desihub / desispec / 934 / 4
23%
main: 38%

LAST BUILD BRANCH: coadd_normalization
DEFAULT BRANCH: main
Ran
Files 76
Run time 5s
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 42.246% (+1.4%) from 40.893%
PYTHON_VERSION=2.7 SETUP_CMD='test --coverage' CONDA_DEPENDENCIES=$CONDA_ALL_DEPENDENCIES PIP_DEPENDENCIES=$PIP_ALL_DEPENDENCIES

push

travis-ci

sbailey 
Bootcalib (#239)

* change threshold definition in find_peaks because not working for sparse fiber (testslit) images

* rm debugging info

* had to adapt threshold for peak finder so that it runs both with full and sparse fiber slit on all cameras b,r,z

* Modify the routines fix_coeff_outliers. One cannot interpolate legendre coeff with fiber id in case the fibers are not periodically spaced on the CCD. The new algorithm uses the X coordinate of the traces instead of the fiber index to interpolation the Y coordinates. The interpolation of the X coefficients is removed.

* fix_ycoeff_outliers : use a tolerance in number of pixels instead of ill defined chi2, iterative loop on fiber rejection

* Important refactoring of bootcalib, with two passes for line identification to take advantage of the number of fibers seeing the same input spectrum. Check consistency of hardcoded line list and NIST ASCII files. Fix bug with duplicate matches.

* do not redo the triplet matching with the full line list if not necessary

* sigma rejection in final polynomial fit

* improve criterion for refitting a fiber by selecting good and bad matched lines based on number of occurence

* add good lines

* do not redo the fiber fit with triplets if 40 good lines or more were found at the first pass (otherwise the number of combinations is too large

* now use an external file  data/arc_lines/goodlines_vacuum.ascii for the list of good lines instead of having them hardcoded

* rm fix_poor_solutions

* remove use_previous_wave

* mv load_gdarc_lines routine to line list section of code

* fix numeric issue when comparing wavelength to get index

* add option to use a custom set of good lines

* adapt tests to new algorithms

* add information in header from arc and continuum exposures

* add possibility to down-select the number of measured lines for the triplet matching algo

* adapt test to new function calls

* restrict the number of good lines to run faster

* add option to write the list of matched lines

3558 of 8422 relevant lines covered (42.25%)

0.42 hits per line

Source Files on job 934.4 (PYTHON_VERSION=2.7 SETUP_CMD='test --coverage' CONDA_DEPENDENCIES=$CONDA_ALL_DEPENDENCIES PIP_DEPENDENCIES=$PIP_ALL_DEPENDENCIES)