environment must be created before plans could be executed. To restart the environment, close the existing environment first by sending ‘environment_close’ request. To destroy ‘frozen’ environment, use ‘environment_destroy’ method.

Parameters

lock_key: str (optional)

Lock key. The API fails if the environment is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

The request only initiates the sequence of creating a new environment. It may take a long time to start the environment. Monitor ‘manager_state’ and ‘worker_environment_exists’ status fields (‘status’ method) to detect when the operation is completed: ‘manager_state’ should have the value ‘opening_environment’ while operation is in progress and change to ‘idle’ when the operation completes and ‘worker_environment_exists’ is set to True if environment was created successfully.

‘environment_close’

Method

‘environment_close’

Description

Initiate the operation of closing the existing RE Worker environment. Fails if there is no existing environment or if RE Manager is not in ‘idle’ state. Use ‘environment_destroy’ method to close a non-responsive RE Worker environment.

Parameters

lock_key: str (optional)

Lock key. The API fails if the environment is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

The request initiates the sequence of closing the environment. Monitor ‘manager_state’ and ‘worker_environment_exists’ status fields (see ‘status’ method) to detection when the operation coompletes: ‘manager_state’ is expected to have the value ‘closing_environment’ while operation is in progress and switch to ‘idle’ when the operation completes and ‘worker_environment_exists’ is set False if environment was closed successfully.

‘environment_destroy’

Method

‘environment_destroy’

Description

Initiate the operation of destroying of the existing (unresponsive) RE Worker environment. The operation fails if there is no existing environment. The request is accepted by the server if status fields worker_environment_exists is True or manager_state is ‘creating_environment’, otherwise the request is rejected.

Parameters

lock_key: str (optional)

Lock key. The API fails if the environment is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

The request initiates the sequence of destroying the environment. Monitor ‘manager_state’ and ‘worker_environment_exists’ status fields (see ‘status’ method): ‘manager_state’ is expected to have the value ‘destroying_environment’ while operation is in progress and switch to ‘idle’ when the operation completes and ‘worker_environment_exists’ is set False if environment was destroyed successfully.

‘environment_update’

Method

‘environment_update’

Description

Update the state and cached parameters of the worker environment based on contents of the worker namespace. The updated parameters include the reference to the Run Engine and lists of existing and available plans and devices. The API is intended for using in cases when users bypass RE Manager to modify contents of the namespace, for example by connecting directly to IPython kernel (IPython mode) and executing commands via Jupyter Console.

Parameters

run_in_background: boolean (optional, default False)

Set this parameter True to execute the update in the background thread (while a plan or another foreground task is running). Generally, it is recommended to run the update in the main thread. Developers of data acquisition workflows and/or user specific code are responsible for thread safety.

lock_key: str (optional)

Lock key. The API fails if the environment is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

task_uid: str or None

Task UID can be used to check status of the task and download results once the task is completed (see task_result API).

Execution

The request initiates the update. The update is not instant, especially if the namespace is large. Monitor ‘manager_state’ (foreground task) or use ‘task_uid’ to check if the task execution is completed or the update is successful.

‘queue_mode_set’

Method

‘queue_mode_set’

Description

Sets parameters that define the mode of plan queue execution. The key/value pairs of the ‘mode’ dictionary represent names and values of the parameters that need to be changed. If mode={} then the mode will not be changed. To reset the mode parameters to the built-in default values, set mode=”default”. The request fails if the ‘mode’ parameter is not a dictionary or the ‘default’ string, the dictionary contains unsupported keys (mode parameters) or key values are of unsupported type. Supported mode parameters (dictionary keys): ‘loop’ (True/False) - enables/disables loop mode; ‘ignore_failures’ (True/False) - enables/disables the option to continue execution of the queue after a plan fails.

Parameters

mode: dict or str

the dictionary of queue mode parameters or ‘default’ string. Supported keys of the dictionary: ‘loop’ (boolean), ‘ignore_failures’ (boolean).

lock_key: str (optional)

Lock key. The API fails if the queue is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

Immediate: no follow-up requests are required.

‘queue_get’

Method

‘queue_get’

Description

Returns the items in the plan queue. The returned list of items may contain plans or instructions. Each item is represented as a dictionary. Plans and instructions can be distinguished by checking the value with the key ‘item_type’: ‘plan’ indicates that the item is a plan, while ‘instruction’ indicates that it is an instruction.

The request always succeeds.

Parameters

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

items: list

list of queue items

running_item: dict

parameters of the item representing currently running plan, empty dictionary ({}) is returned if no plan is currently running.

plan_queue_uid: str

current plan queue UID.

Execution

Immediate: no follow-up requests are required.

‘queue_item_add’

Method

‘queue_item_add’

Description

Add item to the queue. The item may be a plan or an instruction. By default the item is added to the back of the queue. Alternatively the item can be placed at the desired position in the queue or before or after one of the existing items.

