pirl-unc / tcrsift / 25855322242
84%

DEFAULT BRANCH: main
Ran
Jobs 4
Files 25
Run time 1min
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: 72.658% (+0.6%) from 72.058%
25855322242

push

github

web-flow 
Rank-family plot bundle: per-sample rank curves, cumulative coverage, (#63)

top-N labeled bars, top-N stacked (#43 items #2/#3/#4/#5)

Adds the four "rank clones within a sample/method" plots from #43 plus
the shared `draw_reference_fractions` log-y utility they all reuse.
Built on a single per-sample-rank kernel (`_per_sample_rank_table`)
that pivots `sample_frequencies` into a sorted-by-frequency view per
sample, with the existing `tier` column propagated for highlight code.

Plots:

1. `plot_rank_curves_per_sample` — log-y rank curves, one panel per
   sample. Tier-selected clones (`tier` column non-null) overlay as
   red dots. Both `log-x` (default) and `linear-x` variants emitted
   by the orchestrator under distinct filenames.

2. `plot_cumulative_coverage` — Σ(top-k frequency) vs k, one line per
   sample on a shared axes. Reference horizontal lines at 50% / 90%
   make clonality differences across samples readable at a glance.

3. `plot_top_n_labeled_bars_per_sample` — top-N clones per sample as
   a labeled bar chart, multi-panel. X labels are CDR3αβ (truncated
   for readability), color encodes tier-selected vs. not, top-N
   leading bars carry a small cell-count annotation when
   `cell_count` is present.

4. `plot_top_n_cumulative_stacked` — per-sample stacked bar of top-N
   contributions. Rank-1 darkest → rank-N lightest under viridis_r
   so the eye reads the leading contributor first. Colorbar shows
   clone rank within sample. Visually conveys mono- vs. oligo-
   clonality across samples.

Shared utility (`draw_reference_fractions`) adds labeled dashed
horizontal lines at configurable fractions (defaults: 10/1/0.1/0.01/
0.001%). Used by plots 1 and 3.

All four plots skip themselves gracefully when `sample_frequencies`
is missing. Tier highlight and cell-count annotations are optional
and degrade without error when the columns aren't present.

`plot_rank_family` orchestrator emits all five PNGs at once and is
wired into the existing... (continued)

5118 of 7044 relevant lines covered (72.66%)

2.91 hits per line

Coverage Regressions

Lines Coverage ∆ File
269
38.28
 15.37% plots.py
Jobs
ID Job ID Ran Files Coverage
1 python-3.11 - 25855322242.1 25
72.65
 GitHub Action Run
2 python-3.10 - 25855322242.2 25
72.65
 GitHub Action Run
3 python-3.12 - 25855322242.3 25
72.65
 GitHub Action Run
4 python-3.9 - 25855322242.4 25
72.63
 GitHub Action Run
Source Files on build 25855322242