Metadata-Version: 2.1
Name: switch-api
Version: 0.2.21
Summary: A complete package for data ingestion into the Switch Automation Platform.
Home-page: UNKNOWN
Author: Switch Automation Pty Ltd.
License: MIT License
Platform: UNKNOWN
Classifier: Development Status :: 2 - Pre-Alpha
Classifier: License :: OSI Approved :: MIT License
Classifier: Intended Audience :: Other Audience
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Natural Language :: English
Requires-Python: >=3.8.*
Description-Content-Type: text/markdown
License-File: LICENCE
License-File: AUTHORS.rst

# Switch Automation library for Python
This is a package for data ingestion into the Switch Automation software platform. 

You can find out more about the platform on [Switch Automation](https://www.switchautomation.com)

## Getting started

### Prerequisites
* Python 3.8 or later is required to use this package. 
* You must have a [Switch Automation user account](https://www.switchautomation.com/our-solution/) to use this package. 

### Install the package
Install the Switch Automation library for Python with [pip](https://pypi.org/project/pip/):

```bash
pip install switch_api
```

# History

## 0.2.21

### Updated
- Fix on `get_data` method in `dataset` module 
  - Sync parameter structure to backend API for `get_data`
  - List of dict containing properties of `name`, `value`, and `type` items
  - `type` property must be one of subset of the new Literal `DATA_SET_QUERY_PARAMETER_TYPES`

## 0.2.20

### Added
- Newly supported Azure Storage Account: GatewayMqttStorage
- An optional property on QueueTask to specific QueueType
  - Default: DataIngestion

## 0.2.19

### Fixed
- Fix on `upsert_timeseries` method in `integration` module
  - Normalized TimestampId and TimestampLocalId seconds
- Minor fix on `upsert_entities_affected` method in `integration` utils module
  - Prevent upsert entities affected count when data feed file status Id is not valid
- Minor fix on `get_metadata_keys` method in `integration` helper module
  - Fix for issue when a portfolio does not contain any values in the ApiMetadata table


## 0.2.18

### Added
- Added new `is_specific_timezone` parameter in `upsert_timeseries` method of `integration` module
  - Accepts a timezone name as the specific timezone used by the source data. 
  - Can either be of type str or bool and defaults to the value of False.
  - Cannot have value if 'is_local_time' is set to True.
  - Retrieve list of available timezones using 'get_timezones' method in `integration` module
  

    | is_specific_timezone  | is_local_time | Description |
    | --------------------- | -------- | ----------- |
    | False  | False  | Datetimes in provided data is already in UTC and should remain as the value of Timestamp. The TimestampLocal (conversion to site-local Timezone) is calculated. |
    | False  | True  | Datetimes in provided data is already in the site-local Timezone & should be used to set the value of the TimestampLocal field. The UTC Timestamp is calculated |
    | Has Value | True  | NOT ALLOWED |
    | Has Value | False  | Both Timestamp and TimestampLocal fields will are calculated. Datetime is converted to UTC then to Local. |
    | True  |   | NOT ALLOWED |
    | '' (empty string)  |   | NOT ALLOWED |
  

### Fixed
- Minor fix on `upsert_tags` and `upsert_device_metadata` methods in `integration` module 
  - List of required_columns was incorrectly being updated when these functions were called
- Minor fix on `upsert_event_work_order_id` method in `integration` module when attempting to update status of an Event

### Updated
- Update on `DiscoveryIntegrationInput` namedtuple - added `job_id`
- Update `upsert_discovered_records` method required columns in `integration` module
  - add required `JobId` column for Data Frame parameter


## 0.2.17
### Fixed
- Fix on `upsert_timeseries()` method in `integration` module for duplicate records in ingestion files
  - records whose Timestamp falls in the exact DST start created 2 records with identical values but different TimestampLocal
    - one has the TimestampLocal of a DST and the other does not

### Updated
- Update on `get_sites()` method in `integration` module for `InstallationCode` column
  - when the `InstallationCode' value is null in the database it returns an empty string
  - `InstallationCode` column is explicity casted to dtype 'str'


