tomography-tutorial.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "3411ff94",
   "metadata": {},
   "source": [
    "This notebook provides some examples on two-qubit tomography. \n",
    "\n",
    "## Initialization\n",
    "\n",
    "### Imports\n",
    "\n",
    "First, we'll do the meeded imports. Note that library's .py files are directly accessed from the folder `solver`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "fd1534bd",
   "metadata": {},
   "outputs": [],
   "source": [
    "from os import path\n",
    "import warnings\n",
    "assert path.exists(\"solver\") \n",
    "assert path.exists(\"solver/utils\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "cf9858bb",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-04-30 12:49:48.628872: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcudart.so.11.0'; dlerror: libcudart.so.11.0: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: :/home/qc7/.mujoco/mjpro150/bin\n",
      "2022-04-30 12:49:48.628960: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.\n",
      "2022-04-30 12:49:51.771942: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: :/home/qc7/.mujoco/mjpro150/bin\n",
      "2022-04-30 12:49:51.772017: W tensorflow/stream_executor/cuda/cuda_driver.cc:269] failed call to cuInit: UNKNOWN ERROR (303)\n",
      "2022-04-30 12:49:51.772074: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:156] kernel driver does not appear to be running on this host (PeKa-II): /proc/driver/nvidia/version does not exist\n",
      "2022-04-30 12:49:51.772481: I tensorflow/core/platform/cpu_feature_guard.cc:151] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
     ]
    }
   ],
   "source": [
    "with warnings.catch_warnings():  # warning still somehow pierces through, so I cleared the cell output\n",
    "    import tensorflow as tf  # tf 2.x\n",
    "    import tensornetwork as tn\n",
    "    tn.set_default_backend(\"tensorflow\")\n",
    "    \n",
    "import numpy as np\n",
    "import QGOpt as qgo\n",
    "import math\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import solver.utils.general_utils as util\n",
    "import solver.noising_tools as ns\n",
    "import solver.utils.channel_utils as c_util\n",
    "import solver.file_management as fm\n",
    "import solver.circuits_generation as cg\n",
    "\n",
    "from solver.QCCalc import QCEvaluator\n",
    "from solver.experiments import ExperimentConductor, assert_psi2_eq_1\n",
    "from solver.utils.misc import unwrap_dict, NconTemplate, INT, FLOAT, COMPLEX, ID_GATE\n",
    "from solver.QCSolver import QGOptSolver, get_complex_channel_form, QGOptSolverDebug"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7b5f2763",
   "metadata": {},
   "source": [
    "### Quantum gates preparation\n",
    "\n",
    "Let us conduct a simple experiment which makes quantum tomography of gates $\\sf RX$, $\\sf RY$ and $\\sf CX$. $\\sf RX$ and $\\sf RY$ represent unitary rotations with angle $\\pi/2$, while $\\sf CX$ represents the controlled NOT gate. These quantum gates can be expressed as unitary operators as follows:\n",
    "$$\n",
    "{\\sf RX} =  \\begin{bmatrix}\n",
    "            \\cos(\\pi/4) & -\\imath \\sin(\\pi/4)\\\\\n",
    "            -\\imath \\sin(\\pi/4) & \\cos(\\pi/4)\n",
    "        \\end{bmatrix},  \n",
    "{\\sf RY} =  \\begin{bmatrix}\n",
    "            \\cos(\\pi/4) & -\\sin(\\pi/4)\\\\\n",
    "            \\sin(\\pi/4) & \\cos(\\pi/4)\n",
    "        \\end{bmatrix},\n",
    "{\\sf CX} =      \n",
    "    \\begin{bmatrix}\n",
    "        1&0&0&0\\\\0&1&0&0\\\\0&0&0&1\\\\0&0&1&0\n",
    "    \\end{bmatrix}.\n",
    "$$\n",
    "\n",
    "We will also need an identity matrix: both for providiing a fictionary measurement gate (SPAM operation) and for internal purposes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "101dcb9c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rx: [[0.707+0.j    0.   -0.707j]\n",
      " [0.   -0.707j 0.707+0.j   ]]\n",
      "Ry: [[ 0.707+0.j -0.707+0.j]\n",
      " [ 0.707+0.j  0.707+0.j]]\n"
     ]
    }
   ],
   "source": [
    "Rx = util.create_unitary_rotation_x(math.pi/2)\n",
    "Ry = util.create_unitary_rotation_y(math.pi/2)\n",
    "print('Rx:', Rx.numpy().round(3))\n",
    "print('Ry:', Ry.numpy().round(3))\n",
    "\n",
    "CX_44 = tf.constant([[1, 0, 0, 0],\n",
