Bug: `camera` raises `IndexError` with 3-column design and non-unit contrast

[robjelier-1]

Bug: camera raises IndexError with 3-column design and non-unit contrast

Version: latest (pip)

Description:
Calling camera with a 3-column design matrix and a contrast vector of length 3 raises an IndexError inside _camera_default.

Minimal reproducible example:

import numpy as np
import edgepython as ep

counts = np.random.poisson(lam=10, size=(1000, 9))
group = np.array(['U'] * 3 + ['D'] * 3 + ['L'] * 3)
design = np.column_stack([
    np.ones(9),
    (group == 'D').astype(float),
    (group == 'L').astype(float)
])
y = ep.make_dgelist(counts=counts, group=group)
y = ep.calc_norm_factors(y)
y = ep.estimate_disp(y, design=design)

index = {'set1': list(range(50))}
res = ep.camera(y, index=index, design=design, contrast=np.array([0, 0.5, 0.5]))

Traceback:

File .../edgepython/gene_sets.py:295, in _camera_default
    if R_reduced[p - 1, p - 1] < 0:
IndexError: index 2 is out of bounds for axis 0 with size 1

Root cause (suspected):
In _camera_default, the design matrix appears to be reduced to a single column before the QR decomposition, so R_reduced is 1×1. With a 3-column design, p-1 = 2 is then out of bounds.

Expected behavior:
camera should handle a multi-column design with a non-unit contrast vector, consistent with the behavior of camera.DGEList in edgeR.