Parameters

item: dict

the dictionary of plan or instruction parameters. Plans are distinguished from instructions based the value of the required parameter ‘item_type’. Currently supported item types are ‘plan’ and ‘instruction’.

user_group: str

the name of the user group (e.g. ‘primary’).

user: str

the name of the user (e.g. ‘Default User’). The name is included in the item metadata and may be used to identify the user who added the item to the queue. It is not passed to the Run Engine or included in run metadata.

pos: int, ‘front’ or ‘back’ (optional)

position of the item in the queue. RE Manager will attempt to insert the item at the specified position. The position may be positive or negative (counted from the back of the queue) integer. If ‘pos’ value is a string ‘front’ or ‘back’, then the item is inserted at the front or the back of the queue.

before_uid, after_uid: str (optional)

insert the item before or after the item with the given item UID.

lock_key: str (optional)

Lock key. The API fails if the queue is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Parameters ‘pos’, ‘before_uid’ and ‘after_uid’ are mutually exclusive.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

qsize: int or None

the number of items in the plan queue after the plan was added if the operation was successful, None otherwise

item: dict or None (optional)

the inserted item. The item contains the assigned item UID. In case of error the item may be returned without modification (with assigned UID). None will be returned if request does not contain item parameters.

Execution

Immediate: no follow-up requests are required.

‘queue_item_add_batch’

Method

‘queue_item_add_batch’

Description

Add a batch of items to the queue. The behavior of the function is identical to queue_item_add, except that it adds the batch of items instead of a single item. The batch may consist of any number of supported items (see queue_item_add method). Each item in the queue must successfully pass validation before any items are added to the queue. If an item fails validation, the whole batch is rejected. In case the batch is rejected, the function returns the detailed report for each item: success status indicating if the item passed validation and error message in case validation failed.

Parameters

items: list

the list containing a batch of items. Each element is a dictionary containing valid set of item parameters (see instructions for queue_item_add API). An empty item list will also be accepted.

user_group: str

the name of the user group (e.g. ‘primary’).

user: str

the name of the user (e.g. ‘Default User’). The name is included in the item metadata and may be used to identify the user who added the item to the queue. It is not passed to the Run Engine or included in run metadata.

pos: int, ‘front’ or ‘back’ (optional)

position of the first item of the batch in the queue. RE Manager will attempt to insert the batch of items so that the first item in the batch is at the specified position. The position may be positive or negative (counted from the back of the queue) integer. If ‘pos’ value is a string ‘front’ or ‘back’, then the items are inserted at the front or the back of the queue.

before_uid, after_uid: str (optional)

insert the batch of items before or after the item with the given item UID.

lock_key: str (optional)

Lock key. The API fails if the queue is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Parameters ‘pos’, ‘before_uid’ and ‘after_uid’ are mutually exclusive.

Returns

success: boolean

indicates if the request was processed successfully. The request fails if any item fails validation. The queue is not expected to be modified if the request fails. If the parameter is True, then validation of all items succeeded: there is no need to verify success status for each item returned in item_list parameter.

msg: str

error message in case of failure, empty string (‘’) otherwise.

qsize: int or None

the number of items in the plan queue after processing the request. The correct queue size may be returned even if the operation fails. In rare failing cases the parameter may return None.

items: list

the list of processed items. Each item contains inserted item (in case of success), passed item or None (in case of error). See notes for return item parameter for queue_add_item API.

results: list

the list of reports for each processed item. The size of the list is equal to the size of the list returned as items parameter. Each element of the list is a dictionary containing the following keys:

  • success - boolean value indicating if the validation of the item was successful.

  • msg - error message in case validation of the item failed.

Execution

Immediate: no follow-up requests are required.

‘queue_item_update’

Method

‘queue_item_update’

Description

Update the existing item in the queue. The method is intended for editing queue items, but may be used for replacing the existing items with completely different ones. The updated item may be a plan or an instruction. The item parameter ‘item_uid’ must be set to a UID of an existing queue item that is expected to be replaced. The method fails if the item UID is not found. By default, the UID of the updated item is not changed and ‘user’ and ‘user_group’ parameters are set to the values provided in the request. The ‘user_group’ is also used for validation of submitted item. If it is preferable to replace the item UID with a new random UID (e.g. if the item is replaced with completely different item), the method should be called with the optional parameter ‘replace=True’.

Parameters

item: dict

the dictionary of plan or instruction parameters. Plans are distinguished from instructions based the value of the required parameter ‘item_type’. Currently supported item types are ‘plan’ and ‘instruction’.

user_group: str

the name of the user group (e.g. ‘primary’).

user: str

the name of the user (e.g. ‘Default User’). The name is included in the item metadata and may be used to identify the user who added the item to the queue. It is not passed to the Run Engine or included in run metadata.

replace: boolean (optional)

replace the updated item UID with the new random UID (True) or keep the original UID (False). Default value is (False).

lock_key: str (optional)

Lock key. The API fails if the queue is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

qsize: int or None

the number of items in the plan queue after the plan was added if the operation was successful, None otherwise

item: dict or None (optional)

the inserted item. The item contains the assigned item UID. In case of error the item may be returned without modification (with assigned UID). None will be returned if request does not contain item parameters.