    "                     [0, 1, 0, 0],\n",
    "                     [0, 0, 0, 1],\n",
    "                     [0, 0, 1, 0]], dtype=COMPLEX)\n",
    "\n",
    "E = tf.eye(2, dtype=COMPLEX)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e05b3fde",
   "metadata": {},
   "source": [
    "Now it is time to convert these unitary operators to quantum channels. Two-qubit gates need to be reshaped first."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "9d196074",
   "metadata": {},
   "outputs": [],
   "source": [
    "CX = util.convert_44_to_2222(CX_44)\n",
    "\n",
    "Rx_channel = c_util.convert_1qmatrix_to_channel(Rx)\n",
    "Ry_channel = c_util.convert_1qmatrix_to_channel(Ry)\n",
    "E_channel = c_util.convert_1qmatrix_to_channel(E)\n",
    "CX_channel = c_util.convert_2qmatrix_to_channel(CX)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "92df4407",
   "metadata": {},
   "source": [
    "Now we need to make set of channels from this. The recommended order is to fill all single-qubit gates, then internal identity gate (**this is important**), then all two-qubit gates."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "40653164",
   "metadata": {},
   "outputs": [],
   "source": [
    "default_pure_channels = {'Rx':Rx_channel,\n",
    "                         'Ry':Ry_channel,\n",
    "                         'M': E_channel,\n",
    "                         ID_GATE:E_channel,\n",
    "                         'CX': CX_channel}                        "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8a3044cd",
   "metadata": {},
   "source": [
    "### Circuits preparation"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "394ce905",
   "metadata": {},
   "source": [
    "We will need a `DataGenerator`, which can (surprisingly) generate quantum circuits, and convert quantum circuits to the tensor network template form. To initialize it, we need the gates' labels, and it's quite important to pass them in the same order as stated in `default_pure_channels` **without the internal `ID_GATE`**. Some day this inconvenient feature will be fixed. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "100ce769",
   "metadata": {},
   "outputs": [],
   "source": [
    "NUM_QUBITS = 2\n",
    "\n",
    "gates_labels = list(default_pure_channels.keys())\n",
    "gates_labels.remove(ID_GATE)\n",
    "\n",
    "gen = cg.DataGenerator(qubits_num=NUM_QUBITS,\n",
    "                       gates_names=gates_labels, # ['Rx', 'Ry', 'M', 'CX']\n",
    "                       single_qub_gates_num=3,\n",
    "                       two_qub_gates_num=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "24c53092",
   "metadata": {},
   "source": [
    "However, let's resort to using just the second option of what `DataGenerator` can do. We will not use any pre-made generators, but rather create the set of circuits ourselves. For simple tomography experiment, one can use different sets of:\n",
    "* state preparation part\n",
    "* middle part a. k. a. gates to be tomographied\n",
    "* measurement gates part\n",
    "\n",
    "TODO: explain why these sets are usually taken for tomography and how exactly they are conducted.\n",
    "\n",
    "The quantum circuits is now defined as sequence of quantum gates applied to a different qubits. In the code, the so-called \"human\" quantum circuit is represented as a list of strings, with each string having format\n",
    "'LABEL_TARGET' for single-qubit gates or 'LABEL_TARGET1_TARGET2' for two-qubit gates. The order of targets is important, and controlling qubit goes **first**.\n",
    "\n",
    "We can write a simple function which generates a list of all available combinations. First, let's generate all circuits without CX gates in it:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "fa6f47cc",
   "metadata": {},
   "outputs": [],
   "source": [
    "def gen_circs_without_CX():\n",
    "    preps = [[],\n",
    "             ['Rx_0', 'Rx_1'],\n",
    "             ['Ry_0', 'Ry_1'],\n",
    "             ['Rx_0', 'Rx_0', 'Rx_1', 'Rx_1']]\n",
    "    \n",
    "    toms = [[],\n",
    "            ['Rx_0', 'Rx_1'],\n",
    "            ['Ry_0', 'Ry_1']]\n",
    "    \n",
    "    meass = [['M_0', 'M_1'],\n",
    "             ['Rx_0', 'Rx_1', 'M_0', 'M_1'],\n",
    "             ['Ry_0', 'Ry_1', 'M_0', 'M_1']]\n",
    "    \n",
    "    ans = []\n",
    "    \n",
