/*
    This file is part of darktable,
    Copyright (C) 2011-2026 darktable developers.

    darktable is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation, either version 3 of the License, or
    (at your option) any later version.

    darktable is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.

    You should have received a copy of the GNU General Public License
    along with darktable.  If not, see <http://www.gnu.org/licenses/>.
*/

#include <gtk/gtk.h>
#include <inttypes.h>
#include <stdlib.h>
#include <string.h>

#include "bauhaus/bauhaus.h"
#include "common/box_filters.h"
#include "common/math.h"
#include "common/opencl.h"
#include "control/control.h"
#include "develop/develop.h"
#include "develop/imageop.h"
#include "develop/imageop_gui.h"
#include "develop/tiling.h"
#include "gui/accelerators.h"
#include "gui/gtk.h"
#include "iop/iop_api.h"


#define MAX_RADIUS 16

DT_MODULE_INTROSPECTION(1, dt_iop_highpass_params_t)

typedef struct dt_iop_highpass_params_t
{
  float sharpness; // $MIN: 0.0 $MAX: 100.0 $DEFAULT: 50.0
  float contrast;  // $MIN: 0.0 $MAX: 100.0 $DEFAULT: 50.0 $DESCRIPTION: "contrast boost"
} dt_iop_highpass_params_t;

typedef struct dt_iop_highpass_gui_data_t
{
  GtkWidget *sharpness, *contrast;
} dt_iop_highpass_gui_data_t;

typedef struct dt_iop_highpass_data_t
{
  float sharpness;
  float contrast;
} dt_iop_highpass_data_t;

typedef struct dt_iop_highpass_global_data_t
{
  int kernel_highpass_invert;
  int kernel_highpass_hblur;
  int kernel_highpass_vblur;
  int kernel_highpass_mix;
} dt_iop_highpass_global_data_t;


const char *name()
{
  return _("highpass");
}

const char **description(dt_iop_module_t *self)
{
  return dt_iop_set_description(self, _("isolate high frequencies in the image"),
                                      _("creative"),
                                      _("linear or non-linear, Lab, scene-referred"),
                                      _("frequential, Lab"),
                                      _("special, Lab, scene-referred"));
}

int flags()
{
  return IOP_FLAGS_INCLUDE_IN_STYLES | IOP_FLAGS_SUPPORTS_BLENDING | IOP_FLAGS_ALLOW_TILING;
}

int default_group()
{
  return IOP_GROUP_EFFECT | IOP_GROUP_EFFECTS;
}

dt_iop_colorspace_type_t default_colorspace(dt_iop_module_t *self,
                                            dt_dev_pixelpipe_t *pipe,
                                            dt_dev_pixelpipe_iop_t *piece)
{
  return IOP_CS_LAB;
}

void tiling_callback(dt_iop_module_t *self, dt_dev_pixelpipe_iop_t *piece,
                     const dt_iop_roi_t *roi_in, const dt_iop_roi_t *roi_out,
                     dt_develop_tiling_t *tiling)
{
  dt_iop_highpass_data_t *d = piece->data;

  const int rad = MAX_RADIUS * (fmin(100.0f, d->sharpness + 1) / 100.0f);
  const int radius = MIN(MAX_RADIUS, ceilf(rad * roi_in->scale / piece->iscale));

  const float sigma = sqrtf((radius * (radius + 1) * BOX_ITERATIONS + 2) / 3.0f);
  const int wdh = ceilf(3.0f * sigma);

  tiling->factor = 2.1f; // in + out + small slice for box_mean
  tiling->factor_cl = 3.0f; // in + out + tmp
  tiling->maxbuf = 1.0f;
  tiling->overhead = 0;
  tiling->overlap = wdh;
  tiling->align = 1;
}


