Jungfraujoch/image_analysis/scale_merge/RotationScaleMergeGPU.h

// SPDX-FileCopyrightText: 2026 Filip Leonarski, Paul Scherrer Institute <filip.leonarski@psi.ch>
// SPDX-License-Identifier: GPL-3.0-only

#pragma once

#include <cstdint>
#include <memory>
#include <optional>
#include <vector>

// GPU engine for the RotationScaleMerge hot loops. The class keeps the per-observation data resident on
// the device as a structure-of-arrays (coalesced) and runs the scaling loop there. The host keeps the
// one-time raw-hkl sort and the per-space-group ASU keying (gemmi); it hands the GPU the dense group ids
// plus a group-ordered permutation so the per-group reduction is a deterministic segmented reduction
// (one block per group, fixed order, no atomics) - matching the run-to-run determinism of the CPU path.
//
// Only compiled when CUDA is available; the header is safe to include unconditionally (the impl behind
// the pimpl is null without CUDA, and RotationScaleMerge falls back to the CPU loops).
class RotationScaleMergeGPU {
public:
    RotationScaleMergeGPU();
    ~RotationScaleMergeGPU();
    RotationScaleMergeGPU(const RotationScaleMergeGPU &) = delete;
    RotationScaleMergeGPU &operator=(const RotationScaleMergeGPU &) = delete;

    // True if a GPU was found and the engine is usable.
    [[nodiscard]] bool Available() const;

    // Upload the immutable per-observation fields (once). Arrays are length n_obs unless noted; the
    // frame CSR (frame_start/frame_count) is length n_frames and indexes the obs arrays in frame order.
    void SetPartials(int n_obs, int n_frames,
                     const float *I, const float *sigma, const float *rlp, const float *partiality,
                     const float *zeta, const uint8_t *on_ice, const int32_t *frame,
                     const float *corr0,
                     const int32_t *frame_start, const int32_t *frame_count);

    // Per space group: the dense ASU-group id per obs, and a group-ordered permutation of the obs whose
    // group >= 0 (group_perm), with its CSR (group_start/group_count, length n_groups) - so each group's
    // observations are a contiguous, fixed-order segment for the reduction.
    void SetGroups(int n_groups, const int32_t *group,
                   const int32_t *group_perm, int n_group_perm,
                   const int32_t *group_start, const int32_t *group_count);

    // Re-upload the working corr (length n_obs) before a scaling pass (the host mutates it via smooth-G
    // between passes). SetPartials uploads the initial corr; this refreshes it.
    void SetCorr(const float *corr);

    // Run `iters` of {reduce group means -> per-frame robust IRLS G -> update corr} on the device,
    // in place on the resident corr. Rotation model (partiality folded via the stored partiality).
    void ScalePartials(int iters, double robust_k, double min_partiality, bool has_d_min);

    // Copy the updated corr back to the host (length n_obs), and the fitted per-frame G (length
    // n_frames, as double) plus the per-frame "was fitted" flag.
    void GetCorr(float *corr_out) const;
    void GetG(double *g_out, uint8_t *scaled_out) const;

    // Apply the smooth-G correction to the resident corr: corr[i] *= ratio[frame[i]] for frames with
    // apply[f] (both length n_frames), matching CPU SmoothG. Keeps corr resident (no round-trip).
    void SmoothCorr(const uint8_t *apply, const double *ratio);

    // --- merge + error-model reductions over the resident, scaled fulls (reuse the fulls group CSR) ---

    // The per-frame cell-consistency mask (length n_frames) used by the merge filter. Uploaded once.
    void SetFrameCellOk(const uint8_t *frame_cell_ok);

    // Per-group inv-var mean (em_mean, length n_groups) + per-full leverage-corrected error-model samples
    // (s2/I2/dev2 + valid flag, length n_fulls), mirroring MergeAndStats' first two error-model loops.
    // Stashes (for_search, masked, ice, min_partiality) for the MergeAccum/MergeRmeas calls that follow.
    void MergeEmSamples(bool for_search, const uint8_t *masked, int n_masked,
                        double ice_half_width_q, double min_partiality,
                        double *em_mean_out, int32_t *cnt_out, double *s2_out, double *I2_out,
                        double *dev2_out, uint8_t *valid_out);