    "    for prep in preps:\n",
    "        for tom in toms:\n",
    "            for meas in meass:\n",
    "                ans.append(prep + tom + meas)\n",
    "                \n",
    "    return ans\n",
    "\n",
    "circs1 = gen_circs_without_CX()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d05a5e9e",
   "metadata": {},
   "source": [
    "We can print these circuits represented by sequences of gates. Note that the labels of gates in string format are constistent with labels we gave in `default_pure_channels` dict as keys. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "8f1e7b8d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dict_keys(['Rx', 'Ry', 'M', '_E', 'CX'])\n",
      "Circuit #8 is represented as ['Ry_0', 'Ry_1', 'Ry_0', 'Ry_1', 'M_0', 'M_1']\n",
      "Circuit #24 is represented as ['Ry_0', 'Ry_1', 'Ry_0', 'Ry_1', 'M_0', 'M_1']\n",
      "Circuit #3 is represented as ['Rx_0', 'Rx_1', 'M_0', 'M_1']\n"
     ]
    }
   ],
   "source": [
    "np.random.seed(100)\n",
    "print(default_pure_channels.keys())\n",
    "\n",
    "for idx in np.random.choice(np.arange(len(circs1)), 3):\n",
    "    print(f'Circuit #{idx} is represented as {circs1[idx]}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "57c665a1",
   "metadata": {},
   "source": [
    "Circuits with CX gate between qubits 0 and 1 are made a little bit differently: the set of preparations and measurements is now extended (it allows asymmetric measurements), but now we have only one choice for middle part."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "df012a40",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "180\n"
     ]
    }
   ],
   "source": [
    "def gen_circs_with_CX():\n",
    "    preps_0 = [[],\n",
    "               ['Rx_0'],\n",
    "               ['Ry_0'],\n",
    "               ['Rx_0', 'Rx_0']]\n",
    "    preps_1 = [[],\n",
    "               ['Rx_1'],\n",
    "               ['Ry_1'],\n",
    "               ['Rx_1', 'Rx_1']]\n",
    "    \n",
    "    meas_0 = [['M_0'],\n",
    "              ['Rx_0', 'M_0'],\n",
    "              ['Ry_0', 'M_0']]\n",
    "    meas_1 = [['M_1'],\n",
    "              ['Rx_1', 'M_1'],\n",
    "              ['Ry_1', 'M_1']]\n",
    "    \n",
    "    ans = []\n",
    "    tup_ans = []\n",
    "    \n",
    "    for prep_0 in preps_0:\n",
    "        for prep_1 in preps_1:\n",
    "            for m_0 in meas_0:\n",
    "                for m_1 in meas_1:\n",
    "                    ans.append(prep_0 + prep_1 + ['CX_0_1'] + m_0 + m_1)\n",
    "                \n",
    "    return ans\n",
    "\n",
    "circs2 = gen_circs_with_CX()\n",
    "\n",
    "circs = circs1 + circs2\n",
    "print(len(circs))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "36557a30",
   "metadata": {},
   "source": [
    "So far so good: we have just created 180 different circuits which allow for quantum gate tomography. Now need to convert it into tensor network templates format. Templates are stored in a dict, so they require some keys to access them. Right now we won't use any names, and just convert the index of a circuit to string - but you can name your circuits arbitratily. \n",
    "\n",
    "Also we add an empty circuit for demonstration purposes:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "34bb1f24",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "A quantum circuit with very simple \"human\" representation \n",
      "['Rx_0', 'Rx_0', 'CX_0_1', 'Ry_0', 'M_0', 'Rx_1', 'M_1']\n",
      "has a very scary tensor network template \n",
      "[[0, 0, 7, 2, 4, 1, 5], [[3, 1], [4, 3], [6, 5, 2, 4], [7, 5], [-1, 7], [9, 6], [-2, 9]], [3, 4, 5, 6, 7, 9], [-1, -2]]\n",
      "A quantum circuit with very simple \"human\" representation \n",
      "['Rx_0', 'Ry_1', 'CX_0_1', 'M_0', 'Ry_1', 'M_1']\n",
      "has a very scary tensor network template \n",
      "[[0, 3, 7, 4, 3, 5], [[3, 1], [4, 2], [6, 5, 4, 3], [-1, 5], [8, 6], [-2, 8]], [3, 4, 5, 6, 8], [-1, -2]]\n",
      "A quantum circuit with very simple \"human\" representation \n",
      "['Rx_0', 'Rx_0', 'Ry_1', 'CX_0_1', 'Rx_0', 'M_0', 'Ry_1', 'M_1']\n",
      "has a very scary tensor network template \n",
      "[[0, 0, 3, 7, 0, 4, 3, 5], [[3, 1], [4, 3], [5, 2], [7, 6, 5, 4], [8, 6], [-1, 8], [10, 7], [-2, 10]], [3, 4, 5, 6, 7, 8, 10], [-1, -2]]\n"
     ]
    }
   ],
   "source": [
