`ocd.storage` – Storage classes¶

This module contains the classes that act as a storage for data needed by OCD to take decisions.

class ocd.storage.OCDVault(conf)[source]¶

Bases: object

The Vault is handed events and knows which one it can store and where. Each variable is stored into a Container instance, or child thereof, and all the instances are exposed as attributes.

Parameters:	conf : `pyhetdex.tools.configuration.ConfigParser` configuration object
Attributes:	conf same as input `topics` Returns the list of topics the vault can handle `container_names` Return the list of container names

create_containers(self)[source]¶

Generator that yields the containers, their names, the topics they handlers and the topic handlers.

Yields:	container_name : string name of the container container : `Container` container that stores the information topic : string name of the topic handled by the container handler : callable function that handle the topic. Typically is `ContainerEventMeta.handle_event()`

store_event(self, topic, event)[source]¶

If the topic is relevant, call the functions/methods to store the event.

Parameters:	topic : string topic of the event event : dict event to handle
Returns:	bool whether the event has been handled or not

topics¶: Returns the list of topics the vault can handle

container_names¶: Return the list of container names

class ocd.storage.MetrologyVault(conf)[source]¶

Bases: ocd.storage.OCDVault

Vault containing metrology data.

Examples

>>> from ocd.config import load_config
>>> from ocd import tcs_proxy
>>> conf = load_config()
>>> tcs_proxy.init(conf)
>>> ocd_vault = MetrologyVault(conf)
>>> ocd_vault.store_event('pas.Guider1.metrology_data',
...                       {'fit.gauss_par(3)': 42,
...                        'fit.gauss_par(4)': 42,
...                        'filter.magnitude': 42,
...                        'photometry.fixed_skymag': 42,
...                        'photometry.kron_skymag': 42,
...                        'photometry.fixed_mag': 42,
...                        'photometry.kron_mag': 42,
...                        'image_quality': 1.2,
...                        'transparency': 0.8,
...                        'plate_scale.x': 0.196,
...                        'plate_scale.y': 0.196,
...                        '__data_time': 43, })
True
>>> # get the Container instance for the fwhm quantity
>>> fwhm_container1 = ocd_vault.fwhm_g1
>>> fwhm_container2 = ocd_vault['fwhm_g1']
>>> fwhm_container1 == fwhm_container2
True

All the relevant metrology values from the pas.Guider{1,2}.metrology_data are stored, however if the values of any of the following keys is true the value is masked:

photometry.object_at_image_border
photometry.object_in_bad_image_region
photometry.star_ambiguous
photometry.star_not_found
photometry.unreliable_background

Todo

fix the transparency_g* description when the missing documentation arrive. See also ContainerTransparency.

Attributes:

fwhm_g1, fwhm_g2 : ContainerFWHM: converts the values of object.fwhm to arcsec using plate_scale.x and plate_scale.y from the pas.Guider1.metrology_data and pas.Guider2.metrology_data event respectively and store it
skymag_g1, skymag_g2 : ContainerEventKey: store the values of the sky magnitude from the pas.Guider1.metrology_data and pas.Guider2.metrology_data event respectively. The name of the key is defined by the photometry_skymag option of the [containers] section of the configuration
transparency_g1, transparency_g2 : ContainerTransparency: compute and store the transparency from the pas.Guider1.metrology_data and pas.Guider2.metrology_data event respectively. It uses the star magnitude, filter.magnitude and illumination correction keywords. The name of the star magnitude is defined by the photometry_trans option of the [containers] section of the configuration; the illumination correction is set to 0.86 and can be modified using the illumination_correction option.

create_containers(self)[source]¶

Generator that yields the containers, their names, the topics they handlers and the topic handlers.

Yields:	container_name : string name of the container container : `Container` container that stores the information topic : string name of the topic handled by the container handler : callable function that handle the topic. Typically is `ContainerEventMeta.handle_event()`

_container_event_key(self, event_key, mask_func=None)[source]¶

Instantiate a ContainerEventKey and returns it.

Parameters:	event_key : string name of the key in the event whose value must be stored mask_func : callable, optional mask function to pass to the ContainerEventKey; defaults to `None`
Returns:	container : `ContainerEventKey`

class ocd.storage.AzimuthVault(conf)[source]¶

Bases: ocd.storage.OCDVault

Vault for storing azimuth.

Attributes:

azimuth : ContainerSkipMasked: store the values of az, if the setup is done ("true"), from the tcs.root.ra_dec event. Only a few azimuth values are retained by OCD, as only the latest one is used.
azimuth_tmp : ContainerSkipMasked: store the values of az, if the setup is not done ("false"), from the tcs.root.ra_dec event. Only a few azimuth values are retained by OCD, as only the latest one is used. This can be used as a fallback in case the azimuth is empty.

create_containers(self)[source]¶

Generator that yields the containers, their names, the topics they handlers and the topic handlers.

Yields:	container_name : string name of the container container : `Container` container that stores the information topic : string name of the topic handled by the container handler : callable function that handle the topic. Typically is `ContainerEventMeta.handle_event()`

class ocd.storage.Container(maxlen=50, delta_timestamp=None)[source]¶

Bases: object

Container to store a list of n values, their time stamps and their mask. Quantities like mean, median and standard deviation are exposed as properties.

Warning

time stamps are assumed to be seconds “since the epoch” as returned e.g. by time.time(). All the tests are run assuming this.

