tutorials/empiar_10064/workflow.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "7678e2ec",
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "sys.path.insert(1,'../../../SUSAN') # Path to SUSAN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "0b927064",
   "metadata": {},
   "outputs": [],
   "source": [
    "# SUSAN requires NUMPY and NUMBA\n",
    "import susan\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# For visualization\n",
    "import napari\n",
    "\n",
    "# To create masks:\n",
    "from scipy.ndimage import gaussian_filter as gauss\n",
    "from scipy.ndimage import binary_dilation as dilat"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "24d5e04a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# %% Create Tomostxt:\n",
    "N = 4\n",
    "K = 61\n",
    "\n",
    "tomos = susan.data.Tomograms(n_tomo=N,n_proj=K)\n",
    "apix  = 2.62\n",
    "tsiz  = np.array((3710,3710,880))\n",
    "for i in range(4):\n",
    "    base = 'data/mixedCTEM_tomo%d' % (i+1)\n",
    "    tomos.tomo_id[i] = i+1\n",
    "    tomos.set_stack  (i,base+'.b1.ali.mrc')\n",
    "    tomos.set_angles (i,base+'.tlt')\n",
    "    tomos.set_defocus(i,base+'.defocus')\n",
    "    tomos.pix_size[i]  = apix\n",
    "    tomos.tomo_size[i] = tsiz\n",
    "tomos.save('tomos_b1.tomostxt')\n",
    "\n",
    "tomos = susan.data.Tomograms(n_tomo=N,n_proj=K)\n",
    "apix  = 2.62*2\n",
    "tsiz  = np.uint32(np.ceil(np.array((3710,3710,880))/2))\n",
    "for i in range(4):\n",
    "    base = 'data/mixedCTEM_tomo%d' % (i+1)\n",
    "    tomos.tomo_id[i] = i+1\n",
    "    tomos.set_stack  (i,base+'.b2.ali.mrc')\n",
    "    tomos.set_angles (i,base+'.tlt')\n",
    "    tomos.set_defocus(i,base+'.defocus')\n",
    "    tomos.pix_size[i]  = apix\n",
    "    tomos.tomo_size[i] = tsiz\n",
    "tomos.save('tomos_b2.tomostxt')\n",
    "\n",
    "tomos = susan.data.Tomograms(n_tomo=N,n_proj=K)\n",
    "apix  = 2.62*4\n",
    "tsiz  = np.uint32(np.ceil(np.array((3710,3710,880))/4))\n",
    "for i in range(4):\n",
    "    base = 'data/mixedCTEM_tomo%d' % (i+1)\n",
    "    tomos.tomo_id[i] = i+1\n",
    "    tomos.set_stack  (i,base+'.b4.ali.mrc')\n",
    "    tomos.set_angles (i,base+'.tlt')\n",
    "    tomos.set_defocus(i,base+'.defocus')\n",
    "    tomos.pix_size[i]  = apix\n",
    "    tomos.tomo_size[i] = tsiz\n",
    "tomos.save('tomos_b4.tomostxt')\n",
    "\n",
    "tomos = susan.data.Tomograms(n_tomo=N,n_proj=K)\n",
    "apix  = 2.62*8\n",
    "tsiz  = np.uint32(np.ceil(np.array((3710,3710,880))/8))\n",
    "for i in range(4):\n",
    "    base = 'data/mixedCTEM_tomo%d' % (i+1)\n",
    "    tomos.tomo_id[i] = i+1\n",
    "    tomos.set_stack  (i,base+'.b8.ali.mrc')\n",
    "    tomos.set_angles (i,base+'.tlt')\n",
    "    tomos.set_defocus(i,base+'.defocus')\n",
    "    tomos.pix_size[i]  = apix\n",
    "    tomos.tomo_size[i] = tsiz\n",
    "tomos.save('tomos_b8.tomostxt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "393fa461",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load DYNAMO table with NUMPY and convert it to PTCLSRAW\n",
    "tbl_B8 = np.loadtxt('picked_b8.tbl',unpack=True)\n",
    "tomoB8 = susan.read('tomos_b8.tomostxt')\n",
    "ptcls  = susan.data.Particles.import_data(tomograms=tomoB8,\n",
    "                                          position=tbl_B8[23:26,:].transpose(),\n",
    "                                          ptcls_id=tbl_B8[0],\n",
    "                                          tomos_id=tbl_B8[19])\n",
    "ptcls.save('prj_001.ptclsraw')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "c2d47136",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Alternative use dynamotable (https://github.com/teamtomo/dynamotable)\n",
    "# install with: pip install dynamotable\n",
    "import dynamotable\n",
    "table = dynamotable.read('picked_b8.tbl')\n",
    "x = table['x'].to_numpy(np.float32)[:,np.newaxis]\n",
    "y = table['y'].to_numpy(np.float32)[:,np.newaxis]\n",
    "z = table['z'].to_numpy(np.float32)[:,np.newaxis]\n",
    "pos = np.hstack((x,y,z))\n",
    "p_id = table['tag'].to_numpy(np.uint32)\n",
    "t_id = table['tomo'].to_numpy(np.uint32)\n",
    "tomoB8 = susan.read('tomos_b8.tomostxt')\n",
    "ptcls  = susan.data.Particles.import_data(tomograms=tomoB8,\n",
    "                                          position=pos,\n",
    "                                          ptcls_id=p_id,\n",
    "                                          tomos_id=t_id)\n",
    "ptcls.save('prj_001.ptclsraw')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "75db7124",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    Volume reconstruction. Box size: 32\n",
      "    - Input files: prj_001.ptclsraw | tomos_b8.tomostxt\n",
      "    - Output prefix: prj_001\n",
      "    - 1 GPU (GPU id: 0), and 1 thread.\n",
      "    - Bandpass: [0.0 - 15.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 4427 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_001_class001.mrc ...  Done.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f5408bdf430>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "avgr = susan.modules.Averager()\n",
    "avgr.list_gpus_ids  = [0]\n",
    "avgr.ctf_correction = 'wiener'\n",
    "avgr.rec_halfsets   = False\n",
    "avgr.reconstruct('prj_001','tomos_b8.tomostxt','prj_001.ptclsraw',32)\n",
    "# Initial reconstruction should looks like a sphere (no angles)\n",
    "v,_ = susan.io.mrc.read('prj_001_class001.mrc')\n",
    "plt.imshow(v[16],'gray')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "0bfa141f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Duplicate reference to run a MRA project\n",
    "ptcls = susan.data.Particles('prj_001.ptclsraw')\n",
    "susan.data.Particles.MRA.duplicate(ptcls,0)\n",
    "ptcls.save('prj_001_mra.ptclsraw')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "970baf03",
   "metadata": {},
   "outputs": [],
   "source": [
    "susan.io.mrc.write(-susan.utils.create_sphere( 4,64),   'gb_b4.mrc'   ,2.62*4)\n",
    "susan.io.mrc.write( susan.utils.create_sphere(24,64),'mask_sph_b4.mrc',2.62*4)\n",
    "refs = susan.data.Reference(n_refs=2)\n",
    "refs.ref[0] = 'emd_3420_b4.mrc'\n",
    "refs.msk[0] = 'mask_sph_b4.mrc'\n",
    "refs.ref[1] = 'gb_b4.mrc'\n",
    "refs.msk[1] = 'mask_sph_b4.mrc'\n",
    "refs.save('prj_001_b4.refstxt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "ca93cdbb",
   "metadata": {},
   "outputs": [],
   "source": [
    "mngr = susan.project.Manager('prj_001_b4',64)\n",
    "\n",
    "mngr.initial_reference = 'prj_001_b4.refstxt'\n",
    "mngr.initial_particles = 'prj_001_mra.ptclsraw'\n",
    "mngr.tomogram_file     = 'tomos_b4.tomostxt'\n",
    "mngr.list_gpus_ids     = [0,1,2,3] # List of GPUs ID to use\n",
    "mngr.aligner.ctf_correction  = 'cfsc'\n",
    "mngr.aligner.allow_drift     = True\n",
    "mngr.averager.ctf_correction = 'phase_flip'\n",
    "mngr.cc_threshold   = 0.9\n",
    "mngr.iteration_type = 3 # 3D Alignment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "bf93b5c5",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "============================\n",
      "Project: prj_001_b4 (Iteration 1)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 64\n",
      "    - Input files: tomos_b4.tomostxt | prj_001_b4.refstxt\n",
      "    - Particles In/Out: prj_001_mra.ptclsraw | prj_001_b4/ite_0001/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 21.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=360.000,30.000 | Inplace=360.000,30.000 | Refinement: 2|1.\n",
      "    - Angles | 552 | 21 | 21 [594].\n",
      "    - Ellipsoid offset (3D): [20.00,20.00,20.00], Step=1.00. Points: 33401\n",
      "    - 4427 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 332.5 seconds (0:05:32.509556).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    4231 particles [   3807].\n",
      "      Half 1:    2123 particles [   1910].\n",
      "      Half 2:    2108 particles [   1897].\n",
      "    Class  2:     196 particles [    176].\n",
      "      Half 1:      91 particles [     82].\n",
      "      Half 2:     105 particles [     94].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 64\n",
      "    - Input files: prj_001_b4/ite_0001/temp.ptclsraw | tomos_b4.tomostxt\n",
      "    - Output prefix: prj_001_b4/ite_0001/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 31.0]  (roll off: 2.00).\n",
      "    - Phase-flip. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3983 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_001_b4/ite_0001/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0001/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0001/map_class001.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0001/map_class002_half1.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0001/map_class002_half2.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0001/map_class002.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 6.3 seconds (0:00:06.257542).\n",
      "  [FSC Calculation] Start (2 references):\n",
      "    - Reference  1:  83.840 angstroms [8 fourier pixels]\n",
      "    - Reference  2: 111.787 angstroms [6 fourier pixels]\n",
      "Iteration 1 Finished [Elapsed time: 343.2 seconds (0:05:43.176257]\n"
     ]
    }
   ],
   "source": [
    "mngr.aligner.bandpass.highpass = 0\n",
    "mngr.aligner.bandpass.lowpass  = 21\n",
    "\n",
    "mngr.aligner.set_angular_search(360,30,360,30)\n",
    "mngr.aligner.refine.levels = 2\n",
    "mngr.aligner.refine.factor = 1\n",
    "mngr.aligner.set_offset_search(20,1)\n",
    "\n",
    "bp = mngr.execute_iteration(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "ae5ecb10",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "============================\n",
      "Project: prj_001_b4 (Iteration 2)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 64\n",
      "    - Input files: tomos_b4.tomostxt | prj_001_b4/ite_0001/reference.refstxt\n",
      "    - Particles In/Out: prj_001_b4/ite_0001/particles.ptclsraw | prj_001_b4/ite_0002/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 21.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=0.000,1.000 | Inplace=0.000,1.000 | Refinement: 0|1.\n",
      "    - Angles | 1 [1].\n",
      "    - Ellipsoid offset (3D): [12.00,12.00,12.00], Step=1.00. Points: 7153\n",
      "    - 4427 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 6.0 seconds (0:00:06.023605).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2799 particles [   2518].\n",
      "      Half 1:    1394 particles [   1254].\n",
      "      Half 2:    1405 particles [   1264].\n",
      "    Class  2:    1628 particles [   1465].\n",
      "      Half 1:     820 particles [    738].\n",
      "      Half 2:     808 particles [    727].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 64\n",
      "    - Input files: prj_001_b4/ite_0002/temp.ptclsraw | tomos_b4.tomostxt\n",
      "    - Output prefix: prj_001_b4/ite_0002/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 31.0]  (roll off: 2.00).\n",
      "    - Phase-flip. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3983 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_001_b4/ite_0002/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0002/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0002/map_class001.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0002/map_class002_half1.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0002/map_class002_half2.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0002/map_class002.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 6.1 seconds (0:00:06.083259).\n",