#ifdef HAVE_OPENCL
int process_cl(dt_iop_module_t *self, dt_dev_pixelpipe_iop_t *piece, cl_mem dev_in, cl_mem dev_out,
               const dt_iop_roi_t *const roi_in, const dt_iop_roi_t *const roi_out)
{
  dt_iop_highpass_data_t *d = piece->data;
  dt_iop_highpass_global_data_t *gd = self->global_data;

  cl_int err = DT_OPENCL_DEFAULT_ERROR;
  cl_mem dev_tmp = NULL;
  cl_mem dev_m = NULL;

  const int devid = piece->pipe->devid;
  const int width = roi_in->width;
  const int height = roi_in->height;


  const int rad = MAX_RADIUS * (fmin(100.0f, d->sharpness + 1) / 100.0f);
  const int radius = MIN(MAX_RADIUS, ceilf(rad * roi_in->scale / piece->iscale));

  /* sigma-radius correlation to match opencl vs. non-opencl. identified by numerical experiments but
   * unproven. ask me if you need details. ulrich */
  const float sigma = sqrtf((radius * (radius + 1) * BOX_ITERATIONS + 2) / 3.0f);
  const int wdh = ceilf(3.0f * sigma);
  const int wd = 2 * wdh + 1;
  const size_t mat_size = (size_t)wd * sizeof(float);
  float *mat = malloc(mat_size);
  float *m = mat + wdh;
  float weight = 0.0f;

  // init gaussian kernel
  for(int l = -wdh; l <= wdh; l++) weight += m[l] = expf(-(l * l) / (2.f * sigma * sigma));
  for(int l = -wdh; l <= wdh; l++) m[l] /= weight;

  // for(int l=-wdh; l<=wdh; l++) printf("%.6f ", (double)m[l]);
  // printf("\n");

  float contrast_scale = ((d->contrast / 100.0f) * 7.5f);

  int hblocksize;
  dt_opencl_local_buffer_t hlocopt
    = (dt_opencl_local_buffer_t){ .xoffset = 2 * wdh, .xfactor = 1, .yoffset = 0, .yfactor = 1,
                                  .cellsize = sizeof(float), .overhead = 0,
                                  .sizex = 1 << 16, .sizey = 1 };

  if(dt_opencl_local_buffer_opt(devid, gd->kernel_highpass_hblur, &hlocopt) == CL_SUCCESS)
    hblocksize = hlocopt.sizex;
  else
    hblocksize = 1;

  int vblocksize;
  dt_opencl_local_buffer_t vlocopt
    = (dt_opencl_local_buffer_t){ .xoffset = 1, .xfactor = 1, .yoffset = 2 * wdh, .yfactor = 1,
                                  .cellsize = sizeof(float), .overhead = 0,
                                  .sizex = 1, .sizey = 1 << 16 };

  if(dt_opencl_local_buffer_opt(devid, gd->kernel_highpass_vblur, &vlocopt) == CL_SUCCESS)
    vblocksize = vlocopt.sizey;
  else
    vblocksize = 1;


  const size_t bwidth = ROUNDUP(width, hblocksize);
  const size_t bheight = ROUNDUP(height, vblocksize);

  size_t sizes[2];
  size_t local[2];

  dev_tmp = dt_opencl_alloc_device(devid, width, height, sizeof(float) * 4);
  if(dev_tmp == NULL) goto error;

  dev_m = dt_opencl_copy_host_to_device_constant(devid, mat_size, mat);
  if(dev_m == NULL) goto error;

  /* invert image */
  err = dt_opencl_enqueue_kernel_2d_args(devid, gd->kernel_highpass_invert, width, height,
    CLARG(dev_in), CLARG(dev_tmp),
    CLARG(width), CLARG(height));
  if(err != CL_SUCCESS) goto error;

  if(rad != 0)
  {
    /* horizontal blur */
    sizes[0] = bwidth;
    sizes[1] = ROUNDUPDHT(height, devid);
    local[0] = hblocksize;
    local[1] = 1;
    err = dt_opencl_enqueue_kernel_2d_local_args(devid, gd->kernel_highpass_hblur, sizes, local,
      CLARG(dev_tmp), CLARG(dev_out),
      CLARG(dev_m), CLARG(wdh), CLARG(width), CLARG(height), CLARG(hblocksize), CLLOCAL((hblocksize + 2 * wdh) * sizeof(float)));
    if(err != CL_SUCCESS) goto error;