Parameters:	maxlen : int, optional maximum length of the underlying storage delta_timestamp : number, optional maximum time difference for the values stored in the class in seconds. It must be a positive number or `None` When adding a new value, discard all the values older than `delta_timestamp` with respect to the value timestamp; if `None` the timestamp is ignored
Raises:	ocd.errors.OCDTypeError if `delta_timestamp` is not a number ocd.errors.OCDValueError if `delta_timestamp` is negative
Attributes:	`latest` Returns the last value added. `latest_masked` Returns the last non masked value. `mean` Return the mean of the stored values or `NaN` `median` Return the mean of the stored values or `NaN` `stddev` Return the standard deviation of the stored values or `NaN` if no `mean_masked` Return the mean of the non masked stored values or `NaN` `median_masked` Return the median of the non masked stored values or `NaN` if no `stddev_masked` Return the standard deviation of the non masked stored values or

add_value(self, value, timestamp, mask=False)[source]¶

Store the new value and time stamp. If a delta_timestamp is given any entry older than timestamp - delta_timestamp are removed

The code assumes that the values are added with increasing time stamp.

Parameters:	value : number new value to store timestamp : float timestamp in seconds associated to the value mask : bool, optional whether the value is to be masked or not. It follows the convention of numpy masked arrays. By default no value is masked

_masked_values¶: Returns a ndarray representation containing only the non masked values.

mean¶: Return the mean of the stored values or NaN if no value is present

median¶: Return the mean of the stored values or NaN if no value is present

stddev¶: Return the standard deviation of the stored values or NaN if no value is present

mean_masked¶: Return the mean of the non masked stored values or NaN if no value is present or the mask is all True

median_masked¶: Return the median of the non masked stored values or NaN if no value is present or the mask is all True

stddev_masked¶: Return the standard deviation of the non masked stored values or NaN if no value is present or the mask is all True

latest¶: Returns the last value added. Raises an AttributeError if no value is found.

latest_masked¶: Returns the last non masked value. Raises an AttributeError if no value is found.

class ocd.storage.ContainerEventMeta(*args, **kwargs)[source]¶

Bases: ocd.storage.Container

ABC class that extend the Container class adding a method to handle a TCS event and store a value derived from it.

Derived classes must reimplement the value_from_event().

Parameters:

time_key : string, optional: name of the keyword containing the time stamp. Default: __data_time
mask_func : callable, optional: function that receives the event passed to handle_event() as input and should return True if the value is to be masked, False otherwise. If not given or None, no value will be masked
args, kwargs :: positional and keyword parameters passed to the parent class

handle_event(self, event)[source]¶

Call value_from_event() to extract a value from the event, get the time stamp and, if required, the mask and save it.

Parameters:	event : dict event to handle

value_from_event(self, event)[source]¶

Extract one or more values from the event and do the necessary calculations. Returns a single value as result, typically a number.

Implement this function in derived classes

Parameters:	event : dict event to handle
Returns:	result : float or other value

class ocd.storage.ContainerEventKey(key, *args, **kwargs)[source]¶

Bases: ocd.storage.ContainerEventMeta

Implementation of ContainerEventMeta class that extract and save a single value for a TCS event.

Parameters:	key : string when handling the event, get the value for the given key and store it args, kwargs : positional and keyword parameters passed to the parent class

value_from_event(self, event)[source]¶: Extract the value of key from the event and return it.

class ocd.storage.ContainerTransparency(conf, *args, **kwargs)[source]¶

Bases: ocd.storage.ContainerEventMeta

Implementation of ContainerEventMeta class that computes the transparency from a TCS event.

On top of the masking described in MetrologyVault, this container also masks negative values of the filter.magnitude value, on the assumption that if no guiding nor photometry is performed on stars with negative magnitude and that the “unset” value is -9.999.

The illumination correction has been included in the guider fotometry.

Warning

We set the default illumination correction to 1. If necessary this value can be overridden with the illumination_correction option of the [containers] configuration option. This might be eventually removed.

Parameters:	conf : `pyhetdex.tools.configuration.ConfigParser` configuration object args, kwargs : same as in `ContainerEventMeta`

value_from_event(self, event)[source]¶

Calculate the transparency as:

\[\frac{10 ^ {-0.4 * (mag - filter\_mag)}}{illum\_corr}\]

where

mag is the measured magnitude of the star and is stored in the key defined by the photometry_trans option of the [containers] section;
filter_mag is the reference magnitude of the star, as communicated to TCS, and is stored in the filter.magnitude key;
illum_corr is the illumination correction for the current position of the tracker; it is set to 1 or the value of the illumination_correction option of the [containers] section;

Parameters:	event : dict event to handle
Returns:	transparency : float

class ocd.storage.ContainerFWHM(conf, *args, **kwargs)[source]¶

Bases: ocd.storage.ContainerEventMeta

Implementation of ContainerEventMeta class that computes the FWHM in arcseconds as:

\[\sqrt{(\sigma_x * pscale_x) * (\sigma_y * pscale_y)}\]

where the sigma_x and sigma_y are the x and y standard deviation, in pixels, from the Gaussian fit of the guide star profile (fit.gauss_par(3) and fit.gauss_par(3)) and pscale_x and pscale_y are the pixels scale (plate_scale.x and plate_scale.y).

If either fit.gauss_par(3) or fit.gauss_par(4) is zero or negative, the value of fwhm is masked.

Parameters:	conf : `pyhetdex.tools.configuration.ConfigParser` configuration object args, kwargs : same as in `ContainerEventMeta`

value_from_event(self, event)[source]¶

Calculate the fwhm in arcseconds.

Parameters:	event : dict event to handle
Returns:	FWHM : float

class ocd.storage.ContainerSkipMasked(key, *args, **kwargs)[source]¶

Bases: ocd.storage.ContainerEventKey

Container to store only values that are not masked.

add_value(self, *args, **kwargs)[source]¶

This method calls Container.add_value() if mask is False. Otherwise does nothing

Parameters:	all : see `Container.add_value()`

Inheritance scheme¶

ocd.storage – Storage classes¶

Inheritance scheme¶

`ocd.storage` – Storage classes¶