      "  [FSC Calculation] Start (2 references):\n",
      "    - Reference  1:  29.162 angstroms [23 fourier pixels]\n",
      "    - Reference  2:  29.162 angstroms [23 fourier pixels]\n",
      "Iteration 2 Finished [Elapsed time: 16.5 seconds (0:00:16.467565]\n"
     ]
    }
   ],
   "source": [
    "mngr.aligner.bandpass.highpass = 0\n",
    "mngr.aligner.bandpass.lowpass  = 21\n",
    "\n",
    "mngr.aligner.set_angular_search(0,1,0,1)\n",
    "mngr.aligner.refine.levels = 0\n",
    "mngr.aligner.refine.factor = 1\n",
    "mngr.aligner.set_offset_search(12,1)\n",
    "\n",
    "bp = mngr.execute_iteration(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "54ad6010",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "============================\n",
      "Project: prj_001_b4 (Iteration 3)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 64\n",
      "    - Input files: tomos_b4.tomostxt | prj_001_b4/ite_0002/reference.refstxt\n",
      "    - Particles In/Out: prj_001_b4/ite_0002/particles.ptclsraw | prj_001_b4/ite_0003/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 21.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=32.000,6.000 | Inplace=32.000,6.000 | Refinement: 2|1.\n",
      "    - Angles | 95 | 21 | 21 [137].\n",
      "    - Ellipsoid offset (3D): [18.00,18.00,18.00], Step=1.00. Points: 24405\n",
      "    - 4427 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 70.7 seconds (0:01:10.743956).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    3299 particles [   2969].\n",
      "      Half 1:    1640 particles [   1476].\n",
      "      Half 2:    1659 particles [   1493].\n",
      "    Class  2:    1128 particles [   1014].\n",
      "      Half 1:     574 particles [    516].\n",
      "      Half 2:     554 particles [    498].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 64\n",
      "    - Input files: prj_001_b4/ite_0003/temp.ptclsraw | tomos_b4.tomostxt\n",
      "    - Output prefix: prj_001_b4/ite_0003/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 31.0]  (roll off: 2.00).\n",
      "    - Phase-flip. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3983 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_001_b4/ite_0003/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0003/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0003/map_class001.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0003/map_class002_half1.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0003/map_class002_half2.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0003/map_class002.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 6.1 seconds (0:00:06.124519).\n",
      "  [FSC Calculation] Start (2 references):\n",
      "    - Reference  1:  26.829 angstroms [25 fourier pixels]\n",
      "    - Reference  2:  47.909 angstroms [14 fourier pixels]\n",
      "Iteration 3 Finished [Elapsed time: 81.2 seconds (0:01:21.151334]\n",
      "============================\n",
      "Project: prj_001_b4 (Iteration 4)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 64\n",
      "    - Input files: tomos_b4.tomostxt | prj_001_b4/ite_0003/reference.refstxt\n",
      "    - Particles In/Out: prj_001_b4/ite_0003/particles.ptclsraw | prj_001_b4/ite_0004/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 21.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=32.000,6.000 | Inplace=32.000,6.000 | Refinement: 2|1.\n",
      "    - Angles | 95 | 21 | 21 [137].\n",
      "    - Ellipsoid offset (3D): [18.00,18.00,18.00], Step=1.00. Points: 24405\n",
      "    - 4427 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 71.0 seconds (0:01:11.022575).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    3555 particles [   3199].\n",
      "      Half 1:    1776 particles [   1598].\n",
      "      Half 2:    1779 particles [   1601].\n",
      "    Class  2:     872 particles [    784].\n",
      "      Half 1:     438 particles [    394].\n",
      "      Half 2:     434 particles [    390].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 64\n",
      "    - Input files: prj_001_b4/ite_0004/temp.ptclsraw | tomos_b4.tomostxt\n",
      "    - Output prefix: prj_001_b4/ite_0004/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 31.0]  (roll off: 2.00).\n",
      "    - Phase-flip. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3983 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_001_b4/ite_0004/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0004/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0004/map_class001.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0004/map_class002_half1.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0004/map_class002_half2.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0004/map_class002.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 6.0 seconds (0:00:06.034767).\n",
      "  [FSC Calculation] Start (2 references):\n",
      "    - Reference  1:  24.841 angstroms [27 fourier pixels]\n",
      "    - Reference  2:  47.909 angstroms [14 fourier pixels]\n",
      "Iteration 4 Finished [Elapsed time: 81.5 seconds (0:01:21.470003]\n",
      "============================\n",
      "Project: prj_001_b4 (Iteration 5)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 64\n",
      "    - Input files: tomos_b4.tomostxt | prj_001_b4/ite_0004/reference.refstxt\n",
      "    - Particles In/Out: prj_001_b4/ite_0004/particles.ptclsraw | prj_001_b4/ite_0005/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 21.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=32.000,6.000 | Inplace=32.000,6.000 | Refinement: 2|1.\n",
      "    - Angles | 95 | 21 | 21 [137].\n",
      "    - Ellipsoid offset (3D): [18.00,18.00,18.00], Step=1.00. Points: 24405\n",
      "    - 4427 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 69.7 seconds (0:01:09.736578).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    3632 particles [   3268].\n",
      "      Half 1:    1834 particles [   1650].\n",
      "      Half 2:    1798 particles [   1618].\n",
      "    Class  2:     795 particles [    715].\n",
      "      Half 1:     380 particles [    342].\n",
      "      Half 2:     415 particles [    373].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 64\n",
      "    - Input files: prj_001_b4/ite_0005/temp.ptclsraw | tomos_b4.tomostxt\n",
      "    - Output prefix: prj_001_b4/ite_0005/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 31.0]  (roll off: 2.00).\n",
      "    - Phase-flip. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3983 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_001_b4/ite_0005/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0005/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0005/map_class001.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0005/map_class002_half1.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0005/map_class002_half2.mrc ...  Done.\n",
      "        Reconstructing prj_001_b4/ite_0005/map_class002.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 6.1 seconds (0:00:06.110330).\n",
      "  [FSC Calculation] Start (2 references):\n",
      "    - Reference  1:  20.960 angstroms [32 fourier pixels]\n",
      "    - Reference  2:  37.262 angstroms [18 fourier pixels]\n",
      "Iteration 5 Finished [Elapsed time: 80.2 seconds (0:01:20.183268]\n"
     ]
    }
   ],
   "source": [
    "mngr.aligner.bandpass.highpass = 0\n",
    "mngr.aligner.bandpass.lowpass  = 21\n",
    "\n",
    "mngr.aligner.set_angular_search(32,6,32,6)\n",
    "mngr.aligner.refine.levels = 2\n",
    "mngr.aligner.refine.factor = 1\n",
    "mngr.aligner.set_offset_search(18,1)\n",
    "\n",
    "for i in range(3,6):\n",
    "    _ = mngr.execute_iteration(i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "313a8a32",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Show class 1 (Ribosome)\n",
    "_ = napari.view_image(-mngr.get_map(5,1))\n",
    "# Show class 2 (Goldbead)\n",
    "_ = napari.view_image(-mngr.get_map(5,2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "e33b5147",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure()\n",
    "_ = plt.hist( mngr.get_cc(5,1), 100 )\n",
    "_ = plt.hist( mngr.get_cc(5,2), 100 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "90cf3a35",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Continue processing only with first class\n",
    "p0 = mngr.get_ptcls(5)\n",
    "p1 = susan.data.Particles.MRA.select_ref(p0,0)\n",
    "p2 = p1.select(p1.ali_cc[0]>0.02)\n",
    "p2.save('prj_002.ptclsraw')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "8a0abf21",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    Volume reconstruction. Box size: 128\n",
      "    - Input files: prj_002.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_002\n",
      "    - 1 GPU (GPU id: 0), and 1 thread.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3461 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_002_class001.mrc ...  Done.\n"
     ]
    }
   ],
   "source": [
    "avgr = susan.modules.Averager()\n",
    "avgr.list_gpus_ids  = [0]\n",
    "avgr.ctf_correction = 'wiener'\n",
    "avgr.rec_halfsets   = False\n",
    "avgr.reconstruct('prj_002','tomos_b2.tomostxt','prj_002.ptclsraw',128)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "b56c3f3f",
   "metadata": {},
   "outputs": [],
   "source": [
    "a = susan.utils.bandpass(susan.read('prj_002_class001.mrc'),lowpass=20) # lowpass in fourier pixels\n",
    "t = np.quantile(a.flatten(),0.1)\n",
    "b = (a<t)*(susan.utils.create_sphere(38,128)>0.5)\n",
    "b = dilat(b,structure=susan.utils.create_sphere(3,8))\n",
    "c = gauss(np.float32(b),1.5)\n",
    "susan.io.mrc.write(c,'mask_b2_01.mrc')\n",
    "\n",
    "refs = susan.data.Reference(n_refs=1)\n",
    "refs.ref[0] = 'prj_002_class001.mrc'\n",
    "refs.msk[0] = 'mask_b2_01.mrc'\n",
    "refs.save('prj_002.refstxt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "365773a7",
   "metadata": {},
   "outputs": [],
   "source": [
    "mngr = susan.project.Manager('prj_002_b2',128)\n",
    "\n",
    "mngr.initial_reference = 'prj_002.refstxt'\n",
    "mngr.initial_particles = 'prj_002.ptclsraw'\n",
    "mngr.tomogram_file     = 'tomos_b2.tomostxt'\n",
    "mngr.list_gpus_ids     = [0,1,2,3]\n",
    "mngr.aligner.ctf_correction  = 'cfsc'\n",
    "mngr.aligner.allow_drift     = True\n",
    "mngr.averager.ctf_correction = 'wiener'\n",
    "mngr.cc_threshold   = 0.9\n",
    "mngr.iteration_type = 3 # 3D Alignment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "1807e9bd",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "============================\n",
      "Project: prj_002_b2 (Iteration 1)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_002.refstxt\n",
      "    - Particles In/Out: prj_002.ptclsraw | prj_002_b2/ite_0001/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 30.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=0.000,1.000 | Inplace=0.000,1.000 | Refinement: 0|2.\n",
      "    - Angles | 1 [1].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 3461 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 6.5 seconds (0:00:06.536336).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    3461 particles [   3114].\n",