    "np.random.seed(1337)\n",
    "\n",
    "ncon_tmpls = gen.get_tmpl_dict_from_human_circs(circs + [[]])  # add an empty circuit\n",
    "\n",
    "for idx in np.random.choice(np.arange(len(circs)), 3):\n",
    "    print(f'A quantum circuit with very simple \"human\" representation \\n{circs[idx]}'\n",
    "    f'\\nhas a very scary tensor network template \\n{ncon_tmpls[str(idx)]}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "03aec358",
   "metadata": {},
   "source": [
    "Right now we will not explain how these tensor network templates work. However, this form is widely used in the package `ncon` from `tensornetwork` library, and if you are familiar with it, you can read the tensor network templates just as \"human\" circuits."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "6e7fd962",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "An empty quantum circuit was converted to [[6, 6], [[-1, 1], [-2, 2]], [], [-1, -2]]\n"
     ]
    }
   ],
   "source": [
    "print(f'An empty quantum circuit was converted to {list(ncon_tmpls.values())[-1]}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9c30d08f",
   "metadata": {},
   "source": [
    "Sometimes, your circuit will have no gates at certain qubits (or it may have no gates at all). DataGenerator notices that and contracts inputs with outputs using the special `ID_GATE` we passed above, which has the index of 6."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f0acfb74",
   "metadata": {},
   "source": [
    "### Noising"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "24ca6df6",
   "metadata": {},
   "source": [
    "The noise config may look quite complicated, but it's easier to grasp when you look closely at its element.\n",
    "Element of a noise config is a tuple contaning\n",
    "* label of a gate\n",
    "* id of a gate; in a single-qubit case it coincides with qubit number and in two-qubit case it can be obtained by enumerating all qubit pairs. E. g. in three-qubit scenario, targer pairs `[(0, 1), (0, 2), (1, 0), (1, 2), (2, 0), (2, 1)]` will be mapped to ids from 0 to 5.\n",
    "* function to noise - it's best to use a pre-defined function, which has four parameters\n",
    "    * parameter for gaussian blur \n",
    "    * parameter for depolarization channel\n",
    "    * parameter for amplitude damping\n",
    "    * parameter for phase damping\n",
    "\n",
    "TODO: tell about noise params more thoroughly."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a277c543",
   "metadata": {},
   "outputs": [],
   "source": [
    "noise_cfg = []\n",
    "\n",
    "# comments are fidelities between noised and ideal versions of gate\n",
    "noise_cfg.append(('Rx', 0, ns.make_1q_4pars_channel, 0.065, 0.03, 0.005, 0.005)) # 0.97135\n",
    "noise_cfg.append(('Rx', 1, ns.make_1q_4pars_channel, 0.05, 0.03, 0.03, 0.0)) # 0.96153\n",
    "\n",
    "noise_cfg.append(('Ry', 0, ns.make_1q_4pars_channel, 0.02, 0.02, 0.02, 0.02)) # 0.9701\n",
    "noise_cfg.append(('Ry', 1, ns.make_1q_4pars_channel, 0.06, 0.01, 0.03, 0.03)) # 0.96826\n",
    "\n",
    "noise_cfg.append(('M', 0, ns.make_1q_4pars_channel, 0.0, 0.04, 0.02, 0.02)) # 0.95567\n",
    "noise_cfg.append(('M', 1, ns.make_1q_4pars_channel, 0.0, 0.03, 0.03, 0.03)) # 0.95583\n",
    "\n",
    "noise_cfg.append(('CX', 0, ns.make_2q_4pars_channel, 0.05, 0.02, 0.02, 0.02)) # 0.93728"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "319f5a54",
   "metadata": {},
   "source": [
    "We can apply this noising function to a gate and see how it changes. For example, we may calculate fidelity between \"pure\" and \"noised\" quantum channels. Notice the small machine calculation error in imaginary part - we know that it should be equal to zero, but nothing is perfect. There are other metrics to compute (e. g. diamond norm). Since diamond norm is calculated via external Qiskit library, its output is quite beautiful without any formatting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "2d56b217",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fidelity is (0.9688945548090188+1.9075878905071783e-08j)\n",
      "Diamond norm is 0.06466075373331996\n"
     ]
    }
   ],
   "source": [
    "pure_channel = Rx_channel\n",
    "noised_channel = ns.make_1q_4pars_channel(pure_channel, [0.05, 0.02, 0.02, 0.02])\n",
    "print('Fidelity is', util.fidel_calc_1q(pure_channel, noised_channel).numpy())\n",