Execution

Immediate: no follow-up requests are required.

‘queue_item_get’

Method

‘queue_item_get’

Description

Read item from the queue. By default the item from the back of the queue is returned. Alternatively the item at the given position or the item with the given UID may be requested.

Parameters

pos: int, ‘front’ or ‘back’ (optional)

position of the item in the queue. RE Manager will attempt to return the item at the specified position. The position may be positive or negative (counted from the back of the queue) integer. If ‘pos’ value is a string ‘front’ or ‘back’, then the item at the front or the back of the queue is returned.

uid: str (optional)

uid of the requested item.

Parameters ‘pos’ and ‘uid’ are mutually exclusive.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

item: dict

the dictionary of item parameters, ({}) if operation failed.

Execution

Immediate: no follow-up requests are required.

‘queue_item_remove’

Method

‘queue_item_remove’

Description

Remove item from the queue. By default the last item in the queue is removed. Alternatively the position or UID of the item can be specified.

Parameters

pos: int, ‘front’ or ‘back’ (optional)

position of the item in the queue. RE Manager will attempt to remove the item at the specified position. The position may be positive or negative (counted from the back of the queue) integer. If ‘pos’ value is a string ‘front’ or ‘back’, then the item is removed from the front or the back of the queue.

uid: str (optional)

uid of the requested item.

lock_key: str (optional)

Lock key. The API fails if the queue is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Parameters ‘pos’ and ‘uid’ are mutually exclusive.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

item: dict

the dictionary of parameters of the removed item, ({}) if operation failed.

qsize: int or None

the number of items in the plan queue after the plan was added if the operation was successful, None otherwise

Execution

Immediate: no follow-up requests are required.

‘queue_item_remove_batch’

Method

‘queue_item_remove_batch’

Description

Remove a batch of items from the queue. The batch of items is defined as a list of item UIDs and passed as the parameter ‘uids’. The list of UIDs may be empty. By default, the function does not validate the batch and deletes all batch items found in the queue. Batch validation may be enabled by setting ‘ignore_missing=False’. In this case the method succeeds only if the batch does not contain repeated items and all items are found in the queue. If validation fails then no items are removed from the queue.

Parameters

uids: list(str) (required)

list of UIDs of the items in the batch. The list may not contain repeated UIDs. All UIDs must be present in the queue. The list may be empty.

ignore_missing: boolean (optional, default: True)

if the value is ‘False’, then the method fails if the batch contains repeating items or some of the batch items are not found in the queue. If ‘True’ (default), then the method attempts to remove all items in the batch and ignores missing items. The method returns the list of items that were removed from the queue.

lock_key: str (optional)

Lock key. The API fails if the queue is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

items: list(dict)

the list of items that were removed during the operation. The items in the list are arranged in the order in which they were removed. Returns empty list if the operation fails.

qsize: int or None

the size of the queue or None if operation fails.

Execution

Immediate: no follow-up requests are required.

‘queue_item_move’

Method

‘queue_item_move’

Description

Move item to a different position in the queue.

Parameters

pos: int, ‘front’ or ‘back’

current position of the item in the queue. Integer number can be negative.

uid: str (optional)

uid of the item to move.

pos_dest: int, ‘front’, ‘back’

new position of the item. Integer number can be negative.

before_uid, after_uid: str

UID of an existing item in the queue. The selected item will be moved before or after this item.

lock_key: str (optional)

Lock key. The API fails if the queue is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Parameters ‘pos’ and ‘uid’ are mutually exclusive, but at least one of them must be specified.

Parameters ‘pos_dest’, ‘before_uid’ and ‘after_uid’ are mutually exclusive, but at least one of them must be specified.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

item: dict

the dictionary of parameters of the moved item, ({}) if operation failed.

qsize: int or None

the size of the queue.

Execution

Immediate: no follow-up requests are required.

‘queue_item_move_batch’

Method

‘queue_item_move_batch’

Description

Move a batch of item to a different position in the queue. The method accepts a list of UIDs of the items included in the batch. The UIDs in the list must be unique (not repeated) and items with listed UIDs must exist in the queue. If the list is empty, then operation succeeds and the queue remains unchanged. The destination must be specified using one of the mutually exclusive parameters ‘pos_dest’, ‘before_uid’ or ‘after_uid’. The reference item with the UID of passed with the parameters ‘before_uid’ or ‘after_uid’ must not be in the batch. The parameter ‘reorder’ controls the order of the items in the moved batch and indicates whether items in the batch should be reordered with respect to the order of UIDs in the list ‘uids’. The batch may include any set of non-repeated items from the queue arranged in arbitrary order. By default (reorder=False) the batch is inserted in the specified position as a contiguous sequence of items ordered according to the UIDs in the list ‘uids’. If reorder=True, then the inserted items are ordered according to their original positions in the queue. It is assumed that the method will be mostly used with the default ordering option and user will be responsible for creating properly ordered lists of items. The other option is implemented for the cases when the user may want to submit randomly ordered lists of UIDs, but preserve the original order of the moved batch.