      "      Half 1:    1744 particles [   1569].\n",
      "      Half 2:    1717 particles [   1545].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_002_b2/ite_0001/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_002_b2/ite_0001/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3114 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_002_b2/ite_0001/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0001/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0001/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 6.9 seconds (0:00:06.945720).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  15.244 angstroms [44 fourier pixels]\n",
      "Iteration 1 Finished [Elapsed time: 15.5 seconds (0:00:15.532738]\n",
      "============================\n",
      "Project: prj_002_b2 (Iteration 2)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_002_b2/ite_0001/reference.refstxt\n",
      "    - Particles In/Out: prj_002_b2/ite_0001/particles.ptclsraw | prj_002_b2/ite_0002/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 44.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=5.208,1.302 | Inplace=5.208,1.302 | Refinement: 1|2.\n",
      "    - Angles | 35 | 80 [115].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 3461 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 42.9 seconds (0:00:42.904701).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    3461 particles [   3114].\n",
      "      Half 1:    1744 particles [   1569].\n",
      "      Half 2:    1717 particles [   1545].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_002_b2/ite_0002/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_002_b2/ite_0002/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3114 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_002_b2/ite_0002/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0002/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0002/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 6.8 seconds (0:00:06.783704).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  13.414 angstroms [50 fourier pixels]\n",
      "Iteration 2 Finished [Elapsed time: 51.9 seconds (0:00:51.859580]\n",
      "============================\n",
      "Project: prj_002_b2 (Iteration 3)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_002_b2/ite_0002/reference.refstxt\n",
      "    - Particles In/Out: prj_002_b2/ite_0002/particles.ptclsraw | prj_002_b2/ite_0003/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 46.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.981,1.245 | Inplace=4.981,1.245 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 3461 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 71.9 seconds (0:01:11.891181).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    3461 particles [   3114].\n",
      "      Half 1:    1744 particles [   1569].\n",
      "      Half 2:    1717 particles [   1545].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_002_b2/ite_0003/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_002_b2/ite_0003/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3114 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_002_b2/ite_0003/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0003/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0003/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 7.8 seconds (0:00:07.843899).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  13.151 angstroms [51 fourier pixels]\n",
      "Iteration 3 Finished [Elapsed time: 81.8 seconds (0:01:21.791778]\n",
      "============================\n",
      "Project: prj_002_b2 (Iteration 4)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_002_b2/ite_0003/reference.refstxt\n",
      "    - Particles In/Out: prj_002_b2/ite_0003/particles.ptclsraw | prj_002_b2/ite_0004/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 48.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.774,1.193 | Inplace=4.774,1.193 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 3461 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 72.7 seconds (0:01:12.707313).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    3461 particles [   3114].\n",
      "      Half 1:    1744 particles [   1569].\n",
      "      Half 2:    1717 particles [   1545].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_002_b2/ite_0004/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_002_b2/ite_0004/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3114 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_002_b2/ite_0004/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0004/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0004/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 6.9 seconds (0:00:06.858308).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  12.898 angstroms [52 fourier pixels]\n",
      "Iteration 4 Finished [Elapsed time: 81.5 seconds (0:01:21.528208]\n",
      "============================\n",
      "Project: prj_002_b2 (Iteration 5)\n",
      "  [3D Alignment] Start:\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_002_b2/ite_0004/reference.refstxt\n",
      "    - Particles In/Out: prj_002_b2/ite_0004/particles.ptclsraw | prj_002_b2/ite_0005/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 50.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.583,1.146 | Inplace=4.583,1.146 | Refinement: 1|2.\n",
      "    - Angles | 35 | 80 [115].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 3461 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 43.9 seconds (0:00:43.861179).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    3461 particles [   3114].\n",
      "      Half 1:    1744 particles [   1569].\n",
      "      Half 2:    1717 particles [   1545].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_002_b2/ite_0005/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_002_b2/ite_0005/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3114 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_002_b2/ite_0005/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0005/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0005/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 7.8 seconds (0:00:07.781177).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  12.655 angstroms [53 fourier pixels]\n",
      "Iteration 5 Finished [Elapsed time: 53.8 seconds (0:00:53.782270]\n",
      "============================\n",
      "Project: prj_002_b2 (Iteration 6)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_002_b2/ite_0005/reference.refstxt\n",
      "    - Particles In/Out: prj_002_b2/ite_0005/particles.ptclsraw | prj_002_b2/ite_0006/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 52.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.407,1.102 | Inplace=4.407,1.102 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 3461 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 72.9 seconds (0:01:12.940183).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    3461 particles [   3114].\n",
      "      Half 1:    1744 particles [   1569].\n",
      "      Half 2:    1717 particles [   1545].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_002_b2/ite_0006/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_002_b2/ite_0006/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3114 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_002_b2/ite_0006/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0006/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0006/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 6.6 seconds (0:00:06.636965).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  12.195 angstroms [55 fourier pixels]\n",
      "Iteration 6 Finished [Elapsed time: 81.6 seconds (0:01:21.603410]\n",
      "============================\n",
      "Project: prj_002_b2 (Iteration 7)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_002_b2/ite_0006/reference.refstxt\n",
      "    - Particles In/Out: prj_002_b2/ite_0006/particles.ptclsraw | prj_002_b2/ite_0007/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 54.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.244,1.061 | Inplace=4.244,1.061 | Refinement: 1|2.\n",
      "    - Angles | 35 | 80 [115].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 3461 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 45.0 seconds (0:00:44.952495).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    3461 particles [   3114].\n",
      "      Half 1:    1744 particles [   1569].\n",
      "      Half 2:    1717 particles [   1545].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_002_b2/ite_0007/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_002_b2/ite_0007/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3114 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_002_b2/ite_0007/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0007/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0007/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 7.0 seconds (0:00:06.976573).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.767 angstroms [57 fourier pixels]\n",
      "Iteration 7 Finished [Elapsed time: 53.9 seconds (0:00:53.934067]\n",
      "============================\n",
      "Project: prj_002_b2 (Iteration 8)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_002_b2/ite_0007/reference.refstxt\n",
      "    - Particles In/Out: prj_002_b2/ite_0007/particles.ptclsraw | prj_002_b2/ite_0008/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 56.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.092,1.023 | Inplace=4.092,1.023 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 3461 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 73.7 seconds (0:01:13.713800).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    3461 particles [   3114].\n",
      "      Half 1:    1744 particles [   1569].\n",
      "      Half 2:    1717 particles [   1545].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_002_b2/ite_0008/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_002_b2/ite_0008/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3114 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_002_b2/ite_0008/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0008/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0008/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 6.8 seconds (0:00:06.755189).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.564 angstroms [58 fourier pixels]\n",
      "Iteration 8 Finished [Elapsed time: 82.4 seconds (0:01:22.422619]\n",
      "============================\n",
      "Project: prj_002_b2 (Iteration 9)\n",
      "  [3D Alignment] Start:\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_002_b2/ite_0008/reference.refstxt\n",
      "    - Particles In/Out: prj_002_b2/ite_0008/particles.ptclsraw | prj_002_b2/ite_0009/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 58.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=3.951,0.988 | Inplace=3.951,0.988 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 3461 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 73.4 seconds (0:01:13.392097).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    3461 particles [   3114].\n",
      "      Half 1:    1744 particles [   1569].\n",
