Shepherd - Datalib

This Python Module eases the handling of hdf5-recordings used by the shepherd-testbed. Users can read, validate and create files and also extract, down-sample and plot information.

Installation

Manual

• clone repository
• navigate shell into directory
• activate and update pipenv (optional)
• install module

cd .\shepherd-datalib

pipenv shell
pipenv update

python setup.py install

Via PIP (TODO)

pip3 install shepherd_data

Programming Interface

Basic Usage (recommendation)

import shepherd_data as shpd

with shpd.Reader("./hrv_sawtooth_1h.h5") as db:
    print(f"Mode: {db.get_mode()}")
    print(f"Window: {db.get_window_samples()}")
    print(f"Config: {db.get_config()}")

Available Functionality

• Reader()
  • file can be checked for plausibility and validity (is_valid())
  • internal structure of h5file (get_metadata() or save_metadata() ... to yaml) with lots of additional data
  • access data and various converters, calculators
    • read_buffers() -> generator that provides one buffer per call, can be configured on first call
    • get_calibration_data()
    • get_windows_samples()
    • get_mode()
    • get_config()
    • direct access to root h5-structure via reader['element']
    • converters for raw / physical units: si_to_raw() & raw_to_si()
    • energy() sums up recorded power over time
  • downsample() (if needed) visualize recording (plot_to_file())
• Writer()
  • inherits all functionality from Reader
  • append_iv_data_raw()
  • append_iv_data_si()
  • set_config()
  • set_windows_samples()
• IVonne Reader
  • convert_2_ivcurves() converts ivonne-recording into a shepherd ivcurve
  • upsample_2_isc_voc() TODO: for now a upsampled but unusable version of samples of short-circuit-current and open-circuit-voltage
  • convert_2_ivsamples() already applies a simple harvesting-algo and creates ivsamples
• examples
  • example_convert_ivonne.py converts IVonne recording (jogging_10m.iv) to shepherd ivcurves, NOTE: slow implementation
  • example_extract_logs.py is analyzing all files in directory, saves logging-data and calculates cpu-load and data-rate
  • example_generate_sawtooth.py is using Writer to generate a 60s ramp with 1h repetition and uses Reader to dump metadata of that file
  • example_plot_traces.py demos some mpl-plots with various zoom levels
  • example_repair_recordings.py makes old recordings from shepherd 1.x fit for v2
  • jogging_10m.iv
    • 50 Hz measurement with Short-Circuit-Current and two other parameters
    • recorded with "IVonne"

CLI-Interface

After installing the module the datalib offers some often needed functions:

Validate Recordings

• takes a file or directory as an argument

shepherd-data validate dir_or_file

# examples:
shepherd-data validate ./
shepherd-data validate hrv_saw_1h.h5

Extract IV-Samples to csv

• takes a file or directory as an argument
• can take down-sample-factor as an argument

shepherd-data extract dir_or_file [-f ds_factor] [-s separator_symbol]

# examples:
shepherd-data extract ./
shepherd-data extract hrv_saw_1h.h5 -f 1000 -s;

Extract meta-data and sys-logs

• takes a file or directory as an argument

shepherd-data extract-meta dir_or_file

# examples:
shepherd-data extract-meta ./
shepherd-data extract-meta hrv_saw_1h.h5

Plot IVSamples

• takes a file or directory as an argument
• can take start- and end-time as an argument
• can take image-width and -height as an argument

shepherd-data plot dir_or_file [-s start_time] [-e end_time] [-w plot_width] [-h plot_height] [--multiplot]

# examples:
shepherd-data plot ./ --multiplot
shepherd-data plot hrv_saw_1h.h5 -s10 -e20

Downsample IVSamples (for later GUI-usage, TODO)

• generates a set of downsamplings (20 kHz to 0.1 Hz in x4 to x5 Steps)
• takes a file or directory as an argument
• can take down-sample-factor as an argument

shepherd-data downsample dir_or_file [-f ds_factor]