Parameters

uids: list(str) (required)

list of UIDs of the items in the batch. The list may not contain repeated UIDs. All UIDs must be present in the queue. The list may be empty.

pos_dest: ‘front’ or ‘back’

new position of the item. Only string values “front” and “back” are accepted.

before_uid, after_uid: str

UID of an existing item in the queue. The selected item will be moved before or after this item. The item with the specified UID may not be included in the batch.

reorder: boolean (optional, default: False)

Arranged moved items in the order of UIDs in the ‘uids’ list (False) or according to the original item positions in the queue (True).

lock_key: str (optional)

Lock key. The API fails if the queue is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Parameters ‘pos_dest’, ‘before_uid’ and ‘after_uid’ are mutually exclusive, but at least one of them must be specified.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

items: list(dict)

the list of items that were moved during the operation. The items in the list are arranged in the order in which they are inserted in the queue. Returns empty list if the operation fails.

qsize: int or None

the size of the queue or None if operation fails.

Execution

Immediate: no follow-up requests are required.

‘queue_item_execute’

Method

‘queue_item_execute’

Description

Immediately start execution of the submitted item. The item may be a plan or an instruction. The request fails if item execution can not be started immediately (RE Manager is not in IDLE state, RE Worker environment does not exist, etc.). If the request succeeds, the item is executed once. The item is not added to the queue if it can not be immediately started and it is not pushed back into the queue in case its execution fails/stops. If the queue is in the LOOP mode, the executed item is not added to the back of the queue after completion. The API request does not alter the sequence of enqueued plans.

The API is primarily intended for implementing of interactive workflows, in which users are controlling the experiment using client GUI application and user actions (such as mouse click on a plot) are converted into the requests to execute plans in RE Worker environment. Interactive workflows may be used for calibration of the instrument, while the queue may be used to run sequences of scheduled experiments.

The item is not added to the queue or change the existing queue. The API modifies plan_queue_uid status parameter, which is used for monitoring updates of the queue and running items. If the item is a plan, the results of execution are added to plan history as usual. The respective history item could be accessed to check if the plan was executed successfully.

The API does not start execution of the queue. Once execution of the submitted item is finished, RE Manager is switched to the IDLE state.

Parameters

item: dict

the dictionary of plan or instruction parameters. Plans are distinguished from instructions based the value of the required parameter ‘item_type’. Currently supported item types are ‘plan’ and ‘instruction’.

user_group: str

the name of the user group (e.g. ‘primary’).

user: str

the name of the user (e.g. ‘Default User’). The name is included in the item metadata and may be used to identify the user who added the item to the queue. It is not passed to the Run Engine or included in run metadata.

lock_key: str (optional)

Lock key. The API fails if the environment is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

qsize: int or None

the number of items in the plan queue, None if request fails otherwise.

item: dict or None (optional)

the inserted item. The item contains the assigned item UID. In case of error the item may be returned without modification (with assigned UID). None is returned if request does not contain item parameters.

Execution

Immediate: no follow-up requests are required.

‘queue_clear’

Method

‘queue_clear’

Description

Remove all items from the plan queue. The currently running plan does not belong to the queue and is not affected by this operation. If the plan fails or its execution is stopped, it will be pushed to the beginning of the queue.

Parameters

lock_key: str (optional)

Lock key. The API fails if the queue is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

Immediate: no follow-up requests are required.

‘queue_start’

Method

‘queue_start’

Description

Start execution of the queue. The operation succeeds only if RE Manager is in ‘idle’ state and RE Worker environment exists. This operation only initiates the process of starting the execution of the queue.

Parameters

lock_key: str (optional)

Lock key. The API fails if the environment is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

The request initiates the operation of starting the queue. Verify that the queue is running by checking the status fields: ‘manager_state’ is expected to have the value ‘executing_queue’ and ‘running_item_uid’ should return UID of the running plan. It is possible that the queue execution is successfully started, but immediately stopped (queue is empty, queue contains a very short plan, the first plan in the queue fails to start, the first item in the queue is ‘queue_stop’ instruction etc.). RE Manager is expected to handle those cases in orderly way, but the client should be capable of detecting and handling those events as well.