      "      Half 2:    1717 particles [   1545].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_002_b2/ite_0009/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_002_b2/ite_0009/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3114 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_002_b2/ite_0009/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0009/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0009/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 7.0 seconds (0:00:07.013995).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.179 angstroms [60 fourier pixels]\n",
      "Iteration 9 Finished [Elapsed time: 82.4 seconds (0:01:22.358285]\n",
      "============================\n",
      "Project: prj_002_b2 (Iteration 10)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_002_b2/ite_0009/reference.refstxt\n",
      "    - Particles In/Out: prj_002_b2/ite_0009/particles.ptclsraw | prj_002_b2/ite_0010/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 60.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=3.819,0.955 | Inplace=3.819,0.955 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 3461 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 73.9 seconds (0:01:13.869059).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    3461 particles [   3114].\n",
      "      Half 1:    1744 particles [   1569].\n",
      "      Half 2:    1717 particles [   1545].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_002_b2/ite_0010/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_002_b2/ite_0010/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3114 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_002_b2/ite_0010/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0010/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_002_b2/ite_0010/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 6.9 seconds (0:00:06.856384).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  10.995 angstroms [61 fourier pixels]\n",
      "Iteration 10 Finished [Elapsed time: 82.8 seconds (0:01:22.820768]\n"
     ]
    }
   ],
   "source": [
    "mngr.aligner.bandpass.highpass = 0\n",
    "\n",
    "mngr.aligner.refine.factor = 2\n",
    "mngr.aligner.set_offset_search(6,1)\n",
    "\n",
    "# close-loop iterations\n",
    "lp = 30\n",
    "for i in range(1,11):\n",
    "    ang_stp = np.rad2deg(np.arctan2(1,lp))\n",
    "    mngr.aligner.bandpass.lowpass = lp\n",
    "    if i == 1:\n",
    "        mngr.aligner.set_angular_search(0,1,0,1)\n",
    "        mngr.aligner.refine.levels = 0\n",
    "    else:\n",
    "        mngr.aligner.set_angular_search(4*ang_stp,ang_stp,4*ang_stp,ang_stp)\n",
    "        mngr.aligner.refine.levels = 1\n",
    "    # mngr.execute_iteration returns the estimated resolution, in fourier pixels,\n",
    "    # using mngr.fsc_threshold as threshold (default: 0.143)\n",
    "    bp = mngr.execute_iteration(i)\n",
    "    if i == 1:\n",
    "        lp = bp\n",
    "    else:\n",
    "        # Enforce a gradual increase in the lowpass\n",
    "        lp = min(lp+2,bp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "73e2e5b7",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    Volume reconstruction. Box size: 128\n",
      "    - Input files: prj_003.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_003\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 3461 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_003_class001.mrc ...  Done.\n",
      "        Reconstructing prj_003_class002.mrc ...  Done.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "p1 = mngr.get_ptcls(10)\n",
    "_ = plt.hist(p1.ali_cc[0],200)\n",
    "susan.data.Particles.MRA.duplicate(p1,0)\n",
    "p1.ref_cix[:] = p1.ali_cc[0]<0.05\n",
    "p1.save('prj_003.ptclsraw')\n",
    "\n",
    "avgr = susan.modules.Averager()\n",
    "avgr.list_gpus_ids  = [0,1,2,3]\n",
    "avgr.ctf_correction = 'wiener'\n",
    "avgr.rec_halfsets   = False\n",
    "avgr.reconstruct('prj_003','tomos_b2.tomostxt','prj_003.ptclsraw',128)\n",
    "\n",
    "a = gauss(np.float32(susan.utils.create_sphere(39,128)),2)\n",
    "susan.io.mrc.write(a,'mask_b2_02.mrc')\n",
    "\n",
    "refs = susan.data.Reference(n_refs=2)\n",
    "refs.ref[0] = 'prj_003_class001.mrc'\n",
    "refs.msk[0] = 'mask_b2_02.mrc'\n",
    "refs.ref[1] = 'prj_003_class002.mrc'\n",
    "refs.msk[1] = 'mask_b2_02.mrc'\n",
    "refs.save('prj_003.refstxt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "bba8f650",
   "metadata": {},
   "outputs": [],
   "source": [
    "mngr = susan.project.Manager('prj_003_b2',128)\n",
    "\n",
    "mngr.initial_reference = 'prj_003.refstxt'\n",
    "mngr.initial_particles = 'prj_003.ptclsraw'\n",
    "mngr.tomogram_file     = 'tomos_b2.tomostxt'\n",
    "mngr.list_gpus_ids     = [0,1,2,3]\n",
    "mngr.aligner.ctf_correction  = 'cfsc'\n",
    "mngr.aligner.allow_drift     = True\n",
    "mngr.averager.ctf_correction = 'wiener'\n",
    "mngr.cc_threshold   = 0.4\n",
    "mngr.iteration_type = 3 # 3D Alignment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "4153b414",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "============================\n",
      "Project: prj_003_b2 (Iteration 1)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_003.refstxt\n",
      "    - Particles In/Out: prj_003.ptclsraw | prj_003_b2/ite_0001/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 30.0]  (roll off: 15.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=8.000,2.000 | Inplace=8.000,2.000 | Refinement: 0|1.\n",
      "    - Angles | 100 [100].\n",
      "    - Ellipsoid offset (3D): [5.00,5.00,5.00], Step=1.00. Points: 515\n",
      "    - 3461 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 69.1 seconds (0:01:09.094652).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2802 particles [   1121].\n",
      "      Half 1:    1402 particles [    561].\n",
      "      Half 2:    1400 particles [    560].\n",
      "    Class  2:     659 particles [    264].\n",
      "      Half 1:     342 particles [    137].\n",
      "      Half 2:     317 particles [    127].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_003_b2/ite_0001/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_003_b2/ite_0001/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1385 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_003_b2/ite_0001/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0001/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0001/map_class001.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0001/map_class002_half1.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0001/map_class002_half2.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0001/map_class002.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 6.9 seconds (0:00:06.934713).\n",
      "  [FSC Calculation] Start (2 references):\n",
      "    - Reference  1:  13.414 angstroms [50 fourier pixels]\n",
      "    - Reference  2:  47.909 angstroms [14 fourier pixels]\n",
      "Iteration 1 Finished [Elapsed time: 79.4 seconds (0:01:19.396267]\n",
      "============================\n",
      "Project: prj_003_b2 (Iteration 2)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_003_b2/ite_0001/reference.refstxt\n",
      "    - Particles In/Out: prj_003_b2/ite_0001/particles.ptclsraw | prj_003_b2/ite_0002/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 30.0]  (roll off: 15.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=8.000,2.000 | Inplace=8.000,2.000 | Refinement: 0|1.\n",
      "    - Angles | 100 [100].\n",
      "    - Ellipsoid offset (3D): [5.00,5.00,5.00], Step=1.00. Points: 515\n",
      "    - 3461 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 70.1 seconds (0:01:10.064416).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2848 particles [   1140].\n",
      "      Half 1:    1442 particles [    577].\n",
      "      Half 2:    1406 particles [    563].\n",
      "    Class  2:     613 particles [    246].\n",
      "      Half 1:     302 particles [    121].\n",
      "      Half 2:     311 particles [    125].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_003_b2/ite_0002/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_003_b2/ite_0002/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1386 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_003_b2/ite_0002/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0002/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0002/map_class001.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0002/map_class002_half1.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0002/map_class002_half2.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0002/map_class002.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 7.2 seconds (0:00:07.157828).\n",
      "  [FSC Calculation] Start (2 references):\n",
      "    - Reference  1:  13.414 angstroms [50 fourier pixels]\n",
      "    - Reference  2:  47.909 angstroms [14 fourier pixels]\n",
      "Iteration 2 Finished [Elapsed time: 80.5 seconds (0:01:20.511999]\n",
      "============================\n",
      "Project: prj_003_b2 (Iteration 3)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_003_b2/ite_0002/reference.refstxt\n",
      "    - Particles In/Out: prj_003_b2/ite_0002/particles.ptclsraw | prj_003_b2/ite_0003/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 30.0]  (roll off: 15.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=8.000,2.000 | Inplace=8.000,2.000 | Refinement: 0|1.\n",
      "    - Angles | 100 [100].\n",
      "    - Ellipsoid offset (3D): [5.00,5.00,5.00], Step=1.00. Points: 515\n",
      "    - 3461 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 70.4 seconds (0:01:10.399419).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2803 particles [   1122].\n",
      "      Half 1:    1417 particles [    567].\n",
      "      Half 2:    1386 particles [    555].\n",
      "    Class  2:     658 particles [    264].\n",
      "      Half 1:     327 particles [    131].\n",
      "      Half 2:     331 particles [    133].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_003_b2/ite_0003/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_003_b2/ite_0003/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1386 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_003_b2/ite_0003/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0003/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0003/map_class001.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0003/map_class002_half1.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0003/map_class002_half2.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0003/map_class002.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 7.2 seconds (0:00:07.176977).\n",
      "  [FSC Calculation] Start (2 references):\n",
      "    - Reference  1:  13.414 angstroms [50 fourier pixels]\n",
      "    - Reference  2:  41.920 angstroms [16 fourier pixels]\n",
      "Iteration 3 Finished [Elapsed time: 80.8 seconds (0:01:20.771102]\n",
      "============================\n",
      "Project: prj_003_b2 (Iteration 4)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_003_b2/ite_0003/reference.refstxt\n",
      "    - Particles In/Out: prj_003_b2/ite_0003/particles.ptclsraw | prj_003_b2/ite_0004/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 30.0]  (roll off: 15.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=8.000,2.000 | Inplace=8.000,2.000 | Refinement: 0|1.\n",
      "    - Angles | 100 [100].\n",
