Proto commits in RespiraWorks/Ventilator

These 43 commits are when the Protocol Buffers files have changed:

2022-08-11

Automated generation of communications protocol bindings "fixed bad link and updated dependencies; updates #1263" "dependencies; updates #1263" "dependencies; updates #1263" "script should work now; updates #1263" "show pip locations; updates #1263" "get dir a different way; updates #1263" "get dir a different way; updates #1263" "make dir for python lib; updates #1263" "simplify; updates #1263" "controller and gui should build; updates #1263" "controller and gui should build; updates #1263" fix gui common.cmake "gui should build now, and coverage fixes; updates #1263" "compactified comments in proto definition, updated precommit versions; updates #1263" "fixed proto autoformatting, enabled python filename checking; updates #1263" "updated dragged-in nanopb code; updates #1263" Update software/common/README.md Co-authored-by: Jérôme Travert <asmodai27@users.noreply.github.com> "streamlined installation and made protocol generation automatic; updates #1263" "optional conversion of enum values; updates #1263" "some documentation updates and todo; updates #1263"

The documentation is generated from this commit.

2022-07-15

proto reformat

Added new modes in firmware & in GUI, tetsed basically. Committing to git for backup purpose

2020-08-12

Add oxygen sensor support Add unit tests and fix broken ones change dbg_fio2 to 0-1 instead of 0-100% Adress @Martukas review: - Apply Google style for constant names in the files I am modifying - use enum class for Sensor enum - add O2 sensor datasheet More Google style formatting of constants added doc as source for 0.06V/fio2 sensor sensitivity (also changed from 0.061V/fio2 to match doc). Note, this is a stand in value for sensitivity until we have a self-test calibration procedure. This is captured in issue https://github.com/RespiraWorks/Ventilator/issues/859 fix unit test with consistent O2 sensor sensitivity Feed fio2 to the GUI small change to coverage script

Add oxygen sensor support Add unit tests and fix broken ones change dbg_fio2 to 0-1 instead of 0-100% Adress @Martukas review: - Apply Google style for constant names in the files I am modifying - use enum class for Sensor enum - add O2 sensor datasheet More Google style formatting of constants added doc as source for 0.06V/fio2 sensor sensitivity (also changed from 0.061V/fio2 to match doc). Note, this is a stand in value for sensitivity until we have a self-test calibration procedure. This is captured in issue https://github.com/RespiraWorks/Ventilator/issues/859 fix unit test with consistent O2 sensor sensitivity Feed fio2 to the GUI small change to coverage script

2020-07-20

Adds consolidated CoVent links to root README Puts software in its own folder; Updates root README Moves system-design folder contents into repo root adds pre-commit formatting Fixes borked links fixes more borked links Converts PCBreathe & SysDesign repo links Changes build to assembly to avoid some sort of conflict Fixes borked links Adds link to status report Covent doc Separates requirements & moves status report to design folder changes assemlby to manufacturing forgot to save a thing fixed submodule issue and updated paths in circleci script moved controller test script and updated dirs more dir updates in scripts updte controller tests docker script dirs cleaned up main readme typos, badges and other readme updates moved platformio.ini and .clang-tidy updated references after repository rename/merge and references to tickets added screenshot of GUI to software docs

Adds consolidated CoVent links to root README Puts software in its own folder; Updates root README Moves system-design folder contents into repo root adds pre-commit formatting Fixes borked links fixes more borked links Converts PCBreathe & SysDesign repo links Changes build to assembly to avoid some sort of conflict Fixes borked links Adds link to status report Covent doc Separates requirements & moves status report to design folder changes assemlby to manufacturing forgot to save a thing fixed submodule issue and updated paths in circleci script moved controller test script and updated dirs more dir updates in scripts updte controller tests docker script dirs cleaned up main readme typos, badges and other readme updates moved platformio.ini and .clang-tidy updated references after repository rename/merge and references to tickets added screenshot of GUI to software docs

2020-06-19

Send flow correction ml/s to GUI. Will be used for disconnect detection. I observe with the fan that a leak is about -1600ml/s. Even when I set PIP/PEEP to 5, we still open up the valve all the way. I think it would be safe to set the disconnect detection threshold around -500ml/s. GPC: flow-correction-in-proto