‘queue_stop’

Method

‘queue_stop’

Description

Request RE Manager to stop execution of the queue after completion of the currently running plan. The request succeeds only if the queue is currently running (‘manager_state’ status field has value ‘executing_queue’). The ‘queue_stop_pending’ status field can be used at any time to verify if the request is pending.

Parameters

lock_key: str (optional)

Lock key. The API fails if the environment is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

Immediate: no follow-up requests are required.

‘queue_stop_cancel’

Method

‘queue_stop_cancel’

Description

Cancel the pending request to stop execution of the queue after the currently running plan.

The request always succeeds.

Parameters

lock_key: str (optional)

Lock key. The API fails if the environment is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

Immediate: no follow-up requests are required.

‘queue_autostart’

Method

‘queue_autostart’

Description

Enable/disable autostart mode. In autostart mode, the queue execution is automatically started whenever the queue contains items and the manager and the environment are ready to execute plans. Client applications may check if the manager is in autostart mode using queue_autostart_enabled parameter of RE Manager status (see ‘status’ API).The autostart mode is not disabled after the queue runs out of items, but is automatically restarted once new items are added. The autostart mode is disabled if the API is called with enable=False, or if the queue or the running plan is stopped due to the following events:

  • ‘queue_stop’ queue instruction was executed;

  • the queue was stopped using ‘queue_stop’ API (the API call itself does not disable autostart, the queue remains in autostart mode if queue_stop was cancelled before the queue was stopped);

  • running plan was stopped/aborted/halted;

  • running plan failed, unless the ‘ignore_failures’ queue mode is enabled (see ‘queue_mode_set’ API).

The autostart mode may be enabled/disabled at any time. If the queue contains items, but queue can not be started immediately, e.g. due to environment being closed, the manager periodically checks the state of the manager and the worker and will attempt to start the queue once the manager and the worker are ready. If the queue runs out of items, the manager returns to IDLE state and will accept requests to execute foreground tasks: run a plan (‘queue_item_execute’ API), a function (‘function_execute’ API) or a script (‘script_upload’ API). If plans are added to the queue while the manager is busy, the queue is is automatically started once the task is complete.

If a plan is paused by the manager and resumed (and runs to completion) or stopped in Jupyter console, the autostart mode is not disabled and the next plan is started. If the plan is aborted or halted in Jupyter Console, then the autostart is disabled. This behavior may change in the future.

The request always succeeds.

Parameters

enable: boolean

pass True to enable the ‘autostart’ mode and False to disable it.

lock_key: str (optional)

Lock key. The API fails if the environment is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

Immediate: no follow-up requests are required.

‘re_pause’

Method

‘re_pause’

Description

Request Run Engine to pause currently running plan. The request will fail if RE Worker environment does not exist or no plan is currently running. The request only initates the sequence of pausing the plan.

If deferred pause is requested past the last checkpoint of the plan, the plan is run to completion and the queue is stopped. The stopped queue can not be resumed using re_resume method, instead queue_start method should be used to restart the queue. Check manager_state status flag to determine if the queue is stopped (‘idle’ state) or Run Engine is paused (‘paused’ state).

The pause_pending status flag is set if pause request is successfully passed to Run Engine. It may take significant time for deferred pause to be processed. The flag is cleared once the pending pause request is processed (the plan is paused or plan is completed and the queue is stopped).

Parameters

option: ‘immediate’ or ‘deferred’ (optional)

pause the plan immediately (roll back to the previous checkpoint) or continue to the next checkpoint. Default: ‘deferred’.

lock_key: str (optional)

Lock key. The API fails if the environment is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

The request only initiates the operation of pausing the plan. Wait until the plan is paused by polling ‘manager_state’ status field (expected value is ‘paused’).

‘re_resume’, ‘re_stop’, ‘re_abort’, ‘re_halt’

Method

‘re_resume’, ‘re_stop’, ‘re_abort’, ‘re_halt’

Description

Request Run Engine to resume, stop, abort or halt a paused plan. Fails if RE Worker environment does not exist or ‘manager_state’ status field value is not ‘paused’.

Parameters

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

lock_key: str (optional)

Lock key. The API fails if the environment is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Execution

The request only initiates the operation. Wait until the plan is paused by monitoring ‘manager_state’ status field (expected value is ‘executing_queue’ if execution is resumed or ‘idle’ if execution was stopped).

‘re_runs’

Method

‘re_runs’

Description

Request the list of active runs generated by the currently executed plans. The full list of active runs includes the runs that are currently open (‘open’ runs) and the runs that were already closed (‘closed’ runs). Simple single-run plans will have at most one run in the list. Monitor ‘run_list_uid’ RE Manager status field and retrieve the updated list once UID is changed. The UID of the retrieved list is included in the returned parameters

Parameters

option: ‘active’, ‘open’ or ‘closed’ (optional)

select between full list of ‘active’ (default) runs, the list of ‘open’ or ‘closed’ runs.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

run_list: list(dict)

the requested list of runs, list items are dictionaries with keys ‘uid’ (str), ‘scan_id’ (int), ‘is_open’ (boolean) and ‘exit_status’ (str or None). See Bluesky documentation for ‘exit_status’ values.

run_list_uid: str

UID of the returned run list, identical to the RE Manager status field with the same name.