## 0.2.16
### Added

- Added new 5 minute interval for `EXPECTED_DELIVERY` Literal in `automation` module
  - support for data feed deployments Email, FTP, Upload, and Timer
  - usage: expected_delivery='5min'
  
### Fixed

- Minor fix on `upsert_timeseries()` method using `data_feed_file_status_id` parameter in `integration` module.
  - `data_feed_file_status_id` parameter value now synced between process records and ingestion files when supplied

### Updated

- Reduced ingestion files records chunking by half in `upsert_timeseries()` method in `integration` module.
  - from 100k records chunk down to 50k records chunk

## 0.2.15

### Updated

- Optimized `upsert_timeseries()` method memory upkeep in `integration` module.

## 0.2.14

### Fixed

- Minor fix on `invalid_file_format()` method creating structured logs in `error_handlers` module.

## 0.2.13

### Updated

- Freeze Pandera[io] version to 0.7.1
  - PandasDtype has been deprecated since 0.8.0

### Compatibility

- Ensure local environment is running Pandera==0.7.1 to match cloud container state
- Downgrade/Upgrade otherwise by running:
  - pip uninstall pandera
  - pip install switch_api

## 0.2.12

### Added

- Added `upsert_tags()` method to the `integration` module.
  - Upsert tags to existing sites, devices, and sensors
  - Upserting of tags are categorised by the tagging level which are Site, Device, and Sensor level
  - Input dataframe requires `Identifier' column whose value depends on the tagging level specified
    - For Site tag level, InstallationIds are expected to be in the `Identifier` column
    - For Device tag level, DeviceIds are expected to be in the `Identifier` column
    - For Sensor tag level, ObjectPropertyIds are expected to be in the `Identifier` column
- Added `upsert_device_metadata()` method to the `integration` module.
  - Upsert metadata to existing devices

### Usage

- `upsert_tags()`
  - sw.integration.upsert_tags(api_inputs=api_inputs, df=raw_df, tag_level='Device')
  - sw.integration.upsert_tags(api_inputs=api_inputs, df=raw_df, tag_level='Sensor')
  - sw.integration.upsert_tags(api_inputs=api_inputs, df=raw_df, tag_level='Site')
- `upsert_device_metadata()`
  - sw.integration.upsert_device_metadata(api_inputs=api_inputs, df=raw_df)

## 0.2.11

### Added

- New `cache` module that handles cache data related transactions
  - `set_cache` method that stores data to cache
  - `get_cache` method that gets stored data from cache
  - Stored data can be scoped / retrieved into three categories namely Task, Portfolio, and DataFeed scopes
    - For Task scope,
      - Data cache can be retrieved by any Portfolio or Datafeed that runs in same Task
      - provide TaskId (self.id when calling from the driver)
    - For DataFeed scope,
      - Data cache can be retrieved (or set) within the Datafeed deployed in portfolio
      - Provide UUID4 for local testing. api_inputs.data_feed_id will be used when running in the cloud.
    - For Portfolio scope:
      - Data cache can be retrieved (or set) by any Datafeed deployed in portfolio
      - scope_id will be ignored and api_inputs.api_project_id will be used.

## 0.2.10

### Fixed

- Fixed issue with `upsert_timeseries_ds()` method in the `integration` module where required fields such as
  `Timestamp`, `ObjectPropertyId`, `Value` were being removed.

## 0.2.9

### Added

- Added `upsert_timeseries()` method to the `integration` module.
  - Data ingested into table storage in addition to ADX Timeseries table
  - Carbon calculation performed where appropriate
    - Please note: If carbon or cost are included as fields in the `Meta` column then no carbon / cost calculation will be performed

### Changed

- Added `DriverClassName` to required columns for `upsert_discovered_records()` method in the `integration` module

### Fixed

- A minor fix to 15-minute interval in `upsert_timeseries_ds()` method in the `integration` module.