      "    - Ellipsoid offset (3D): [5.00,5.00,5.00], Step=1.00. Points: 515\n",
      "    - 3461 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 70.6 seconds (0:01:10.629810).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2771 particles [   1108].\n",
      "      Half 1:    1398 particles [    559].\n",
      "      Half 2:    1373 particles [    549].\n",
      "    Class  2:     690 particles [    277].\n",
      "      Half 1:     346 particles [    139].\n",
      "      Half 2:     344 particles [    138].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_003_b2/ite_0004/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_003_b2/ite_0004/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1385 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_003_b2/ite_0004/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0004/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0004/map_class001.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0004/map_class002_half1.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0004/map_class002_half2.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0004/map_class002.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 6.7 seconds (0:00:06.705936).\n",
      "  [FSC Calculation] Start (2 references):\n",
      "    - Reference  1:  13.414 angstroms [50 fourier pixels]\n",
      "    - Reference  2:  33.536 angstroms [20 fourier pixels]\n",
      "Iteration 4 Finished [Elapsed time: 80.5 seconds (0:01:20.494464]\n",
      "============================\n",
      "Project: prj_003_b2 (Iteration 5)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 128\n",
      "    - Input files: tomos_b2.tomostxt | prj_003_b2/ite_0004/reference.refstxt\n",
      "    - Particles In/Out: prj_003_b2/ite_0004/particles.ptclsraw | prj_003_b2/ite_0005/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 30.0]  (roll off: 15.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=8.000,2.000 | Inplace=8.000,2.000 | Refinement: 0|1.\n",
      "    - Angles | 100 [100].\n",
      "    - Ellipsoid offset (3D): [5.00,5.00,5.00], Step=1.00. Points: 515\n",
      "    - 3461 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 70.1 seconds (0:01:10.129087).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   1126].\n",
      "      Half 1:    1418 particles [    567].\n",
      "      Half 2:    1397 particles [    559].\n",
      "    Class  2:     646 particles [    259].\n",
      "      Half 1:     326 particles [    131].\n",
      "      Half 2:     320 particles [    128].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 128\n",
      "    - Input files: prj_003_b2/ite_0005/temp.ptclsraw | tomos_b2.tomostxt\n",
      "    - Output prefix: prj_003_b2/ite_0005/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 63.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1385 Particles (2 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_003_b2/ite_0005/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0005/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0005/map_class001.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0005/map_class002_half1.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0005/map_class002_half2.mrc ...  Done.\n",
      "        Reconstructing prj_003_b2/ite_0005/map_class002.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 7.3 seconds (0:00:07.265045).\n",
      "  [FSC Calculation] Start (2 references):\n",
      "    - Reference  1:  13.414 angstroms [50 fourier pixels]\n",
      "    - Reference  2:  31.939 angstroms [21 fourier pixels]\n",
      "Iteration 5 Finished [Elapsed time: 80.7 seconds (0:01:20.676656]\n"
     ]
    }
   ],
   "source": [
    "mngr.aligner.bandpass.highpass = 0\n",
    "mngr.aligner.bandpass.rolloff  = 15\n",
    "mngr.aligner.bandpass.lowpass  = 30\n",
    "mngr.aligner.set_angular_search(8,2,8,2)\n",
    "mngr.aligner.refine.levels = 0\n",
    "mngr.aligner.refine.factor = 1\n",
    "mngr.aligner.set_offset_search(5,1)\n",
    "\n",
    "for i in range(1,6):\n",
    "    _ = mngr.execute_iteration(i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "50339175",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    Volume reconstruction. Box size: 256\n",
      "    - Input files: prj_004.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_004\n",
      "    - 1 GPU (GPU id: 0), and 1 thread.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_004_class001.mrc ...  Done.\n"
     ]
    }
   ],
   "source": [
    "ptcls = mngr.get_ptcls(5)\n",
    "p1 = susan.data.Particles.MRA.select_ref(ptcls,0)\n",
    "p1.save('prj_004.ptclsraw')\n",
    "\n",
    "avgr = susan.modules.Averager()\n",
    "avgr.list_gpus_ids  = [0]\n",
    "avgr.ctf_correction = 'wiener'\n",
    "avgr.rec_halfsets   = False\n",
    "avgr.reconstruct('prj_004','tomos_b1.tomostxt','prj_004.ptclsraw',256)\n",
    "\n",
    "a = susan.utils.bandpass(susan.read('prj_004_class001.mrc'),lowpass=40)\n",
    "t = np.quantile(a.flatten(),0.1)\n",
    "b = (a<t)*(susan.utils.create_sphere(70,256)>0.5)\n",
    "b = dilat(b,structure=susan.utils.create_sphere(6,14))\n",
    "c = gauss(np.float32(b),2)\n",
    "susan.io.mrc.write(c,'mask_b1_01.mrc')\n",
    "\n",
    "refs = susan.data.Reference(n_refs=1)\n",
    "refs.ref[0] = 'prj_004_class001.mrc'\n",
    "refs.msk[0] = 'mask_b1_01.mrc'\n",
    "refs.save('prj_004.refstxt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "706232e1",
   "metadata": {},
   "outputs": [],
   "source": [
    "mngr = susan.project.Manager('prj_004_b1',256)\n",
    "\n",
    "mngr.initial_reference = 'prj_004.refstxt'\n",
    "mngr.initial_particles = 'prj_004.ptclsraw'\n",
    "mngr.tomogram_file     = 'tomos_b1.tomostxt'\n",
    "mngr.list_gpus_ids     = [0,1,2,3]\n",
    "mngr.aligner.ctf_correction  = 'cfsc'\n",
    "mngr.aligner.allow_drift     = True\n",
    "mngr.averager.ctf_correction = 'wiener'\n",
    "mngr.cc_threshold   = 0.9\n",
    "mngr.iteration_type = 3 # 3D Alignment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "328d99af",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "============================\n",
      "Project: prj_004_b1 (Iteration 1)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_004.refstxt\n",
      "    - Particles In/Out: prj_004.ptclsraw | prj_004_b1/ite_0001/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 40.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=0.000,1.000 | Inplace=0.000,1.000 | Refinement: 0|2.\n",
      "    - Angles | 1 [1].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 17.0 seconds (0:00:17.019592).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   2533].\n",
      "      Half 1:    1418 particles [   1276].\n",
      "      Half 2:    1397 particles [   1257].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_004_b1/ite_0001/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_004_b1/ite_0001/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2533 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_004_b1/ite_0001/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0001/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0001/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 17.4 seconds (0:00:17.423349).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  12.421 angstroms [54 fourier pixels]\n",
      "Iteration 1 Finished [Elapsed time: 38.5 seconds (0:00:38.455655]\n",
      "============================\n",
      "Project: prj_004_b1 (Iteration 2)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_004_b1/ite_0001/reference.refstxt\n",
      "    - Particles In/Out: prj_004_b1/ite_0001/particles.ptclsraw | prj_004_b1/ite_0002/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 42.5]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=5.392,1.348 | Inplace=5.392,1.348 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 196.6 seconds (0:03:16.577467).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   2533].\n",
      "      Half 1:    1418 particles [   1276].\n",
      "      Half 2:    1397 particles [   1257].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_004_b1/ite_0002/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_004_b1/ite_0002/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2533 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_004_b1/ite_0002/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0002/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0002/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 15.9 seconds (0:00:15.885731).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.368 angstroms [59 fourier pixels]\n",
      "Iteration 2 Finished [Elapsed time: 216.8 seconds (0:03:36.770546]\n",
      "============================\n",
      "Project: prj_004_b1 (Iteration 3)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_004_b1/ite_0002/reference.refstxt\n",
      "    - Particles In/Out: prj_004_b1/ite_0002/particles.ptclsraw | prj_004_b1/ite_0003/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 45.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=5.092,1.273 | Inplace=5.092,1.273 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 199.2 seconds (0:03:19.241951).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   2533].\n",
      "      Half 1:    1418 particles [   1276].\n",
      "      Half 2:    1397 particles [   1257].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_004_b1/ite_0003/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_004_b1/ite_0003/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2533 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_004_b1/ite_0003/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0003/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0003/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 15.9 seconds (0:00:15.943003).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.179 angstroms [60 fourier pixels]\n",
      "Iteration 3 Finished [Elapsed time: 219.9 seconds (0:03:39.928470]\n",
      "============================\n",
      "Project: prj_004_b1 (Iteration 4)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_004_b1/ite_0003/reference.refstxt\n",
      "    - Particles In/Out: prj_004_b1/ite_0003/particles.ptclsraw | prj_004_b1/ite_0004/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 47.5]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.824,1.206 | Inplace=4.824,1.206 | Refinement: 1|2.\n",
      "    - Angles | 35 | 80 [115].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 116.3 seconds (0:01:56.286484).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   2533].\n",
      "      Half 1:    1418 particles [   1276].\n",
      "      Half 2:    1397 particles [   1257].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_004_b1/ite_0004/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_004_b1/ite_0004/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2533 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_004_b1/ite_0004/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0004/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0004/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 13.0 seconds (0:00:13.026965).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.179 angstroms [60 fourier pixels]\n",
      "Iteration 4 Finished [Elapsed time: 133.6 seconds (0:02:13.550218]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "============================\n",