    /* vertical blur */
    sizes[0] = ROUNDUPDWD(width, devid);
    sizes[1] = bheight;
    local[0] = 1;
    local[1] = vblocksize;
    err = dt_opencl_enqueue_kernel_2d_local_args(devid, gd->kernel_highpass_vblur, sizes, local,
      CLARG(dev_out), CLARG(dev_tmp),
      CLARG(dev_m), CLARG(wdh), CLARG(width), CLARG(height), CLARG(vblocksize), CLLOCAL((vblocksize + 2 * wdh) * sizeof(float)));
    if(err != CL_SUCCESS) goto error;
  }

  /* mixing tmp and in -> out */
  err = dt_opencl_enqueue_kernel_2d_args(devid, gd->kernel_highpass_mix, width, height,
    CLARG(dev_in), CLARG(dev_tmp), CLARG(dev_out),
    CLARG(width), CLARG(height), CLARG(contrast_scale));

error:
  dt_opencl_release_mem_object(dev_m);
  dt_opencl_release_mem_object(dev_tmp);
  free(mat);
  return err;
}
#endif

static void _blend(const float *const restrict in,
                   float *const restrict out,
                   const double contrast_scale,
                   const size_t npixels)
{
  /* Blend the inverted blurred L channel with the original input.  Because we packed the L values */
  /* and are inserting the result in the same buffer containing the L values, we need to work in */
  /* reverse order */
  /* We can only do the final 3/4 in parallel here, because updating the first quarter in one thread */
  /* would clobber values still needed by other threads. */
  DT_OMP_FOR()
  for(size_t k = npixels - 1; k >= npixels/4; k--)
  {
    dt_aligned_pixel_t hipass = { 0.0f, 0.0f, 0.0f, 0.0f };  // a=b=0 to desaturate, alpha doesn't matter
   // Mix out and in
    const float L = (out[k] + in[4*k]) - 100.0f;
    hipass[0] = CLAMP((L * contrast_scale) + 50.0f, 0.0f, 100.0f);
    copy_pixel(out + 4*k, hipass);
  }
}

void process(dt_iop_module_t *self,
             dt_dev_pixelpipe_iop_t *piece,
             const void *const ivoid,
             void *const ovoid,
             const dt_iop_roi_t *const roi_in,
             const dt_iop_roi_t *const roi_out)
{
  if(!dt_iop_have_required_input_format(4 /*we need full-color pixels*/, self, piece->colors,
                                        ivoid, ovoid, roi_in, roi_out))
    return;
  dt_iop_highpass_data_t *data = piece->data;
  const float *const in = (float *)ivoid;
  float *out = (float *)ovoid;

/* create inverted image and then blur */
/* since we use only the L channel, pack the values together instead of every fourth float */
/* to reduce cache pressure and memory bandwidth during the blur operation */
  const size_t npixels = (size_t)roi_out->height * roi_out->width;

  DT_OMP_FOR()
  for(size_t k = 0; k < (size_t)npixels; k++)
    out[k] = 100.0f - LCLIP(in[4 * k]); // only L in Lab space

  const int rad = MAX_RADIUS * (fmin(100.0, data->sharpness + 1) / 100.0);
  const int radius = MIN(MAX_RADIUS, ceilf(rad * roi_in->scale / piece->iscale));

  /* horizontal blur out into out */
  const int range = 2 * radius + 1;
  const int hr = range / 2;

  dt_box_mean(out, roi_out->height, roi_out->width, 1, hr, BOX_ITERATIONS);