## 0.2.8

### Changed

- For the `EventWorkOrderTask` class in the `pipeline` module, the `check_work_order_input_valid()` and the
  `generate_work_order()` methods expect an additional 3 keys to be included by default in the dictionary passed to
  the `work_order_input` parameter:
  - `InstallationId`
  - `EventLink`
  - `EventSummary`

### Fixed

- Issue with the header/payload passed to the API within the `upsert_event_work_order_id()`
  function of the `integration` module.

## 0.2.7

### Added

- New method, `deploy_as_on_demand_data_feed()` added to the `Automation` class of the `pipeline` module
  - this new method is only applicable for tasks that subclass the `EventWorkOrderTask` base class.

### Changed

- The `data_feed_id` is now a required parameter, not optional, for the following methods on the `Automation` class of
  the `pipeline` module:
  - `deploy_on_timer()`
  - `deploy_as_email_data_feed()`
  - `deploy_as_ftp_data_feed()`
  - `deploy_as_upload_data_feed()`
- The `email_address_domain` is now a required parameter, not optional, for the `deploy_as_email_data_feed()` method
  on the `Automation` class of the `pipeline` module.

### Fixed

- issue with payload on `switch_api.pipeline.Automation.register_task()` method for `AnalyticsTask` and
  `EventWorkOrderTask` base classes.

## 0.2.6

### Fixed

- Fixed issues on 2 methods in the `Automation` class of the `pipeline` module:
  - `delete_data_feed()`
  - `cancel_deployed_data_feed()`

### Added

In the `pipeline` module:

- Added new class `EventWorkOrderTask`
  - This task type is for generation of work orders in 3rd party systems via the Switch Automation Platform's Events UI.

### Changed

In the `pipeline` module:

- `AnalyticsTask` - added a new method & a new abstract property:
  - `analytics_settings_definition` abstract property - defines the required inputs (& how these are displayed in the
    Switch Automation Platform UI) for the task to successfully run
  - added `check_analytics_settings_valid()` method that should be used to validate the
    `analytics_settings` dictionary passed to the `start()` method contains the required keys for the task to
    successfully run (as defined by the `analytics_settings_definition`)

In the `error_handlers` module:

- In the `post_errors()` function, the parameter `errors_df` is renamed to `errors` and now accepts strings in
  addition to pandas.DataFrame

### Removed

Due to cutover to a new backend, the following have been removed:

- `run_clone_modules()` function from the `analytics` module
- the entire `platform_insights` module including the :
  - `get_current_insights_by_equipment()` function

## 0.2.5

### Added

- The `Automation` class of the `pipeline` module has 2 new methods added: -`delete_data_feed()`
  - Used to delete an existing data feed and all related deployment settings
  - `cancel_deployed_data_feed()`
    - used to cancel the specified `deployment_type` for a given `data_feed_id`
    - replaces and expands the functionality previously provided in the `cancel_deployed_timer()` method which has been
      removed.

### Removed

- Removed the `cancel_deployed_timer()` method from the `Automation` class of the `pipeline` module
  - this functionality is available through the new `cancel_deployed_data_feed()` method when `deployment_type`
    parameter set to `['Timer']`

## 0.2.4

### Changed

- New parameter `data_feed_name` added to the 4 deployment methods in the `pipeline` module's `Automation` class
  - `deploy_as_email_data_feed()`
  - `deploy_as_ftp_data_feed()`
  - `deploy_as_upload_data_feed()`
  - `deploy_on_timer()`

## 0.2.3

### Fixed

- Resolved minor issue on `register_task()` method for the `Automation` class in the `pipeline` module.

## 0.2.2

### Fixed

- Resolved minor issue on `upsert_discovered_records()` function in `integration` module related to device-level
  and sensor-level tags.

## 0.2.1

### Added

- New class added to the `pipeline` module
  - `DiscoverableIntegrationTask` - for API integrations that are discoverable.
    - requires `process()` & `run_discovery()` abstract methods to be created when sub-classing
    - additional abstract property, `integration_device_type_definition`, required compared to base `Task`
- New function `upsert_discovered_records()` added to the `integration` module
  - Required for the `DiscoverableIntegrationTask.run_discovery()` method to upsert discovery records to Build -
    Discovery & Selection UI

### Fixed

- Set minimum msal version required for the switch_api package to be installed.

