MPIfR-BDG issueshttps://gitlab.mpcdf.mpg.de/groups/mpifr-bdg/-/issues2024-03-12T10:30:38Zhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/49EDD Fits Simulator Output2024-03-12T10:30:38ZTobias WinchenEDD Fits Simulator OutputProvide a backend simulator for Effelsberg. The simualtor should accepty SCPI input and produce output to the fits writer (possibly generated by np rand.)Provide a backend simulator for Effelsberg. The simualtor should accepty SCPI input and produce output to the fits writer (possibly generated by np rand.)Tobias WinchenTobias Winchenhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/48Zoom Bands in Fits Interface2024-03-12T10:27:06ZTobias WinchenZoom Bands in Fits InterfaceThe user should be able to specify zoom bands, so that the fits interface generates packages with according sections per zoom band.
In a first setup, the bands are specified with channel numbers. In a later iteration (when the front-end...The user should be able to specify zoom bands, so that the fits interface generates packages with according sections per zoom band.
In a first setup, the bands are specified with channel numbers. In a later iteration (when the front-end controller is there), the bands are specified with frequency ranges.Tobias WinchenTobias Winchenhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/47Decouple SubprocessMonitor between parent and child2024-03-01T12:18:32ZTobias WinchenDecouple SubprocessMonitor between parent and childCurrent setup requires that the monitor is instanciated in the child, but the error handler is in the parentCurrent setup requires that the monitor is instanciated in the child, but the error handler is in the parenthttps://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/46[REF] provision_tests.load_inventory and master_controller.get_ansible_invent...2024-02-29T08:43:35ZNiclas Esser[REF] provision_tests.load_inventory and master_controller.get_ansible_inventory has code duplicationBoth functions have the same purpose and can be generalized.Both functions have the same purpose and can be generalized.Niclas EsserNiclas Esserhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/45[BUG] A deconfigure of the pulsar pipeline does not kill "dspsr" when it is i...2024-02-21T10:43:07ZNiclas Esser[BUG] A deconfigure of the pulsar pipeline does not kill "dspsr" when it is in measuring stateWhen the pulsar pipeline is in the state of `measuring` and a `deconfigure` is called, the pipeline is not sufficiently cleaned (e.g. `dspsr` is not killed).
Another re-configure does not solve the issue.
Solution: Call `measurement-stop...When the pulsar pipeline is in the state of `measuring` and a `deconfigure` is called, the pipeline is not sufficiently cleaned (e.g. `dspsr` is not killed).
Another re-configure does not solve the issue.
Solution: Call `measurement-stop` in the the `deconfigure` function
This issue might be related to [issue-20](https://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/20)https://gitlab.mpcdf.mpg.de/mpifr-bdg/edd_dbbc/-/issues/7[BUG] DDC plotter stops streaming spetra to redis2024-02-15T08:50:29ZNiclas Esser[BUG] DDC plotter stops streaming spetra to redisThe DDC plotter stops streaming at some point; the reason is unknown but needs fixing. The data rates are probably too high, so the plotter can not keep up with capturing and processing. An approach is to trash data if the plotter is not...The DDC plotter stops streaming at some point; the reason is unknown but needs fixing. The data rates are probably too high, so the plotter can not keep up with capturing and processing. An approach is to trash data if the plotter is not fast enough.Niclas EsserNiclas Esserhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/edd_dbbc/-/issues/5[BUG] VDIF packer on one NUMA node2024-02-15T07:31:22ZNiclas Esser[BUG] VDIF packer on one NUMA nodeEspecially for low data rates, the VDIFPacker pipeline has a minor resource usage and can run on a single NUMA node. In the [hotfix](https://gitlab.mpcdf.mpg.de/mpifr-bdg/edd_dbbc/-/commit/5efb610fb5c06f91d3412df7567ab9c07aa737a7), the V...Especially for low data rates, the VDIFPacker pipeline has a minor resource usage and can run on a single NUMA node. In the [hotfix](https://gitlab.mpcdf.mpg.de/mpifr-bdg/edd_dbbc/-/commit/5efb610fb5c06f91d3412df7567ab9c07aa737a7), the VDIF packer got the ability to run on a single NUMA node, but the pipeline still spawns two `mkrecv` processes to capture two polarizations independently. It is preferable to capture both polarizations with a single `mkrecv` process because the state transition (measuring -> ready) will reduce significantly (less buffers, less processes) and (most important) `mkrecv` processes won't interfere.Niclas EsserNiclas Esserhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/edd_dbbc/-/issues/4[FEA] FFT-based DDC2024-02-15T07:19:52ZNiclas Esser[FEA] FFT-based DDCCurrently, a Polyphase resampler is used as the base of the DDC. The polyphase resampler allows rational downsampling and is more flexible than an FFT-based downsampler. However, when an integer downsampling is required, an FFT-based can...Currently, a Polyphase resampler is used as the base of the DDC. The polyphase resampler allows rational downsampling and is more flexible than an FFT-based downsampler. However, when an integer downsampling is required, an FFT-based can be significantly faster and provide more independent DDC bands and many frequency-continuous DDC bands.