    "print('Diamond norm is', util.diamond_norm_1q(pure_channel, noised_channel).numpy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "5086f51e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fidelity is (0.9372799568088622+4.759170873593774e-18j)\n",
      "Diamond norm is 0.13671446696269413\n"
     ]
    }
   ],
   "source": [
    "pure_channel = CX_channel\n",
    "noised_channel = ns.make_2q_4pars_channel(pure_channel, [0.05, 0.02, 0.02, 0.02])\n",
    "print('Fidelity is', util.fidel_calc_2q(pure_channel, noised_channel).numpy())\n",
    "print('Diamond norm is', util.diamond_norm_2q(pure_channel, noised_channel).numpy())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "08abbbb7",
   "metadata": {},
   "source": [
    "## Conducting experiment"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5d017576",
   "metadata": {},
   "source": [
    "We have a class called ExperimentConductor contaning all the parameters for our experiment. It has the:\n",
    "* ideal gate set ${\\bf G}^{\\rm ideal}$\n",
    "* noise configuration we have just made\n",
    "* experiment name for working with files (more on that later)\n",
    "* number of qubits\n",
    "* learning rate\n",
    "* regularization coefficients for one-qubit and two-qubit gates\n",
    "* sample size a. k. a. shots number for sample generation (8192 is a realistic number you can get at real quantum PC)\n",
    "\n",
    "Since we're not conducting any complicated experiments, the full functionality of this class will not be unleashed. However, it's convenient to store all the information together in order not to lose track. Notice the learning rate, regularizations and iterations hyperparameteres - they are all set here. These are not optimal but get the job done. Feel free to experiment!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "cf635d9b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:5 out of the last 5 calls to <function make_1q_4pars_channel at 0x7f37b8533c10> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
      "WARNING:tensorflow:6 out of the last 6 calls to <function make_1q_4pars_channel at 0x7f37b8533c10> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n"
     ]
    }
   ],
   "source": [
    "with warnings.catch_warnings():\n",
    "    exp_test = ExperimentConductor(pure_channels_set=default_pure_channels,\n",
    "                                   noise_cfg=noise_cfg,\n",
    "                                   exp_name='tomo_tutorial',\n",
    "                                   qubits_num=NUM_QUBITS,\n",
    "                                   lr=0.002,  \n",
    "                                   lmbd1=100,\n",
    "                                   lmbd2=100,\n",
    "                                   iterations=300,\n",
    "                                   sample_size=8192)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c6d2a7bf",
   "metadata": {},
   "source": [
    "First, let's use solver as a sample generator in order to generate the outcomes for each circuit. In the current version of the program, the number of outcomes (`sample_size`) is shared across all circuits.   \n",
    "We are using the debug version of QGOptSolver since it gives more info. Pay attention to the variable `noise_iter0` - it determines the initial approximation. If you suspect your data to be noisy, increase it. It is **SUPER** important for convergence to set this above zero in order not to get stuck in a local minima! "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "ed4da81c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "generating samples... 25% 50% 75% Done!\n"
     ]
    }
   ],
   "source": [
    "QC_t = QGOptSolverDebug(qubits_num=NUM_QUBITS,\n",
    "                        single_qub_gates_names={'Rx', 'Ry', 'M'},\n",
    "                        two_qub_gates_names={'CX'},\n",
    "                        pure_channels_set=default_pure_channels,\n",
    "                        compress_samples=True,\n",
    "                        noise_params=exp_test.noise_params,\n",
    "                        noise_iter0=0.07)\n",
    "\n",
    "for name, tmpl in ncon_tmpls.items():\n",
    "    QC_t.add_circuit(tn_template=tmpl, name=name)\n",
    "    \n",
    "QC_t.generate_all_samples(v=False, smpl_size=exp_test.sample_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b44b27f",