      "Project: prj_004_b1 (Iteration 5)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_004_b1/ite_0004/reference.refstxt\n",
      "    - Particles In/Out: prj_004_b1/ite_0004/particles.ptclsraw | prj_004_b1/ite_0005/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 50.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.583,1.146 | Inplace=4.583,1.146 | Refinement: 1|2.\n",
      "    - Angles | 35 | 80 [115].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 116.3 seconds (0:01:56.291829).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   2533].\n",
      "      Half 1:    1418 particles [   1276].\n",
      "      Half 2:    1397 particles [   1257].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_004_b1/ite_0005/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_004_b1/ite_0005/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2533 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_004_b1/ite_0005/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0005/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0005/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 15.8 seconds (0:00:15.820850).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.368 angstroms [59 fourier pixels]\n",
      "Iteration 5 Finished [Elapsed time: 136.9 seconds (0:02:16.860528]\n",
      "============================\n",
      "Project: prj_004_b1 (Iteration 6)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_004_b1/ite_0005/reference.refstxt\n",
      "    - Particles In/Out: prj_004_b1/ite_0005/particles.ptclsraw | prj_004_b1/ite_0006/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 52.5]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.365,1.091 | Inplace=4.365,1.091 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 200.9 seconds (0:03:20.922924).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   2533].\n",
      "      Half 1:    1418 particles [   1276].\n",
      "      Half 2:    1397 particles [   1257].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_004_b1/ite_0006/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_004_b1/ite_0006/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2533 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_004_b1/ite_0006/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0006/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0006/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 16.2 seconds (0:00:16.160745).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.368 angstroms [59 fourier pixels]\n",
      "Iteration 6 Finished [Elapsed time: 221.4 seconds (0:03:41.380007]\n"
     ]
    }
   ],
   "source": [
    "mngr.aligner.bandpass.highpass = 0\n",
    "\n",
    "mngr.aligner.refine.factor = 2\n",
    "mngr.aligner.set_offset_search(6,1)\n",
    "\n",
    "lp = 40\n",
    "for i in range(1,7):\n",
    "    ang_stp = np.rad2deg(np.arctan2(1,lp))\n",
    "    mngr.aligner.bandpass.lowpass = lp\n",
    "    if i == 1:\n",
    "        mngr.aligner.set_angular_search(0,1,0,1)\n",
    "        mngr.aligner.refine.levels = 0\n",
    "    else:\n",
    "        mngr.aligner.set_angular_search(4*ang_stp,ang_stp,4*ang_stp,ang_stp)\n",
    "        mngr.aligner.refine.levels = 1\n",
    "    bp = mngr.execute_iteration(i)\n",
    "    lp = min(lp+2.5,bp)\n",
    "    # Denoise reference with FOM [Sindelar and Grigorieff, 2012]\n",
    "    m = mngr.get_names_map(i)\n",
    "    v,a = susan.io.mrc.read(m)\n",
    "    v_f = susan.utils.apply_FOM(v,mngr.get_fsc(i))\n",
    "    susan.io.mrc.write(v_f,m,a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "541afb33",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "============================\n",
      "Project: prj_004_b1 (Iteration 7)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_004_b1/ite_0006/reference.refstxt\n",
      "    - Particles In/Out: prj_004_b1/ite_0006/particles.ptclsraw | prj_004_b1/ite_0007/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 48.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.774,1.193 | Inplace=4.774,1.193 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 196.6 seconds (0:03:16.623416).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   2533].\n",
      "      Half 1:    1418 particles [   1276].\n",
      "      Half 2:    1397 particles [   1257].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_004_b1/ite_0007/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_004_b1/ite_0007/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2533 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_004_b1/ite_0007/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0007/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0007/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 15.3 seconds (0:00:15.285539).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.179 angstroms [60 fourier pixels]\n",
      "Iteration 7 Finished [Elapsed time: 216.1 seconds (0:03:36.059023]\n",
      "============================\n",
      "Project: prj_004_b1 (Iteration 8)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_004_b1/ite_0007/reference.refstxt\n",
      "    - Particles In/Out: prj_004_b1/ite_0007/particles.ptclsraw | prj_004_b1/ite_0008/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 50.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.583,1.146 | Inplace=4.583,1.146 | Refinement: 1|2.\n",
      "    - Angles | 35 | 80 [115].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 116.1 seconds (0:01:56.141317).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   2533].\n",
      "      Half 1:    1418 particles [   1276].\n",
      "      Half 2:    1397 particles [   1257].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_004_b1/ite_0008/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_004_b1/ite_0008/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2533 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_004_b1/ite_0008/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0008/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0008/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 15.4 seconds (0:00:15.355942).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.368 angstroms [59 fourier pixels]\n",
      "Iteration 8 Finished [Elapsed time: 136.6 seconds (0:02:16.561435]\n",
      "============================\n",
      "Project: prj_004_b1 (Iteration 9)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_004_b1/ite_0008/reference.refstxt\n",
      "    - Particles In/Out: prj_004_b1/ite_0008/particles.ptclsraw | prj_004_b1/ite_0009/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 52.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.407,1.102 | Inplace=4.407,1.102 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 199.6 seconds (0:03:19.590638).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   2533].\n",
      "      Half 1:    1418 particles [   1276].\n",
      "      Half 2:    1397 particles [   1257].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_004_b1/ite_0009/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_004_b1/ite_0009/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2533 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_004_b1/ite_0009/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0009/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0009/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 15.3 seconds (0:00:15.260888).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.368 angstroms [59 fourier pixels]\n",
      "Iteration 9 Finished [Elapsed time: 219.4 seconds (0:03:39.374022]\n",
      "============================\n",
      "Project: prj_004_b1 (Iteration 10)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_004_b1/ite_0009/reference.refstxt\n",
      "    - Particles In/Out: prj_004_b1/ite_0009/particles.ptclsraw | prj_004_b1/ite_0010/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 54.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.244,1.061 | Inplace=4.244,1.061 | Refinement: 1|2.\n",
      "    - Angles | 35 | 80 [115].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 116.2 seconds (0:01:56.223954).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   2533].\n",
      "      Half 1:    1418 particles [   1276].\n",
      "      Half 2:    1397 particles [   1257].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_004_b1/ite_0010/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_004_b1/ite_0010/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2533 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_004_b1/ite_0010/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0010/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0010/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 15.1 seconds (0:00:15.080402).\n",
      "  [FSC Calculation] Start (1 reference):\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    - Reference  1:  11.368 angstroms [59 fourier pixels]\n",
      "Iteration 10 Finished [Elapsed time: 135.7 seconds (0:02:15.728464]\n",
      "============================\n",
      "Project: prj_004_b1 (Iteration 11)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_004_b1/ite_0010/reference.refstxt\n",
      "    - Particles In/Out: prj_004_b1/ite_0010/particles.ptclsraw | prj_004_b1/ite_0011/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 56.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.092,1.023 | Inplace=4.092,1.023 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 201.0 seconds (0:03:20.982263).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   2533].\n",
      "      Half 1:    1418 particles [   1276].\n",
      "      Half 2:    1397 particles [   1257].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_004_b1/ite_0011/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_004_b1/ite_0011/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2533 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_004_b1/ite_0011/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0011/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0011/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 13.9 seconds (0:00:13.939543).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.179 angstroms [60 fourier pixels]\n",
      "Iteration 11 Finished [Elapsed time: 219.0 seconds (0:03:39.010953]\n",
      "============================\n",
      "Project: prj_004_b1 (Iteration 12)\n",
      "  [3D Alignment] Start:\n",
      "  Volume 3D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_004_b1/ite_0011/reference.refstxt\n",
      "    - Particles In/Out: prj_004_b1/ite_0011/particles.ptclsraw | prj_004_b1/ite_0012/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 58.0]  (roll off: 2.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=3.951,0.988 | Inplace=3.951,0.988 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Ellipsoid offset (3D): [6.00,6.00,6.00], Step=1.00. Points: 925\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [3D Alignment] Finished. Elapsed time: 201.3 seconds (0:03:21.345617).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   2533].\n",
      "      Half 1:    1418 particles [   1276].\n",