Execution

Immediate: no follow-up requests are required.

‘script_upload’

Method

‘script_upload’

Description

Upload and execute script in RE Worker namespace. The script may add, change or replace objects defined in the namespace, including plans and devices. Dynamic modification of the worker namespace may be used to implement more flexible workflows. The API call updates the lists of existing and allowed plans and devices if necessary. Changes in the lists will be indicated by changed list UIDs. Use ‘task_result’ API to check if the script was loaded correctly. Note, that if the task fails, the script is still executed to the point where the exception is raised and the respective changes to the environment are applied.

Parameters

script: str

The string that contains the Python script. The rules for the script are the same as for Bluesky startup scripts. The script can use objects already existing in the RE Worker namespace.

update_lists: boolean (optional, default True)

Update lists of existing and available plans and devices after execution of the script. It is required to update the lists if the script adds or modifies plans and/or devices in RE Worker namespace, but it is more efficient to disable the update for other scripts, e.g. the scripts that print or modify variables in the namespace during iteractive debug session.

update_re: boolean (optional, default False)

The uploaded scripts may replace Run Engine (‘RE’) and Data Broker (‘db’) instances in the namespace. In most cases this operation should not be allowed, therefore it is disabled by default (update_re is False), i.e. if the script creates new RE and db objects, those objects are discarded. Set this parameter True to allow the server to replace RE and db objects. This parameter has no effect if the script is not creating new instances of RE and/or db.

run_in_background: boolean (optional, default False)

Set this parameter True to upload and execute the script in the background (while a plan or another foreground task is running). Generally, it is not recommended to update RE Worker namespace in the background. Background tasks are executed in separate threads and only thread-safe scripts should be uploaded in the background. Developers of data acquisition workflows and/or user specific code are responsible for thread safety.

lock_key: str (optional)

Lock key. The API fails if the environment is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

task_uid: str or None

Task UID can be used to check status of the task and download results once the task is completed (see task_result API).

Execution

The method initiates the operation. Monitor task_results_uid status field and call task_result API to check for success.

‘function_execute’

Method

‘function_execute’

Description

Start execution of a function in RE Worker namespace. The function must be defined in the namespace (in startup code or a script uploaded using script_upload method. The function may be executed as a foreground task (only if RE Manager and RE Worker environment are idle) or as a background task. Background tasks are executed in separate threads and may consume processing or memory resources and interfere with running plans or other tasks. RE Manager does not guarantee thread safety of the user code running in the background. Developers of startup code are fully responsible for preventing threading issues.

The method allows to pass parameters (args and kwargs) to the function. Once the task is completed, the results of the function execution, including the return value, can be loaded using task_result method. If the task fails, the return value is None and the error message and traceback are included in the result. The data types of parameters and return values must be JSON serializable. The task fails if the return value can not be serialized.

The method only initiates execution of the function. If the request is successful (success=True), the server starts the task, which attempts to execute the function with given name and parameters. The function may still fail start (e.g. if the user is permitted to execute function with the given name, but the function is not defined in the namespace). Use ‘task_result’ method with the returned task_uid to check the status of the tasks and load the result upon completion.

Parameters

item: dict

the dictionary that contains function name and parameters. The structure of dictionary is identical to item representing a plan or an instruction, except that item_type is ‘function’.

user_group: str

the name of the user group (e.g. ‘primary’).

user: str

the name of the user (e.g. ‘Default User’). The name is included in the item metadata and may be used to identify the user who submitted the item.

run_in_background: boolean (optional, default False)

Set this parameter True to start a background task. Background tasks can be started and executed while a plan or another foreground task is running. If workflow requires executing background tasks, user code should be analyzed for thread safety to ensure there are no potential threading issues. Developers of data acquisition workflows and/or user specific code are responsible for thread safety.

lock_key: str (optional)

Lock key. The API fails if the environment is locked and no valid key is submitted with the request. See documentation on ‘lock’ API for more details.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

item: dict or None (optional)

the item with information on the function. The item contains the assigned item UID. In case of error the item may be returned without modification. None will be returned if request does not contain item parameters. In current implementation the assigned item_uid is equal to task_uid, but it may change in the future.

task_uid: str or None

task UID can be used to check status of the task and download results once the task is completed (see task_result API). None is returned if the request fails.

Execution

The method initiates the operation. Monitor task_results_uid status field and call task_result API to check for success.