Send flow correction ml/s to GUI. Will be used for disconnect detection. I observe with the fan that a leak is about -1600ml/s. Even when I set PIP/PEEP to 5, we still open up the valve all the way. I think it would be safe to set the disconnect detection threshold around -500ml/s. GPC: flow-correction-in-proto

2020-06-18

Allows setting FiO2 % from GUI

Allows setting FiO2 % from GUI

Adds FiO2 to the proto

Adds FiO2 to the proto

2020-06-15

Changes ventilation mode property to be an enum. Adds HFNC mode to the proto as a formality.

Changes ventilation mode property to be an enum. Adds HFNC mode to the proto as a formality.

Change breath_id to uint64 based on uptime

Change breath_id to uint64 based on uptime

Adds breath_id to network protocol

Adds breath_id to network protocol

2020-06-14

Removes alarms from network protocol

Removes alarms from network protocol

Delete unused VentParams.rise_time_ms. I didn't realize that we had this param and, in an earlier change, hardcoded it to 100. I guess that's...fine? We don't expect to change it from the GUI. GPC: delete-rise-time

Delete unused VentParams.rise_time_ms. I didn't realize that we had this param and, in an earlier change, hardcoded it to 100. I guess that's...fine? We don't expect to change it from the GUI. GPC: delete-rise-time

ControllerStatus.fan_setpoint -> pressure_setpoint. Simple rename, no functional change. GPC: fan-setpoint

ControllerStatus.fan_setpoint -> pressure_setpoint. Simple rename, no functional change. GPC: fan-setpoint

2020-06-13

First pass at Pressure Assist Mode. Note: updated proto file but also updated network_protocol.pb.h manually (couldn't get nanopb to run...)

First pass at Pressure Assist Mode. Note: updated proto file but also updated network_protocol.pb.h manually (couldn't get nanopb to run...)

2020-06-12

Rename SensorReadings proto -> SensorsProto. Preparation for reclaiming the name "SensorReadings" as a C++ class. This is inconsistent with the names of the other proto messages, but given that we're going to have this C++ class, I think it will be really hard to remember which is which if one of them doesn't include the word "proto". GPC: rename-sensor-readings

Rename SensorReadings proto -> SensorsProto. Preparation for reclaiming the name "SensorReadings" as a C++ class. This is inconsistent with the names of the other proto messages, but given that we're going to have this C++ class, I think it will be really hard to remember which is which if one of them doesn't include the word "proto". GPC: rename-sensor-readings

2020-05-14

Tweak SensorReadings proto. - Rename pressure_cm_h2o -> patient_pressure_cm_h2o - Add {inflow,outflow}_pressure_cm_h2o. These values are primarily useful for debugging, but it's fine to send them everywhere. - Tweak the DEV_MODE serial printf. Now it's a series of printfs, so that it's easy to comment one out. Volume is commented out in the printf because without something to fix the linear component of the volume, it quickly overwhelms the other values in the graph, rendering everything else out-of-scale. GPC: sensor-readings

Tweak SensorReadings proto. - Rename pressure_cm_h2o -> patient_pressure_cm_h2o - Add {inflow,outflow}_pressure_cm_h2o. These values are primarily useful for debugging, but it's fine to send them everywhere. - Tweak the DEV_MODE serial printf. Now it's a series of printfs, so that it's easy to comment one out. Volume is commented out in the printf because without something to fix the linear component of the volume, it quickly overwhelms the other values in the graph, rendering everything else out-of-scale. GPC: sensor-readings

2020-04-29

New blower FSM. The moment we've all been waiting for... - Create blower_fsm, which separates out controlling the pressure setpoint from the PID loop that drives the fan to achieve this setpoint. - Rename control to blower_pid now that its role is smaller. - Add setpoint field to ControllerStatus proto so the GUI can observe the pressure we're trying to drive. Fixes issue #17. GPC: blower-fsm

New blower FSM. The moment we've all been waiting for... - Create blower_fsm, which separates out controlling the pressure setpoint from the PID loop that drives the fan to achieve this setpoint. - Rename control to blower_pid now that its role is smaller. - Add setpoint field to ControllerStatus proto so the GUI can observe the pressure we're trying to drive. Fixes issue #17. GPC: blower-fsm