      "      Half 2:    1397 particles [   1257].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_004_b1/ite_0012/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_004_b1/ite_0012/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2533 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_004_b1/ite_0012/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0012/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_004_b1/ite_0012/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 14.0 seconds (0:00:13.999849).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  10.995 angstroms [61 fourier pixels]\n",
      "Iteration 12 Finished [Elapsed time: 219.9 seconds (0:03:39.907489]\n"
     ]
    }
   ],
   "source": [
    "mngr.aligner.bandpass.highpass = 0\n",
    "\n",
    "mngr.aligner.refine.levels = 1\n",
    "mngr.aligner.refine.factor = 2\n",
    "mngr.aligner.set_offset_search(6,1)\n",
    "\n",
    "lp = 48\n",
    "for i in range(7,13):\n",
    "    ang_stp = np.rad2deg(np.arctan2(1,lp))\n",
    "    mngr.aligner.bandpass.lowpass = lp\n",
    "    mngr.aligner.set_angular_search(4*ang_stp,ang_stp,4*ang_stp,ang_stp)\n",
    "    bp = mngr.execute_iteration(i)\n",
    "    lp = min(lp+2,bp)\n",
    "    # Denoise reference with FOM [Sindelar and Grigorieff, 2012]\n",
    "    m = mngr.get_names_map(i)\n",
    "    v,a = susan.io.mrc.read(m)\n",
    "    v_f = susan.utils.apply_FOM(v,mngr.get_fsc(i))\n",
    "    # l0-norm: Using M-sparse constraint [Blumensath and Davies, 2008]\n",
    "    # protein occupies 5% of the box\n",
    "    v_f = susan.utils.denoise_l0(v_f,l0_lambda=0.05)\n",
    "    susan.io.mrc.write(v_f,m,a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "fe309609",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f2c72f5ab20>]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot( mngr.get_fsc(1) )\n",
    "plt.plot( mngr.get_fsc(6) )\n",
    "plt.plot( mngr.get_fsc(12) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "c7f9420b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ptcls = mngr.get_ptcls(12)\n",
    "_ = plt.hist(np.minimum(ptcls.ali_cc[0],0.025),100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "0f3b7d86",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_005.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_005\n",
      "    - 1 GPU (GPU id: 0), and 1 thread.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener SSNR: S=1.00 F=0.80. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_005_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_005_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_005_class001.mrc ...  Done.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f2c6e6fdee0>]"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "p1 = mngr.get_ptcls(12)\n",
    "p1.save('prj_005.ptclsraw')\n",
    "\n",
    "avgr = susan.modules.Averager()\n",
    "avgr.list_gpus_ids  = [0]\n",
    "avgr.ctf_correction = 'wiener_ssnr'\n",
    "avgr.ssnr.S = 1\n",
    "avgr.ssnr.F = 0.8\n",
    "avgr.rec_halfsets   = True\n",
    "avgr.reconstruct('prj_005','tomos_b1.tomostxt','prj_005.ptclsraw',256)\n",
    "\n",
    "v,a = susan.io.mrc.read('prj_005_class001.mrc')\n",
    "fsc = susan.utils.fsc_get('prj_005_class001_half1.mrc','prj_005_class001_half2.mrc','mask_b1_01.mrc')\n",
    "v_f = susan.utils.apply_FOM(v,fsc)\n",
    "v_f = susan.utils.denoise_l0(v_f,l0_lambda=0.04) # protein occupies 5% of the box\n",
    "susan.io.mrc.write(v_f,'prj_005_class001_denoised.mrc',a)\n",
    "\n",
    "refs = susan.data.Reference(n_refs=1)\n",
    "refs.ref[0] = 'prj_005_class001_denoised.mrc'\n",
    "refs.msk[0] = 'mask_b1_01.mrc'\n",
    "refs.save('prj_005.refstxt')\n",
    "\n",
    "plt.plot(fsc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b929fd4f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 2D refinements\n",
    "mngr = susan.project.Manager('prj_005_b1',256)\n",
    "\n",
    "mngr.initial_reference = 'prj_005.refstxt'\n",
    "mngr.initial_particles = 'prj_005.ptclsraw'\n",
    "mngr.tomogram_file     = 'tomos_b1.tomostxt'\n",
    "mngr.list_gpus_ids     = [0,1,2,3]\n",
    "mngr.aligner.ctf_correction  = 'cfsc'\n",
    "mngr.aligner.allow_drift     = True\n",
    "mngr.averager.ctf_correction = 'wiener'\n",
    "mngr.cc_threshold   = 0.6\n",
    "mngr.iteration_type = 2 # 2D refinements"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "b1532fce",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "============================\n",
      "Project: prj_005_b1 (Iteration 1)\n",
      "  [2D Alignment] Start:\n",
      "  Volume 2D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_005.refstxt\n",
      "    - Particles In/Out: prj_005.ptclsraw | prj_005_b1/ite_0001/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 40.0]  (roll off: 10.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=5.728,1.432 | Inplace=5.728,1.432 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Circular offset (2D): [6.00,6.00]. Points: 1469\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [2D Alignment] Finished. Elapsed time: 102.0 seconds (0:01:42.008293).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   1689].\n",
      "      Half 1:    1418 particles [    851].\n",
      "      Half 2:    1397 particles [    838].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_005_b1/ite_0001/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_005_b1/ite_0001/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1689 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_005_b1/ite_0001/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0001/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0001/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 16.1 seconds (0:00:16.107418).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.564 angstroms [58 fourier pixels]\n",
      "Iteration 1 Finished [Elapsed time: 122.5 seconds (0:02:02.499187]\n",
      "============================\n",
      "Project: prj_005_b1 (Iteration 2)\n",
      "  [2D Alignment] Start:\n",
      "  Volume 2D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_005_b1/ite_0001/reference.refstxt\n",
      "    - Particles In/Out: prj_005_b1/ite_0001/particles.ptclsraw | prj_005_b1/ite_0002/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 42.0]  (roll off: 10.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=5.456,1.364 | Inplace=5.456,1.364 | Refinement: 1|2.\n",
      "    - Angles | 35 | 80 [115].\n",
      "    - Circular offset (2D): [6.00,6.00]. Points: 1469\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [2D Alignment] Finished. Elapsed time: 38.5 seconds (0:00:38.495750).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   1689].\n",
      "      Half 1:    1418 particles [    851].\n",
      "      Half 2:    1397 particles [    838].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_005_b1/ite_0002/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_005_b1/ite_0002/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1689 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_005_b1/ite_0002/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0002/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0002/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 13.5 seconds (0:00:13.526202).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.767 angstroms [57 fourier pixels]\n",
      "Iteration 2 Finished [Elapsed time: 56.6 seconds (0:00:56.608932]\n",
      "============================\n",
      "Project: prj_005_b1 (Iteration 3)\n",
      "  [2D Alignment] Start:\n",
      "  Volume 2D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_005_b1/ite_0002/reference.refstxt\n",
      "    - Particles In/Out: prj_005_b1/ite_0002/particles.ptclsraw | prj_005_b1/ite_0003/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 44.0]  (roll off: 10.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=5.208,1.302 | Inplace=5.208,1.302 | Refinement: 1|2.\n",
      "    - Angles | 35 | 80 [115].\n",
      "    - Circular offset (2D): [6.00,6.00]. Points: 1469\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [2D Alignment] Finished. Elapsed time: 38.0 seconds (0:00:38.016479).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   1689].\n",
      "      Half 1:    1418 particles [    851].\n",
      "      Half 2:    1397 particles [    838].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_005_b1/ite_0003/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_005_b1/ite_0003/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1689 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_005_b1/ite_0003/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0003/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0003/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 15.7 seconds (0:00:15.668393).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  12.421 angstroms [54 fourier pixels]\n",
      "Iteration 3 Finished [Elapsed time: 58.6 seconds (0:00:58.566859]\n",
      "============================\n",
      "Project: prj_005_b1 (Iteration 4)\n",
      "  [2D Alignment] Start:\n",
      "  Volume 2D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_005_b1/ite_0003/reference.refstxt\n",
      "    - Particles In/Out: prj_005_b1/ite_0003/particles.ptclsraw | prj_005_b1/ite_0004/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 46.0]  (roll off: 10.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.981,1.245 | Inplace=4.981,1.245 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Circular offset (2D): [6.00,6.00]. Points: 1469\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [2D Alignment] Finished. Elapsed time: 98.3 seconds (0:01:38.324607).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   1689].\n",
      "      Half 1:    1418 particles [    851].\n",
      "      Half 2:    1397 particles [    838].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_005_b1/ite_0004/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_005_b1/ite_0004/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1689 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_005_b1/ite_0004/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0004/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0004/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 17.2 seconds (0:00:17.197074).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.767 angstroms [57 fourier pixels]\n",
      "Iteration 4 Finished [Elapsed time: 119.7 seconds (0:01:59.700679]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "============================\n",
      "Project: prj_005_b1 (Iteration 5)\n",
      "  [2D Alignment] Start:\n",
      "  Volume 2D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_005_b1/ite_0004/reference.refstxt\n",
      "    - Particles In/Out: prj_005_b1/ite_0004/particles.ptclsraw | prj_005_b1/ite_0005/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 48.0]  (roll off: 10.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.774,1.193 | Inplace=4.774,1.193 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Circular offset (2D): [6.00,6.00]. Points: 1469\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [2D Alignment] Finished. Elapsed time: 103.0 seconds (0:01:43.036650).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   1689].\n",
      "      Half 1:    1418 particles [    851].\n",
      "      Half 2:    1397 particles [    838].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_005_b1/ite_0005/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_005_b1/ite_0005/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1689 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_005_b1/ite_0005/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0005/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0005/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 16.3 seconds (0:00:16.273194).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.767 angstroms [57 fourier pixels]\n",