The DDC processor should auto-select between both approaches if it is not explicitly set.Niclas EsserNiclas Esserhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/edd_dbbc/-/issues/3[FEA] Optimized convolve kernel2024-02-15T07:14:44ZNiclas Esser[FEA] Optimized convolve kernelThe convolve kernel, which convolves the Hilbert filter with the downsampled signal, is a pure time domain convolution leading to O(N*M) costs. Probably an FFT-based convolution is faster for larger data sets O(N * log(M)).The convolve kernel, which convolves the Hilbert filter with the downsampled signal, is a pure time domain convolution leading to O(N*M) costs. Probably an FFT-based convolution is faster for larger data sets O(N * log(M)).Niclas EsserNiclas Esserhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/edd_dbbc/-/issues/2[FEA] Merged channels in single VDIF packet2024-02-15T07:09:21ZNiclas Esser[FEA] Merged channels in single VDIF packetCurrently, the VDIF packer cannot merge multiple channels into one VDIF packet. Hence, for each channel, a vdif_packer + dada buffers are required. It would be more efficient and desirable to pack channels into a single packet.Currently, the VDIF packer cannot merge multiple channels into one VDIF packet. Hence, for each channel, a vdif_packer + dada buffers are required. It would be more efficient and desirable to pack channels into a single packet.Niclas EsserNiclas Esserhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/edd_moonbounce/-/issues/1Moonbounce pipeline goes into an error state on c"configure" with the latest ...2024-02-09T10:11:33ZNiclas EsserMoonbounce pipeline goes into an error state on c"configure" with the latest CICD docker imagehttps://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/44Use async redis for datastore2024-02-05T09:02:02ZTobias WinchenUse async redis for datastorehttps://redis-py.readthedocs.io/en/stable/examples/asyncio_examples.htmlhttps://redis-py.readthedocs.io/en/stable/examples/asyncio_examples.htmlhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/43[FEA] Handle Stokes and DP streams simultaneously in data_plotter.py2024-02-02T20:51:27ZNiclas Esser[FEA] Handle Stokes and DP streams simultaneously in data_plotter.pyWhen both DP and Stokes spectrometer stream to one end-point, the data arrays are messed up![Screenshot_from_2024-02-02_21-50-44](/uploads/70254ba05475eb4b196839404f164838/Screenshot_from_2024-02-02_21-50-44.png)When both DP and Stokes spectrometer stream to one end-point, the data arrays are messed up![Screenshot_from_2024-02-02_21-50-44](/uploads/70254ba05475eb4b196839404f164838/Screenshot_from_2024-02-02_21-50-44.png)Niclas EsserNiclas Esserhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/42[BUG] Full Stokes Spectrometer accepts only a dictionary as input_data_streams2024-02-02T15:58:01ZNiclas Esser[BUG] Full Stokes Spectrometer accepts only a dictionary as input_data_streamsAll pipelines expect a list of dictionaries as `input_data_streams` except the gated_full_stokes_spectrometer.
It is preferred that all pipelines have the same interface for `input_data_streams`. However, changing the it also leads to ad...All pipelines expect a list of dictionaries as `input_data_streams` except the gated_full_stokes_spectrometer.
It is preferred that all pipelines have the same interface for `input_data_streams`. However, changing the it also leads to adjustments in the site-specific edd_provision-repository.Niclas EsserNiclas Esserhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/40Periodic data loss in HDF5 Writer2024-02-02T14:17:09ZTobias WinchenPeriodic data loss in HDF5 Writer![Screenshot_20240129_164227](/uploads/fab00498151124a71867e6bd332780f7/Screenshot_20240129_164227.png)
Periodicity of 10 seconds possibly related to data plotting. Visible in Dualpol+Stokes modus at SKAMPI.![Screenshot_20240129_164227](/uploads/fab00498151124a71867e6bd332780f7/Screenshot_20240129_164227.png)
Periodicity of 10 seconds possibly related to data plotting. Visible in Dualpol+Stokes modus at SKAMPI.Niclas EsserNiclas Esserhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/39[FEA] Handle multiple readers in DbMonitor._stdout_parser()2023-05-31T11:45:58ZNiclas Esser[FEA] Handle multiple readers in DbMonitor._stdout_parser()Currently, the standard output parser of the DbMonitor class (defined in mpikat.utils.dada_tools) can only handle a single reader.Currently, the standard output parser of the DbMonitor class (defined in mpikat.utils.dada_tools) can only handle a single reader.https://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/38Watchdog limits spectrometer integration tiome to 10 sec2023-02-21T13:12:38ZTobias WinchenWatchdog limits spectrometer integration tiome to 10 secThe watchdog should consider the integration time to allow longer integrations if neededThe watchdog should consider the integration time to allow longer integrations if neededhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/37[FEA] Refactor mkrecv_sensor2023-02-17T16:29:15ZNiclas Esser[FEA] Refactor mkrecv_sensorRefactoring the `mkrecv_stdout_parser.py` to contain all sensor values.Refactoring the `mkrecv_stdout_parser.py` to contain all sensor values.https://gitlab.mpcdf.mpg.de/mpifr-bdg/edd_dbbc/-/issues/1[FEA] Create a bandflip feature2023-02-09T10:12:26ZNiclas Esser[FEA] Create a bandflip featureThe lower sideband is flipped and potentially needs to be flipped back.The lower sideband is flipped and potentially needs to be flipped back.Niclas EsserNiclas Esserhttps://gitlab.mpcdf.mpg.de/mpifr-bdg/mpikat/-/issues/35Ensure correct dada key generation2023-10-16T12:49:00ZNiclas EsserEnsure correct dada key generationThe function `occupied()` in `mpikat/utils/dada_tools/ringbuffer.py` has an optional argument `exclude=set()`. It must be None.The function `occupied()` in `mpikat/utils/dada_tools/ringbuffer.py` has an optional argument `exclude=set()`. It must be None.