  // combine the contrast factor from user settings with the averaging
  // factor to save a multiplication per pixel
  const float contrast_scale = ((data->contrast / 100.0f) * 7.5f) * 0.5f;

  _blend(in, out, contrast_scale, npixels);    // only does final 3/4 of given pixels, so repeat
  _blend(in, out, contrast_scale, npixels/4);  // only does final 3/4 of given pixels
  /* process the remaining sixteenth of the pixels */
  for(ssize_t k = npixels/16 - 1; k >= 0; k--)
  {
    dt_aligned_pixel_t hipass = { 0.0f, 0.0f, 0.0f, 0.0f };  // a=b=0 to desaturate, alpha doesn't matter
    // Mix out and in
    const float L = (out[k] + in[4*k]) - 100.0f;
    hipass[0] = CLAMP((L * contrast_scale) + 50.0f, 0.0f, 100.0f);
    copy_pixel(out + 4*k, hipass);
  }
}

void commit_params(dt_iop_module_t *self, dt_iop_params_t *p1, dt_dev_pixelpipe_t *pipe,
                   dt_dev_pixelpipe_iop_t *piece)
{
  dt_iop_highpass_params_t *p = (dt_iop_highpass_params_t *)p1;
  dt_iop_highpass_data_t *d = piece->data;

  d->sharpness = p->sharpness;
  d->contrast = p->contrast;
}

void init_pipe(dt_iop_module_t *self, dt_dev_pixelpipe_t *pipe, dt_dev_pixelpipe_iop_t *piece)
{
  piece->data = calloc(1, sizeof(dt_iop_highpass_data_t));
}

void cleanup_pipe(dt_iop_module_t *self, dt_dev_pixelpipe_t *pipe, dt_dev_pixelpipe_iop_t *piece)
{
  free(piece->data);
  piece->data = NULL;
}

void init_global(dt_iop_module_so_t *self)
{
  const int program = 4; // highpass.cl, from programs.conf
  dt_iop_highpass_global_data_t *gd = malloc(sizeof(dt_iop_highpass_global_data_t));
  self->data = gd;
  gd->kernel_highpass_invert = dt_opencl_create_kernel(program, "highpass_invert");
  gd->kernel_highpass_hblur = dt_opencl_create_kernel(program, "highpass_hblur");
  gd->kernel_highpass_vblur = dt_opencl_create_kernel(program, "highpass_vblur");
  gd->kernel_highpass_mix = dt_opencl_create_kernel(program, "highpass_mix");
}

void cleanup_global(dt_iop_module_so_t *self)
{
  dt_iop_highpass_global_data_t *gd = self->data;
  dt_opencl_free_kernel(gd->kernel_highpass_invert);
  dt_opencl_free_kernel(gd->kernel_highpass_hblur);
  dt_opencl_free_kernel(gd->kernel_highpass_vblur);
  dt_opencl_free_kernel(gd->kernel_highpass_mix);
  free(self->data);
  self->data = NULL;
}


void gui_init(dt_iop_module_t *self)
{
  dt_iop_highpass_gui_data_t *g = IOP_GUI_ALLOC(highpass);

  g->sharpness = dt_bauhaus_slider_from_params(self, N_("sharpness"));
  dt_bauhaus_slider_set_format(g->sharpness, "%");
  gtk_widget_set_tooltip_text(g->sharpness, _("the sharpness of highpass filter"));

  g->contrast = dt_bauhaus_slider_from_params(self, "contrast");
  dt_bauhaus_slider_set_format(g->contrast, "%");
  gtk_widget_set_tooltip_text(g->contrast, _("the contrast of highpass filter"));
}

// clang-format off
// modelines: These editor modelines have been set for all relevant files by tools/update_modelines.py
// vim: shiftwidth=2 expandtab tabstop=2 cindent
// kate: tab-indents: off; indent-width 2; replace-tabs on; indent-mode cstyle; remove-trailing-spaces modified;
// clang-format on