   "metadata": {},
   "source": [
    "Now we need to setup our manifold for quantum channels optimization and grab optimizer itself from `QGOpt` library."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "3dbc3bff",
   "metadata": {},
   "outputs": [],
   "source": [
    "manif = qgo.manifolds.ChoiMatrix()\n",
    "opt_t = qgo.optimizers.RAdam(manif, exp_test.lr)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ffc75cc9",
   "metadata": {},
   "source": [
    "We are ready to minimize the loss hard-coded in the solver:   \n",
    "TODO: talk about loss.  \n",
    "See article for details.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "f3c58eb3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 out of 300 iterations passed\n",
      "30 out of 300 iterations passed\n",
      "60 out of 300 iterations passed\n",
      "90 out of 300 iterations passed\n",
      "120 out of 300 iterations passed\n",
      "150 out of 300 iterations passed\n",
      "180 out of 300 iterations passed\n",
      "210 out of 300 iterations passed\n",
      "240 out of 300 iterations passed\n",
      "270 out of 300 iterations passed\n"
     ]
    }
   ],
   "source": [
    "loss_dynamics_t, norms_dict, _ = QC_t.train_optimizer(opt=opt_t,\n",
    "                                                      lmbd1=exp_test.lmbd1,\n",
    "                                                      lmbd2=exp_test.lmbd2,\n",
    "                                                      iters=exp_test.iters,\n",
    "                                                      v=1,\n",
    "                                                      fid_ctr=1, # we'd like to calculate fidelities every step\n",
    "                                                      norm_ctr=-1)  # we don't calculate L1 norms here"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b9eb96d8",
   "metadata": {},
   "source": [
    "We can look at our loss dymamics to see that learning process was OK:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "9f59d41d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f37b83b4790>]"
      ]
     },
     "execution_count": 19,
     "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(loss_dynamics_t)  "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1a8dec86",
   "metadata": {},
   "source": [
    "Could use even smaller learning rate, but whatever."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4a5051cc",
   "metadata": {},
   "source": [
    "## Results"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4f68c14e",
   "metadata": {},
   "source": [
    "The debug version of the solver has the ability to extract norms $\\left( \\| {\\bf G}^{\\rm est}(l,t) - {\\bf G}^{\\rm true}(l,t) \\|, \\; \\| {\\bf G}^{\\rm est}(l,t) - {\\bf G}^{\\rm ideal}(l,t) \\| \\right) $ during each iteration of algorithm. These norms will provide us with beautiful graphs. For most of the gates, diamond norm in extracted, and for measurement gate ${\\sf M}$ it is POVM norm (see article for details).  \n",
    "\n",
    "Let's plot these norms vs. algorithm iteration number to prove that our algorithm did a good job. Remember, since we were the ones who generated the noise, we have information about norms between true vestions and estimated versions of quantum gates $\\| {\\bf G}^{\\rm est}(l,t) - {\\bf G}^{\\rm true}(l,t) \\| $. Of course, in real experiment we would be able to see only norms between ideal and estimated versions.\n",
    "\n",
    "First, we extract the norms between true and ideal quantum gates; these norms will appear as horizontal lines a. k. a. answers. Our goal is to make ideal-estimated norm $\\| {\\bf G}^{\\rm est}(l,t) - {\\bf G}^{\\rm ideal}(l,t) \\| $ as close as possible to true-ideal norm $\\| {\\bf G}^{\\rm true}(l,t) - {\\bf G}^{\\rm ideal}(l,t) \\| $. This process is related to minimizing the true-estimated norm. Note that unlike in the article, here the number of circuis stays the same! X axis represents the number of iterations which is just a hyperparameter in the big experiment."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "ff28c932",
   "metadata": {},
   "outputs": [],
   "source": [
    "from collections import defaultdict\n",
    "HLINES = defaultdict(list)\n",
    "\n",
    "for label in QC_t.hidden_gates_dict:\n",