# examples:
shepherd-data downsample ./ 
shepherd-data downsample hrv_saw_1h.h5 -f 1000

Data-Layout and Design choices

Modes and Datatypes

• Mode harvester recorded a harvesting-source like solar with one of various algorithms
  • Datatype ivsample is directly usable by shepherd, input for virtual source / converter
  • Datatype ivcurve is directly usable by shepherd, input for a virtual harvester (output are ivsamples)
  • Datatype isc_voc is specially for solar-cells and needs to be (at least) transformed into ivcurves later
• Mode emulator replayed a harvester-recording through a virtual converter and supplied a target while recording the power-consumption
  • Datatype ivsample is the only output of this mode

Compression & Beaglebone

• supported are uncompressed, lzf and gzip with level 1 (order of recommendation)
  • lzf seems better-suited due to lower load, or if space isn't a constraint: uncompressed (None as argument)
  • note: lzf seems to cause trouble with some third party hdf5-tools
  • compression is a heavy load for the beaglebone, but it got more performant with recent python-versions
• size-experiment A: 24 h of ramping / sawtooth (data is repetitive with 1 minute ramp)
  • gzip-1: 49'646 MiB -> 588 KiB/s
  • lzf: 106'445 MiB -> 1262 KiB/s
  • uncompressed: 131'928 MiB -> 1564 KiB/s
• cpu-load-experiments (input is 24h sawtooth, python 3.10 with most recent libs as of 2022-04)
  • warning: gpio-traffic and other logging-data can cause lots of load

  emu_120s_gz1_to_gz1.h5 	-> emulator, cpu_util [%] = 65.59, data-rate =  352.0 KiB/s
  emu_120s_gz1_to_lzf.h5 	-> emulator, cpu_util [%] = 57.37, data-rate =  686.0 KiB/s
  emu_120s_gz1_to_unc.h5 	-> emulator, cpu_util [%] = 53.63, data-rate = 1564.0 KiB/s
  emu_120s_lzf_to_gz1.h5 	-> emulator, cpu_util [%] = 63.18, data-rate =  352.0 KiB/s
  emu_120s_lzf_to_lzf.h5 	-> emulator, cpu_util [%] = 58.60, data-rate =  686.0 KiB/s
  emu_120s_lzf_to_unc.h5 	-> emulator, cpu_util [%] = 55.75, data-rate = 1564.0 KiB/s
  emu_120s_unc_to_gz1.h5 	-> emulator, cpu_util [%] = 63.84, data-rate =  351.0 KiB/s
  emu_120s_unc_to_lzf.h5 	-> emulator, cpu_util [%] = 57.28, data-rate =  686.0 KiB/s
  emu_120s_unc_to_unc.h5 	-> emulator, cpu_util [%] = 51.69, data-rate = 1564.0 KiB/s 

Open Tasks

• implementations for this lib
  • generalize up- and down-sampling, use out_sample_rate instead of ds-factor
    • lib samplerate (tested) -> promising, but designed for float32 and range of +-1.0
    • lib resampy (tested) -> could be problematic with slice-iterator
    • https://stackoverflow.com/questions/29085268/resample-a-numpy-array
    • scipy.signal.resample, https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.resample.html
    • scipy.signal.decimate, https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.decimate.html
    • scipy.signal.resample_poly, https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.resample_poly.html#scipy.signal.resample_poly
    • timestamps could be regenerated with np.arange( tmin, tmax, 1e9/samplerate)
  • generalize converters (currently in IVonne)
    • isc&voc <-> ivcurve
    • ivcurve -> ivsample
  • plotting and downsampling for IVCurves ()
  • plotting more generalized (power, cpu-util, ..., if IV then offer power as well)
  • some metadata is calculated wrong (non-scalar datasets)
  • unittests
• main shepherd-code
  • proper validation first
  • update commentary
  • pin-description should be in yaml (and other descriptions for cpu, io, ...)
  • datatype-hint in h5-file (ivcurve, ivsample, isc_voc), add mechanism to prevent misuse
  • hostname for emulation
  • full and minimal config into h5
  • use the datalib as a base
  • isc-voc-harvesting
  • directly process isc-voc -> resample to ivcurve?