Description
Filtering before csv write using an EMAP query would reduce disk and CPU usage on the GAE, and is one possible implementation of #79
Norad status is a flowsheet in EMAP, so it could be used for this.
Will also need ventilator status which could be inferred (with difficulty, see comments) from recent interchange messages.
Please note #15 as another, less painful way to reduce disk usage. These two methods would be additive if both implemented.
Definition of Done
Dependencies
Comments
I would rather only do this if absolutely necessary. It is going to be fiddly and creates a lot of different combinations to test.
The waveform controller handles incoming interchange messages one by one, and currently the only way to look at what messages have been recently received is to look at the just-written CSVs. Unfortunately, incoming message order is not completely guaranteed, so you could receive some ECG data and discard it before you receive the message telling you that the patient is ventilated.
These difficulties are probably fixable if we really need to do it this way, but it's not the path of least resistance.
If the goal is reducing disk usage I would suggest making the CSVs smaller, which would be much easier since they're very redundant as things stand.
Description
Filtering before csv write using an EMAP query would reduce disk and CPU usage on the GAE, and is one possible implementation of #79
Norad status is a flowsheet in EMAP, so it could be used for this.
Will also need ventilator status which could be inferred (with difficulty, see comments) from recent interchange messages.
Please note #15 as another, less painful way to reduce disk usage. These two methods would be additive if both implemented.
Definition of Done
Dependencies
Comments
I would rather only do this if absolutely necessary. It is going to be fiddly and creates a lot of different combinations to test.
The waveform controller handles incoming interchange messages one by one, and currently the only way to look at what messages have been recently received is to look at the just-written CSVs. Unfortunately, incoming message order is not completely guaranteed, so you could receive some ECG data and discard it before you receive the message telling you that the patient is ventilated.
These difficulties are probably fixable if we really need to do it this way, but it's not the path of least resistance.
If the goal is reducing disk usage I would suggest making the CSVs smaller, which would be much easier since they're very redundant as things stand.