traveltimes_prediction.configs package¶
Submodules¶
traveltimes_prediction.configs.db_settings module¶
Module db_settings.
Contains the settings of database connections.
Connections used for data sources are following:
- For DET1s, BCKs - connection_string_data_bs3
- For DET2 - connection_string_data_bs2
- For referential TTs - connection_string_tt_real
Connections used as data storages:
- Predictions` results - connection_string_results
- Trained models - connection_string_models
traveltimes_prediction.configs.general_settings module¶
Module general_settings
.
Contain settings about the prediction & training intervals, paths to dumped data, calculator.
-
traveltimes_prediction.configs.general_settings.
all_models
= {<class 'traveltimes_prediction.models.cluster_model.ClusterModel'>, <class 'traveltimes_prediction.models.combined_model.CombinedModel'>, <class 'traveltimes_prediction.models.time_domain_model.TimeDomainModel'>}¶ Allowed models to be used by prediction system
-
traveltimes_prediction.configs.general_settings.
dumped_features_path
= ''¶ Path to dumped features` folder - folder, where the dumped features are stored - to avoid frequent heavy DB queries
-
traveltimes_prediction.configs.general_settings.
max_traveltime
= 7200¶ Max allowed traveltime predicted - used for coercing nonsense values produced by models
-
traveltimes_prediction.configs.general_settings.
prediction_first_run_offset
= 120¶ Offset in seconds from start of the first training, when prediction will be run for the first time
-
traveltimes_prediction.configs.general_settings.
prediction_interval_seconds
= 60¶ Prediction interval in seconds
-
traveltimes_prediction.configs.general_settings.
real_tt_calculator_path
= ''¶ Path to
calculator of real traveltimes
- it is used for the calculation of the ground truth necessary for the training
-
traveltimes_prediction.configs.general_settings.
training_data_interval
= 30¶ Data to be retrieved for training in days
-
traveltimes_prediction.configs.general_settings.
training_interval_seconds
= 86400¶ Training interval in seconds
traveltimes_prediction.configs.sections_settings module¶
Module section_settings.
Contains settings for all used sections, list of sections which should be maintained and definitions of sensors.
Settings of sections are in sections_settings
constant. For each section, multiple attributes has to be specified as follows:
- Input sensors - list of sensors, which are at the beginning of the section.
- Output sensors - list of sensors that are at the end of the section.
- Inner sensors - list of sensor groups, which are between Input and Output. Each group of sensors is located at one slice.
- Forbidden models - set of model classes, that are not used for creation of the models of section.
- Models` config - configurations of the models (dicts)
- Sensor types - types of sensors that are used on this section.
- Timezone - timezone of section -> e.g. ‘Europe/Prague’, ‘Europe/London’ and others.
Sensors` names per each SENSOR_TYPE
are defined in constant sensors
.
-
class
traveltimes_prediction.configs.sections_settings.
SENSOR_TYPE
[source]¶ Bases:
object
Class that defines the sensors` types.
-
SENSOR_DET1
= 'det1'¶
-
SENSOR_DET2
= 'det2'¶
-
-
traveltimes_prediction.configs.sections_settings.
sections_to_maintain
= ['TEST-TEST']¶ List of sections that should be maintained