      "Iteration 5 Finished [Elapsed time: 123.4 seconds (0:02:03.425735]\n",
      "============================\n",
      "Project: prj_005_b1 (Iteration 6)\n",
      "  [2D Alignment] Start:\n",
      "  Volume 2D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_005_b1/ite_0005/reference.refstxt\n",
      "    - Particles In/Out: prj_005_b1/ite_0005/particles.ptclsraw | prj_005_b1/ite_0006/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 50.0]  (roll off: 10.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.583,1.146 | Inplace=4.583,1.146 | Refinement: 1|2.\n",
      "    - Angles | 35 | 80 [115].\n",
      "    - Circular offset (2D): [6.00,6.00]. Points: 1469\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [2D Alignment] Finished. Elapsed time: 40.7 seconds (0:00:40.653330).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   1689].\n",
      "      Half 1:    1418 particles [    851].\n",
      "      Half 2:    1397 particles [    838].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_005_b1/ite_0006/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_005_b1/ite_0006/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1689 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_005_b1/ite_0006/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0006/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0006/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 14.4 seconds (0:00:14.433773).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  11.368 angstroms [59 fourier pixels]\n",
      "Iteration 6 Finished [Elapsed time: 59.0 seconds (0:00:59.035102]\n",
      "============================\n",
      "Project: prj_005_b1 (Iteration 7)\n",
      "  [2D Alignment] Start:\n",
      "  Volume 2D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_005_b1/ite_0006/reference.refstxt\n",
      "    - Particles In/Out: prj_005_b1/ite_0006/particles.ptclsraw | prj_005_b1/ite_0007/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 52.0]  (roll off: 10.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.407,1.102 | Inplace=4.407,1.102 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Circular offset (2D): [6.00,6.00]. Points: 1469\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [2D Alignment] Finished. Elapsed time: 97.6 seconds (0:01:37.586172).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   1689].\n",
      "      Half 1:    1418 particles [    851].\n",
      "      Half 2:    1397 particles [    838].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_005_b1/ite_0007/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_005_b1/ite_0007/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1689 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_005_b1/ite_0007/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0007/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0007/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 16.7 seconds (0:00:16.745419).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  10.995 angstroms [61 fourier pixels]\n",
      "Iteration 7 Finished [Elapsed time: 118.8 seconds (0:01:58.847845]\n",
      "============================\n",
      "Project: prj_005_b1 (Iteration 8)\n",
      "  [2D Alignment] Start:\n",
      "  Volume 2D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_005_b1/ite_0007/reference.refstxt\n",
      "    - Particles In/Out: prj_005_b1/ite_0007/particles.ptclsraw | prj_005_b1/ite_0008/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 54.0]  (roll off: 10.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.244,1.061 | Inplace=4.244,1.061 | Refinement: 1|2.\n",
      "    - Angles | 35 | 80 [115].\n",
      "    - Circular offset (2D): [6.00,6.00]. Points: 1469\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [2D Alignment] Finished. Elapsed time: 40.1 seconds (0:00:40.083993).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   1689].\n",
      "      Half 1:    1418 particles [    851].\n",
      "      Half 2:    1397 particles [    838].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_005_b1/ite_0008/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_005_b1/ite_0008/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1689 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_005_b1/ite_0008/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0008/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0008/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 15.5 seconds (0:00:15.503925).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  10.818 angstroms [62 fourier pixels]\n",
      "Iteration 8 Finished [Elapsed time: 60.4 seconds (0:01:00.365678]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "============================\n",
      "Project: prj_005_b1 (Iteration 9)\n",
      "  [2D Alignment] Start:\n",
      "  Volume 2D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_005_b1/ite_0008/reference.refstxt\n",
      "    - Particles In/Out: prj_005_b1/ite_0008/particles.ptclsraw | prj_005_b1/ite_0009/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 56.0]  (roll off: 10.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=4.092,1.023 | Inplace=4.092,1.023 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Circular offset (2D): [6.00,6.00]. Points: 1469\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [2D Alignment] Finished. Elapsed time: 100.1 seconds (0:01:40.067726).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   1689].\n",
      "      Half 1:    1418 particles [    851].\n",
      "      Half 2:    1397 particles [    838].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_005_b1/ite_0009/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_005_b1/ite_0009/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1689 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_005_b1/ite_0009/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0009/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0009/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 15.9 seconds (0:00:15.878091).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  10.480 angstroms [64 fourier pixels]\n",
      "Iteration 9 Finished [Elapsed time: 119.8 seconds (0:01:59.788281]\n",
      "============================\n",
      "Project: prj_005_b1 (Iteration 10)\n",
      "  [2D Alignment] Start:\n",
      "  Volume 2D alignment. Box size: 256\n",
      "    - Input files: tomos_b1.tomostxt | prj_005_b1/ite_0009/reference.refstxt\n",
      "    - Particles In/Out: prj_005_b1/ite_0009/particles.ptclsraw | prj_005_b1/ite_0010/particles.ptclsraw\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 58.0]  (roll off: 10.00).\n",
      "    - On Substack (CFSC). Normalized to Mean=0, Std=1.\n",
      "    - Angular search: Sym=c1 | Cone=3.951,0.988 | Inplace=3.951,0.988 | Refinement: 1|2.\n",
      "    - Angles | 100 | 100 [200].\n",
      "    - Circular offset (2D): [6.00,6.00]. Points: 1469\n",
      "    - 2815 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Aligning particles: 100.00%                              \n",
      "  [2D Alignment] Finished. Elapsed time: 102.4 seconds (0:01:42.436836).\n",
      "  [Aligned partices] Processing:\n",
      "    Class  1:    2815 particles [   1689].\n",
      "      Half 1:    1418 particles [    851].\n",
      "      Half 2:    1397 particles [    838].\n",
      "  [Aligned partices] Done.\n",
      "  [Reconstruct Maps] Start:\n",
      "    Volume reconstruction (including half-sets). Box size: 256\n",
      "    - Input files: prj_005_b1/ite_0010/temp.ptclsraw | tomos_b1.tomostxt\n",
      "    - Output prefix: prj_005_b1/ite_0010/map\n",
      "    - 4 GPUs (GPU ids: 0,1,2,3), and 4 threads.\n",
      "    - Bandpass: [0.0 - 127.0]  (roll off: 2.00).\n",
      "    - Wiener Inversion. Normalized to Mean=0, Std=1.\n",
      "    - c1 Symmetry. Inversion: Ite=10 Std=0.750000\n",
      "\tLoading data files... Done\n",
      "    - 1689 Particles (1 classes) in 4 tomograms with max 61 projections.\n",
      "        Filling fourier space: 100.00%                           \n",
      "        Reconstructing prj_005_b1/ite_0010/map_class001_half1.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0010/map_class001_half2.mrc ...  Done.\n",
      "        Reconstructing prj_005_b1/ite_0010/map_class001.mrc ...  Done.\n",
      "  [Reconstruct Maps] Finished. Elapsed time: 15.5 seconds (0:00:15.485533).\n",
      "  [FSC Calculation] Start (1 reference):\n",
      "    - Reference  1:  10.162 angstroms [66 fourier pixels]\n",
      "Iteration 10 Finished [Elapsed time: 121.8 seconds (0:02:01.788803]\n"
     ]
    }
   ],
   "source": [
    "mngr.aligner.bandpass.highpass = 0\n",
    "mngr.aligner.bandpass.rolloff  = 10\n",
    "\n",
    "mngr.aligner.refine.levels = 1\n",
    "mngr.aligner.refine.factor = 2\n",
    "mngr.aligner.set_offset_search(6)\n",
    "\n",
    "lp = 40\n",
    "for i in range(1,11):\n",
    "    ang_stp = np.rad2deg(np.arctan2(1,lp))\n",
    "    mngr.aligner.bandpass.lowpass = lp\n",
    "    mngr.aligner.set_angular_search(4*ang_stp,ang_stp,4*ang_stp,ang_stp)\n",
    "    bp = mngr.execute_iteration(i)\n",
    "    lp = min(lp+2,bp)\n",
    "    m = mngr.get_names_map(i)\n",
    "    v,a = susan.io.mrc.read(m)\n",
    "    v_f = susan.utils.apply_FOM(v,mngr.get_fsc(i))\n",
    "    susan.io.mrc.write(v_f,m[:-4]+'_fom.mrc',a) # save FOM\n",
    "    v_f = susan.utils.denoise_l0(v_f,l0_lambda=0.03)\n",
    "    susan.io.mrc.write(v_f,m,a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "dc738d8a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f2064768850>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot( mngr.get_fsc(1),label='Iteration 1' )\n",
    "plt.plot( mngr.get_fsc(10),label='Iteration 10' )\n",
    "plt.axis([0,128+5,-0.05,1.05])\n",
    "plt.yticks([0,0.143,0.5,0.75,1])\n",
    "plt.xticks(np.arange(0,128,16))\n",
    "plt.grid(True)\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2901bea5",
   "metadata": {},
   "outputs": [],
   "source": [
    "_ = napari.view_image(-mngr.get_map(10))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "7f821e1e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(-0.5, 159.5, 159.5, -0.5)"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f,ax = plt.subplots(1,2)\n",
    "f.suptitle('XY plane')\n",
    "v = susan.read('prj_005_b1/ite_0001/map_class001_fom.mrc')\n",
    "ax[0].imshow( v[135,208:48:-1,48:208], 'gray' )\n",
    "ax[0].set_title('Iteration 1')\n",
    "ax[0].axis('off')\n",
    "v = susan.read('prj_005_b1/ite_0010/map_class001_fom.mrc')\n",
    "ax[1].imshow( v[135,208:48:-1,48:208], 'gray' )\n",
    "ax[1].set_title('Iteration 10')\n",
    "ax[1].axis('off')\n",
    "f,ax = plt.subplots(1,2)\n",
    "f.suptitle('XY plane')\n",
    "v = susan.read('prj_005_b1/ite_0001/map_class001_fom.mrc')\n",
    "ax[0].imshow( v[208:48:-1,127,48:208], 'gray' )\n",
    "ax[0].set_title('Iteration 1')\n",
    "ax[0].axis('off')\n",
    "v = susan.read('prj_005_b1/ite_0010/map_class001_fom.mrc')\n",
    "ax[1].imshow( v[208:48:-1,127,48:208], 'gray' )\n",
    "ax[1].set_title('Iteration 10')\n",
    "ax[1].axis('off')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}