2020-04-28

Remove some unused options in network_protocol.proto. - Remove FiO2 control. Ethan said that he anticipates FiO2 will be controlled externally from the ventilator. - Remove PRESSURE_ASSIST / PRESSURE_SUPPORT ventilation modes. I am writing code that wants to switch() on ventilator modes, and getting rid of the unused modes means the switch() can cover all modes without a `default` case. That way when we add a new mode, we'll get a compile error showing everything we have to change. GPC: protobuf-unused

Remove some unused options in network_protocol.proto. - Remove FiO2 control. Ethan said that he anticipates FiO2 will be controlled externally from the ventilator. - Remove PRESSURE_ASSIST / PRESSURE_SUPPORT ventilation modes. I am writing code that wants to switch() on ventilator modes, and getting rid of the unused modes means the switch() can cover all modes without a `default` case. That way when we add a new mode, we'll get a compile error showing everything we have to change. GPC: protobuf-unused

2020-04-22

Resurrects clang-tidy excluding generated code. Moves generated code to common/generated_libs. Fixes up some clang-tidy warnings that popped up while it was disabled.

Resurrects clang-tidy excluding generated code. Moves generated code to common/generated_libs. Fixes up some clang-tidy warnings that popped up while it was disabled.

Send and receive only a single proto message. Switch from a command-based system to a system where each side periodically sends its status to the other. This has a few benefits. - No longer do we have to worry about packets arriving out of order (e.g. due to the first one having a checksum failure) and causing an invalid vent state. - No longer do we have to handle ACKs and retransmits. - We can have more confidence that if we change the GUI, it won't affect the controller, because in all cases, the GUI is sending down exactly the same messages to the controller. - Easier to test, since more of the state of the controller is encapsulated in this one struct. - Easier to add a new parameter -- just add it to the struct, instead of having to add getters/setters for it. - Simpler state machine: Now modules like pid just read whatever is the current state, rather than having to update their internal state whenever something changes. This is the culmination of a lot of work by - Lee Matthews - Albertas Mickėnas - Sarah Evans - Eugene Kirpichov - verityRF - and others

Send and receive only a single proto message. Switch from a command-based system to a system where each side periodically sends its status to the other. This has a few benefits. - No longer do we have to worry about packets arriving out of order (e.g. due to the first one having a checksum failure) and causing an invalid vent state. - No longer do we have to handle ACKs and retransmits. - We can have more confidence that if we change the GUI, it won't affect the controller, because in all cases, the GUI is sending down exactly the same messages to the controller. - Easier to test, since more of the state of the controller is encapsulated in this one struct. - Easier to add a new parameter -- just add it to the struct, instead of having to add getters/setters for it. - Simpler state machine: Now modules like pid just read whatever is the current state, rather than having to update their internal state whenever something changes. This is the culmination of a lot of work by - Lee Matthews - Albertas Mickėnas - Sarah Evans - Eugene Kirpichov - verityRF - and others

2020-04-21

clean up comments, variable and type names add small fixes related to review - comments, var names and types rename ack_handler, improve command message decoder, fix bugs related to it where message length was incorrectly specified convert comment to TODO apply formatting rules to the proto file add comments explaining Commands VentilatorController handles rename AlertLevel type, also add more comments on usage of different data packets pass objects by reference to command and gui ack handlers. Demote command handlers to comms library Consolidate all serdes code to a single file pull ControllerStatus and network-protocol dependency up to pid.c. migrate tests to gtest

clean up comments, variable and type names add small fixes related to review - comments, var names and types rename ack_handler, improve command message decoder, fix bugs related to it where message length was incorrectly specified convert comment to TODO apply formatting rules to the proto file add comments explaining Commands VentilatorController handles rename AlertLevel type, also add more comments on usage of different data packets pass objects by reference to command and gui ack handlers. Demote command handlers to comms library Consolidate all serdes code to a single file pull ControllerStatus and network-protocol dependency up to pid.c. migrate tests to gtest

Changes communication protocol to use Nanopb

Changes communication protocol to use Nanopb