## 0.2.0

Major overhaul done of the switch_api package. A complete replacement of the API used by the package was done.

### Changed

- The `user_id` parameter has been removed from the `switch_api.initialise()` function.
  - Authentication of the user is now done via Switch Platform SSO. The call to initialise will trigger a web browser
    window to open to the platform login screen.
    - Note: each call to initialise for a portfolio in a different datacentre will open up browser and requires user to
      input their username & password.
    - for initialise on a different portfolio within the same datacentre, the authentication is cached so user will not
      be asked to login again.
- `api_inputs` is now a required parameter for the `switch_api.pipeline.Automation.register_task()`
- The `deploy_on_timer()`, `deploy_as_email_data_feed()`, `deploy_as_upload_data_feed()`, and
  `deploy_as_ftp_data_feed()` methods on the `switch_api.pipeline.Automation` class have an added parameter:
  `data_feed_id`
  - This new parameter allows user to update an existing deployment for the portfolio specified in the `api_inputs`.
  - If `data_feed_id` is not supplied, a new data feed instance will be created (even if portfolio already has that
    task deployed to it)

## 0.1.18

### Changed

- removed rebuild of the ObjectProperties table in ADX on call to `upsert_device_sensors()`
- removed rebuild of the Installation table in ADX on call to `upsert_sites()`

## 0.1.17

### Fixed

- Fixed issue with `deploy_on_timer()` method of the `Automation` class in the `pipeline` module.
- Fixed column header issue with the `get_tag_groups()` function of the `integration` module.
- Fixed missing Meta column on table generated via `upsert_workorders()` function of the `integration` module.

### Added

- New method for uploading custom data to blob `Blob.custom_upload()`

### Updated

- Updated the `upsert_device_sensors()` to improve performance and aid release of future functionality.

## 0.1.16

### Added

To the `pipeline` module:

- New method `data_feed_history_process_errors()`, to the `Automation` class.
  - This method returns a dataframe containing the distinct set of error types encountered for a specific
    `data_feed_file_status_id`
- New method `data_feed_history_errors_by_type` , to the `Automation` class.
  - This method returns a dataframe containing the actual errors identified for the specified `error_type` and
    `data_feed_file_status_id`

Additional logging was also incorporated in the backend to support the Switch Platform UI.

### Fixed

- Fixed issue with `register()` method of the `Automation` class in the `pipeline` module.

### Changed

For the `pipeline` module:

- Standardised the following methods of the `Automation` class to return pandas.DataFrame objects.
- Added additional error checks to ensure only allowed values are passed to the various `Automation` class methods
  for the parameters:
  - `expected_delivery`
  - `deploy_type`
  - `queue_name`
  - `error_type`

For the `integration` module:

- Added additional error checks to ensure only allowed values are passed to `post_errors` function for the parameters:
  - `error_type`
  - `process_status`

For the `dataset` module:

- Added additional error check to ensure only allowed values are provided for the `query_language` parameter of the
  `get_data` function.

For the `_platform` module:

- Added additional error checks to ensure only allowed values are provided for the `account` parameter.

## 0.1.14

### Changed

- updated get_device_sensors() to not auto-detect the data type - to prevent issues such as stripping leading zeroes,
  etc from metadata values.

## 0.1.13

### Added

To the `pipeline` module:

- Added a new method, `data_feed_history_process_output`, to the `Automation` class

## 0.1.11

### Changed

- Update to access to `logger` - now available as `switch_api.pipeline.logger()`
- Update to function documentation

## 0.1.10

### Changed

- Updated the calculation of min/max date (for timezone conversions) inside the `upsert_device_sensors` function as
  the previous calculation method will not be supported in a future release of numpy.