‘task_status’

Method

‘task_status’

Description

Returns the status of one or more tasks executed by the worker process. The request must contain one or more valid task UIDs, returned by one of APIs that starts tasks. A single UID may be passed as a string, multiple UIDs must be passed as as a list of strings. If a UID is passed as a string, then the returned status is also a string, if a list of one or more UIDs is passed, then the status is a dictionary that maps task UIDs and their status. The completed tasks are stored at the server at least for the period determined by retention time (currently 120 seconds after completion of the task). The expired results could be automatically deleted at any time and the method will return the task status as ‘not_found’.

Parameters

task_uid: str or list(str)

Task UID or a list of task UIDs.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

task_uid: str, list(str) or None

task UID or a list of task UIDs (expected to be the same as the input parameter). May be None if the request fails.

status: str or dict

status of the task(s) or None if the request (not task) failed. If task_uid is a string representing single UID, then status is a string, which is one of of ‘running’, ‘completed’ or ‘not_found’. If task_uid is a list of strings, then ‘status’ is a dictionary that maps task UIDs to status of the respective tasks.

Execution

Immediate: no follow-up requests are required.

‘task_result’

Method

‘task_result’

Description

Get the status and results of task execution. The completed tasks are stored at the server at least for the period determined by retention time (currently 120 seconds after completion of the task). The expired results could be automatically deleted at any time and the method will return the task status as ‘not_found’.

Parameters

task_uid: str

Task UID.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

task_uid: str or None

task UID (expected to be the same as the input parameter). May be None if the request fails.

status: ‘running’, ‘completed’, ‘not_found’ or None

status of the task or None if the request (not task) failed.

result: dict or None

Dictionary containing the information on a running task, results of the completed task or None if the request failed. The contents of the dictionary depends on the returned ‘status’:

  • ‘running’ - Keys: ‘task_uid’, ‘start_time’ and ‘run_in_background’.

  • ‘completed’ - Keys: ‘task_uid’, ‘success’ (True/False), ‘msg’ (short error message, empty string if execution was successful), ‘traceback’ (full traceback in case of error, empty string otherwise), ‘return_value’ (value returned by the task, e.g. by the executed function, None if the task failed), ‘time_start’ and ‘time_stop’.

  • ‘not_found’ - Empty dictionary.

Execution

Immediate: no follow-up requests are required.

‘lock’

Method

‘lock’

Description

Lock RE Manager with the provided lock key to prevent other clients from modifying the environment, starting plans or tasks or editing the queue. The lock is not intended for access control. The read-only API are not affected by the lock, therefore all monitoring client applications are expected to remain functional after the lock is applied. The lock does not influence internal operation of the manager, e.g. the running queue will continue running and has to be explicitly stopped if needed.

Each lockable API has an optional parameter ‘lock_key’. Passing a valid lock key (used to lock RE Manager) with the API requests allows to control RE Manager while it is locked. This supports the scenarios when a beamline scientists locks RE Manager with a unique code before entering the hutch to change samples or make adjustments and then safely runs a series of calibration or testing plans without interference from automated agents or remote users. A remote operators may still control locked RE Manager if the beamline scientists provides them with the lock key.

The API parameters allow to choose between locking the environment, the queue or both. Locking the environment affects the following API:

Locking the queue affects the following API:

The additional parameters include the name of the user (user, required) who locked RE Manager and the message to other users (note, optional) which may explain the reason why the manager is locked. The user name and the note is returned by lock_info API and included in the ‘Invalid lock key’ error messages.

An emergency lock key may be optionally set using QSERVER_EMERGENCY_LOCK_KEY_FOR_SERVER environment variable. The emergency key could be used to unlock RE Manager if the lock key is lost. The emergency lock key is accepted only by the unlock API.

Note

Restarting RE Manager does not change the lock status. The manager has to be unlocked using the valid lock key or the emergency lock key.

Parameters

lock_key: str

The lock key is an arbitrary non-empty string. Users/clients are expected to keep the key used to lock RE Manager and use it to unlock the manager or make API requests. If the lock key is lost by accident, then RE Manager may be unlocked using the emergency lock key.

environment: boolean (optional, default: False)

Enable lock for the API that control RE Worker environment. The request fails if both environment and queue are missing or False.

queue: boolean (optinal, default: False)

Enable lock for the API that control the queue. The request fails if both environment and queue are missing or False.

user: str

Name of the user who submits the request. The user name is returned as part of lock_info and included in error messages.

note: str or None (optional, default: None)

A text message to other users that explains the reason why RE Manager is locked. The note is returned as part of lock_info and included in error messages. If the value is None, then no message submitted.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

lock_info: dict

Dictionary containing the information on the status of the lock. The dictionary is also returned by lock_info API and includes the following fields:

  • ‘environment’ (boolean) - indicates if the RE Worker environment is locked.

  • ‘queue’ (boolean) - indicates if the queue is locked.

  • ‘user’ (str or None) - the name of the user who locked RE Manager, None if the lock is not set.