    // Per-group merge accumulators (inv-var sums + deterministic half-sets, error-model-corrected sigma
    // from a/b). Outputs length n_groups; rejected[g] counts outliers dropped (reject_median uploaded, NAN
    // where none). Requires MergeEmSamples first (em_mean resident).
    void MergeAccum(double error_model_a, double error_model_b, bool error_model_active,
                    bool reject_outliers, double reject_nsigma, const float *reject_median,
                    double *swI, double *sw, double *swIh0, double *swIh1,
                    double *swh0, double *swh1, int32_t *nh0, int32_t *nh1, double *d_out,
                    int32_t *rejected);

    // Per-group R_meas accumulators (sum|I_corr-merged_I|, sum_I, n, and the usable count for per-shell
    // total_observations); merged_I is uploaded. All arrays length n_groups.
    void MergeRmeas(const double *merged_I, double *absdev, double *sumI, int32_t *n, int32_t *nusable);

    // Post-smooth per-frame diagnostic CC: recompute the group means from the resident (smoothed) corr
    // and the Pearson CC of each frame's I*corr vs its group mean, downloading only the per-frame cc /
    // cc_n (length n_frames). Mirrors ReduceGroupMeans(partials) + FinalizePerFrameScale's CC loop.
    void ComputePartialCC(double min_partiality, double *cc_out, int64_t *cc_n_out);

    // --- 3D combine (partials -> fulls), all on the device ---

    // The per-obs fields the combine needs on top of the scaling inputs (image-local bkg, fractional
    // frame position for event contiguity, resolution). Uploaded once, alongside SetPartials.
    void SetCombineInputs(const float *bkg, const float *image_number, const float *d);

    // The one-time raw-hkl run layout (space-group-independent): the (raw h,k,l, image_number)-sorted
    // permutation of the obs, split into contiguous per-raw-hkl runs. Uploaded once in Ingest.
    void SetRawRuns(int n_runs, int n_perm, const int32_t *perm,
                    const int32_t *rawrun_start, const int32_t *rawrun_count,
                    const int32_t *rawrun_h, const int32_t *rawrun_k, const int32_t *rawrun_l);

    // Combine the resident partials (reading the current resident corr) into fulls on the device,
    // mirroring RotationScaleMerge::Combine: one thread per raw-hkl run splits its usable partials into
    // rocking events (frame gap <= 2), pools background, seeds F, does 3 de-biased Poisson reweights and
    // adds the capture-uncertainty term. rawrun_group (length n_runs) is the current space group's ASU
    // id per raw hkl (it becomes the full's group). Deterministic: fulls are emitted in raw-run-major,
    // event order (a count pass -> host prefix sum -> emit-at-offset), matching the CPU path. Returns the
    // number of fulls (call GetFulls with buffers of that length).
    int Combine(const int32_t *rawrun_group, double min_partiality, double capture_uncertainty_coeff);

    // Download the combined fulls SoA (length = Combine()'s return). The working corr is downloaded
    // separately by GetFullsCorr (it is only meaningful after ScaleFulls; otherwise the caller sets it).
    void GetFulls(int32_t *h, int32_t *k, int32_t *l, float *I, float *sigma, float *d,
                  float *image_number, int32_t *frame, uint8_t *on_ice, int32_t *group) const;

    // --- scale the resident fulls on the device (Unity model), no round-trip ---

    // Download the fulls' frame and ASU-group keys (emit order) so the host can build the frame/group CSRs
    // with a counting sort (deterministic, no GPU stable-sort) and hand them back below.
    void GetFullsKeys(int32_t *frame, int32_t *group) const;

    // The fulls' per-frame CSR: frame_perm groups the emit-ordered fulls by frame (frame_start/count length
    // n_frames index it), so FitPerFrameG can scale the fulls without physically reordering them.
    void SetFullsFrameCSR(const int32_t *frame_perm, int n_perm,
                          const int32_t *frame_start, const int32_t *frame_count);

    // The fulls' per-ASU-group CSR (group-ordered permutation of the fulls with group>=0, + its CSR).
    void SetFullsGroups(const int32_t *gperm, int n_gperm,
                        const int32_t *gstart, const int32_t *gcount);

    // Run `iters` of the Unity scaling loop on the resident fulls (reduce group means -> per-frame IRLS G
    // -> update corr), in place on the fulls' working corr. Requires SetFullsFrameCSR + SetFullsGroups.
    void ScaleFulls(int iters, double robust_k, double min_partiality);

    // Download the fulls' working corr (length = n_fulls), valid after ScaleFulls.
    void GetFullsCorr(float *corr) const;

private:
    struct Impl;
    std::unique_ptr<Impl> impl_;
};