It is useful to analyze injections with Dingo, especially because it is so fast. However, this has not yet been implemented in dingo_pipe. The goal of this PR is to accomplish this.
:innocent: Implemented Features:
Allows specification of injection-parameters in the settings.ini file.
Allows one to change the waveform approximant which is being injected from the one used for training the network
Allows specification of PSD from trigger-time
Allows specification of random seed for noise realization
:smiling_face_with_tear: TODO:
Implement Zero-Noise option
Check that the asd-dataset and trigger-time can't both be specified at the same time. As only one should be used to specify the PSD
For this^: I see we can specify multiple trigger times via: https://github.com/dingo-gw/dingo/blob/main/dingo/pipe/dag_creator.py#L49 which will generate multiple dags for data generation and sampling. I am wondering if this could be a simple way to iterate through multiple different noise realizations and at the end we can merge the result but I'm not quite sure what appending things to the generation_node_list does if it's different data.
It is useful to analyze injections with Dingo, especially because it is so fast. However, this has not yet been implemented in dingo_pipe. The goal of this PR is to accomplish this.
:innocent: Implemented Features:
:smiling_face_with_tear: TODO:
For this^: I see we can specify multiple trigger times via: https://github.com/dingo-gw/dingo/blob/main/dingo/pipe/dag_creator.py#L49 which will generate multiple dags for data generation and sampling. I am wondering if this could be a simple way to iterate through multiple different noise realizations and at the end we can merge the result but I'm not quite sure what appending things to the generation_node_list does if it's different data.