  • ‘note’ (str or None) - the text note left by the user who locked RE Manager, None if the lock is not set.

  • ‘time’ (float or None) - timestamp (time when RE Manager was locked), None if the lock is not set.

  • ‘time_str’ (str) - human-readable representation of the timestamp, empty string if the lock is not set.

  • ‘emergency_lock_key_is_set’ (boolean) - indicates if the optional emergency lock key is set.

lock_info_uid: str

UID of lock_info. The UID is also returned in RE Manager status and could be monitored to detect updates of lock_info.

Execution

Immediate: no follow-up requests are required.

‘lock_info’

Method

‘lock_info’

Description

Load the lock status of RE Manager and optionally validate the lock key. Monitor lock_info_uid field of RE Manager status (see documentation for the ‘status’ API) to detect changes in lock status and download the lock status only when the UID changes.

Parameters

lock_key: str or None (optional, default: None)

The parameter is used to validate the lock key (not the emergency lock key, see the documentation for ‘lock’ API). If the lock key is a string that matches the key used to lock RE Manager, then the request succeeds. If no lock key is passed with the request, the lock key is None or RE Manager is not locked, then no validation is performed and the request always succeeds. The request fails if RE Manager is locked and lock_key is not valid.

Returns

success: boolean

indicates if the request was processed successfully. The request succeeds if RE Manager is unlocked or the lock_key parameter is not specified or None. If lock_key is a string, then the request succeeds if the key is valid and fails otherwise.

msg: str

error message in case of failure, empty string (‘’) otherwise.

lock_info: dict

Dictionary containing the information on the status of the lock. See the documentation on ‘lock’ API for the detailed description.

lock_info_uid: str

UID of lock_info. The UID is also returned in RE Manager status and could be monitored to detect updates of lock_info.

Execution

Immediate: no follow-up requests are required.

‘unlock’

Method

‘unlock’

Description

Unlock RE Manager with the lock key used to lock the manager or the emergency lock key. See the documentation for ‘lock’ API for more details.

Parameters

lock_key: str

The lock key must match the key used to lock RE Manager (with ‘lock’ API) or the emergency lock key.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

lock_info: dict

Dictionary containing the updated information on the lock status after the request is processed. See the documentation on ‘lock’ API for the detailed description.

lock_info_uid: str

UID of lock_info. The UID is also returned in RE Manager status and could be monitored to detect updates of lock_info.

Execution

Immediate: no follow-up requests are required.

‘kernel_interrupt’

Method

‘kernel_interrupt’

Description

Send interrupt request (Ctrl-C) to the running IPython kernel. The API call fails if IPython mode is not enabled or environment does not exist (there is no IPython kernel). The API is primarily intended to interrupt tasks started by clients connected directly to IPython kernel (such as Jupyter Console) and by default it fails if the manager is executing a plan or a task. Set the interrupt_task and/or interrupt_plan parameters True in order to be able to interrupt a running foreground task or a plan (single interrupt initiates deferred pause, two consecutive interrupts initiate immediate pause). Note, that ‘re_pause’ API is more reliable method of pausing the plan.

Parameters

interrupt_task: boolean (optional, default: False)

Allow interrupting a foreground task (e.g. a function or a script) started by RE Manager.

interrupt_plan: boolean (optional, default: False)

Allow interrupting a running plan. By default the API fails if a plan is running in the worker environment (Run Engine is in the ‘running’ state), whether the plan was started by RE Manager or by connecting directly to the kernel (e.g. using Jupyter Console). Note, that using ‘re_pause’ API is the preferred way of pausing a running plan even if the plan was started bypassing the manager.

task_uid: str

Task UID.

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

Immediate: no follow-up requests are required.

‘manager_stop’

Method

‘manager_stop’

Description

Exit RE Manager application. Clients will probably not need to initiate exit remotely, but ability to do so is extremely useful for automated testing.

Parameters

option: ‘safe_on’ or ‘safe_off’ (optional)

if the option of ‘safe_on’ is selected (default), then the request fails unless ‘manager_state’ status field is ‘idle’ (no plans are running). If the option is ‘safe_off’ then RE Worker environment is destroyed (worker process is terminated and all data that was not saved is discarded).

Returns

success: boolean

indicates if the request was processed successfully.

msg: str

error message in case of failure, empty string (‘’) otherwise.

Execution

The request only initiates the operation of exiting RE Manager. If the request succeeds it may be expected that RE Manager application will eventually be exited and stops it stops responding to requests.

‘manager_kill’

Method

‘manager_kill’

Description

Freezes RE Manager process by stopping the asyncio event loop. By design, the manager process is expected to restart after 5 seconds of inactivity. The restart of the manager process should not affect the state of the queue or running plans. This function is implemented for testing purposes only. There is no practical reason for a client application to send this request.

Parameters

Returns

RE Manager will not respond to the request.

Execution

Immediate: no follow-up requests are required.