### Fixed

- Fixed issue with retrieval of tag groups and tags via the functions:
  - `get_sites`
  - `get_device_sensors`

## 0.1.9

### Added

- New module `platform_insights`

In the `integration` module:

- New function `get_sites` added to lookup site information (optionally with site-level tags)
- New function `get_device_sensors` added to assist with lookup of device/sensor information, optionally including
  either metadata or tags.
- New function `get_tag_groups` added to lookup list of sensor-level tag groups
- New function `get_metadata_keys` added to lookup list of device-level metadata keys

### Changed

- Modifications to connections to storage accounts.
- Additional parameter `queue_name` added to the following methods of the `Automation` class of the `pipeline`
  module:
  - `deploy_on_timer`
  - `deploy_as_email_data_feed`
  - `deploy_as_upload_data_feed`
  - `deploy_as_ftp_data_feed`

### Fixed

In the `pipeline` module:

- Addressed issue with the schema validation for the `upsert_workorders` function

## 0.1.8

### Changed

In the `integrations` module:

- Updated to batch upserts by DeviceCode to improve reliability & performance of the `upsert_device_sensors` function.

### Fixed

In the `analytics` module:

- typing issue that caused error in the import of the switch_api package for python 3.8

## 0.1.7

### Added

In the `integrations` module:

- Added new function `upsert_workorders`
  - Provides ability to ingest work order data into the Switch Automation Platform.
  - Documentation provides details on required & optional fields in the input dataframe and also provides information
    on allowed values for some fields.
  - Two attributes available for function, added to assist with creation of scripts by providing list of required &
    optional fields:
    - `upsert_workorders.df_required_columns`
    - `upsert_workorders.df_optional_columns`
- Added new function `get_states_by_country`:
  - Retrieves the list of states for a given country. Returns a dataframe containing both the state name and
    abbreviation.
- Added new function `get_equipment_classes`:
  - Retrieves the list of allowed values for Equipment Class.
    - EquipmentClass is a required field for the upsert_device_sensors function

### Changed

In the `integrations` module:

- For the `upsert_device_sensors` function:
  - New attributes added to assist with creation of tasks:
    - `upsert_device_sensors.df_required_columns` - returns list of required columns for the input `df`
  - Two new fields required to be present in the dataframe passed to function by parameter `df`:
    - `EquipmentClass`
    - `EquipmentLabel`
  - Fix to documentation so required fields in documentation match.
- For the `upsert_sites` function:
  - New attributes added to assist with creation of tasks:
    - `upsert_sites.df_required_columns` - returns list of required columns for the input `df`
    - `upsert_sites.df_optional_columns` - returns list of required columns for the input `df`
- For the `get_templates` function:
  - Added functionality to filter by type via new parameter `object_property_type`
  - Fixed capitalisation issue where first character of column names in dataframe returned by the function had been
    converted to lowercase.
- For the `get_units_of_measure` function:
  - Added functionality to filter by type via new parameter `object_property_type`
  - Fixed capitalisation issue where first character of column names in dataframe returned by the function had been
    converted to lowercase.

In the `analytics` module:

- Modifications to type hints and documentation for the functions:
  - `get_clone_modules_list`
  - `run_clone_modules`
- Additional logging added to `run_clone_modules`

## 0.1.6

### Added

- Added new function `upsert_timeseries_ds()` to the `integrations` module

### Changed

- Additional logging added to `invalid_file_format()` function from the `error_handlers` module.

### Removed

- Removed `append_timeseries()` function

## 0.1.5

### Fixed

- bug with `upsert_sites()` function that caused optional columns to be treated as required columns.

### Added

Added additional functions to the `error_handlers` module:

- `validate_datetime()` - which checks whether the values of the datetime column(s) of the source file are valid. Any
  datetime errors identified by this function should be passed to the `post_errors()` function.
- `post_errors()` - used to post errors (apart from those identified by the `invalid_file_format()` function) to
  the data feed dashboard.

## 0.1.4

### Changed

Added additional required properties to the Abstract Base Classes (ABC): Task, IntegrationTask, AnalyticsTask,
LogicModuleTask. These properties are:

- Author
- Version

Added additional parameter `query_language` to the `switch.integration.get_data()` function. Allowed values for this
parameter are:

- `sql`
- `kql`

Removed the `name_as_filename` and `treat_as_timeseries` parameter from the following functions:

- `switch.integration.replace_data()`
- `switch.integration.append_data()`
- `switch.integration.upload_data()`


