The KF tracking dump reader now follows the parameters dumped by kf_tracking.cc
lib/gps_l1_ca_kf_read_tracking_dump.py:
- Rewrite using the same _RECORD_FORMAT / _FIELD_NAMES architecture as
dll_pll_veml_read_tracking_dump.py: one struct.unpack per record and
EOF handling, dropping the v1..v22 / bytes_shift / seek.
gps_l1_ca_kf_plot_sample.py:
- map innovation -> carr_error; drop r_noise_cov
lib/plotKalman.py, lib/plotTracking.py:
- Skip the "Estimated Noise Variance" panel when r_noise_cov is absent.
- Replace fig.canvas.set_window_title() with
fig.canvas.manager.set_window_title(). The canvas method was deprecated
in matplotlib 3.4 and is gone in current matplotlib (3.6.3 here), where
it raised AttributeError. Ref: https://github.com/raysect/source/issues/383
Signed-off-by: minhaj <minhaj.sixbyte@gmail.com>
two fields (TOW_ms and WN). Probably because this file was written at a time when
these fields were not present in the tracking dump.
This commit fixes that, and plot_tracking_quality_indicator.py is operational now.
This script now also lays out a _RECORD_FORMAT such that it is easily maintainable
if the structure changes again in future in the dll_pll_veml_tracking.cc
Thanks!
Signed-off-by: minhaj <minhaj.sixbyte@gmail.com>
- Added a plot_positive_acq filter option, to enable plotting only the positive
acquisitions.
- Improved the file path processing. Previous version used hardcoded character
offset that will not work if someone sets custom dump prefix. Now we use
python's built-in pathlib and split-parse using the "_" separator.
- The scripts used to use two different ways of encoding system,channel,chips,
one with python dict another with massive if-else nest. This has been uniformed.
- The script used to assume all files in the dump folder would be acquisition dump
but now we filter for acq dumps using pathlib's glob()
Signed-off-by: minhaj <minhaj.sixbyte@gmail.com>
minhaj