    "    sublist = QC_t.hidden_gates_dict[label]\n",
    "    for idx, gate in enumerate(sublist):\n",
    "        if label == 'CX':\n",
    "            HLINES[label].append(util.diamond_norm_2q(gate, default_pure_channels[label]))\n",
    "        elif label == 'M':\n",
    "            HLINES[label].append(util.get_povm_dist(gate, default_pure_channels[label]))\n",
    "        else:\n",
    "            HLINES[label].append(util.diamond_norm_1q(gate, default_pure_channels[label]))\n",
    "            \n",
    "COLORS = ['tab:blue', 'tab:orange', 'tab:green', 'tab:red', 'tab:purple', 'tab:gray']"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eaa2b565",
   "metadata": {},
   "source": [
    "### Plotting norms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "c8b4a46c",
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_diamond_norms(ax1, ax2):\n",
    "    colors_idx = [0, 0]\n",
    "    \n",
    "    for key in norms_dict.keys():\n",
    "        label, idx, norm_type = key\n",
    "\n",
    "        if label in ['Rx', 'Ry', 'CX']:\n",
    "            if label == 'CX' and idx == 1:\n",
    "                continue\n",
    "                \n",
    "            if norm_type == 'i':\n",
    "                ax1.plot(np.abs(norms_dict[label, idx, norm_type]),\n",
    "                         label=f'{label}{idx}',\n",
    "                         color=COLORS[colors_idx[0]])\n",
    "                ax1.axhline(y=HLINES[label][idx],\n",
    "                            color=COLORS[colors_idx[0]],\n",
    "                            linestyle=':')\n",
    "                colors_idx[0] += 1\n",
    "\n",
    "            else:\n",
    "                ax2.plot(np.abs(norms_dict[label, idx, norm_type]),\n",
    "                         label=f'{label}{idx}',\n",
    "                         color=COLORS[colors_idx[1]])\n",
    "                colors_idx[1] += 1\n",
    "            \n",
    "    ax1.set_title('Diamond norm(ideal, est)', {\"size\":18})\n",
    "    ax1.set_xlabel('Algorithm iteration number', {\"size\":16})\n",
    "\n",
    "    ax2.set_title('Diamond norm(true, est)', {\"size\":18})\n",
    "    ax2.set_xlabel('Algorithm iteration number', {\"size\":16})\n",
    "\n",
    "    ax1.legend(prop={'size': 14})\n",
    "    ax2.legend(prop={'size': 14})\n",
    "        \n",
    "        # ax.set_ylim(0.92, 0.98)\n",
    "        \n",
    "def plot_povm_norms(ax1, ax2):\n",
    "    colors_idx = [0, 0]\n",
    "    \n",
    "    for key in norms_dict.keys():\n",
    "        label, idx, norm_type = key\n",
    "\n",
    "        if label == 'M':\n",
    "            if norm_type == 'i':\n",
    "                ax1.plot(np.abs(norms_dict[label, idx, norm_type]),\n",
    "                         label=f'{label}{idx}',\n",
    "                         color=COLORS[colors_idx[0]])\n",
    "                ax1.axhline(y=HLINES[label][idx],\n",
    "                            color=COLORS[colors_idx[0]],\n",
    "                            linestyle=':')\n",
    "                colors_idx[0] += 1\n",
    "\n",
    "            else:\n",
    "                ax2.plot(np.abs(norms_dict[label, idx, norm_type]),\n",
    "                         label=f'{label}{idx}',\n",
    "                         color=COLORS[colors_idx[1]])\n",
    "                colors_idx[1] += 1\n",
    "            \n",
    "    ax1.set_title('POVM norm(ideal, est)', {\"size\":18})\n",
    "    ax1.set_xlabel('Algorithm iteration number', {\"size\":16})\n",
    "\n",
    "    ax2.set_title('POVM norm(true, est)', {\"size\":18})\n",
    "    ax2.set_xlabel('Algorithm iteration number', {\"size\":16})\n",
    "\n",
    "    ax1.legend(prop={'size': 14})\n",
    "    ax2.legend(prop={'size': 14})\n",
    "        \n",
    "        # ax.set_ylim(0.92, 0.98)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "c67c5b2c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x864 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axarr = plt.subplots(2, 2, figsize=(16, 12))\n",
    "ax1, ax2, ax3, ax4 = [a for sublist in axarr for a in sublist]\n",
    "plot_diamond_norms(ax1, ax2)\n",
    "plot_povm_norms(ax3, ax4)\n",
    "fig.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b0c6b71b",
   "metadata": {},
   "source": [
    "As you can see, our alrogithm achieved decent results while using not much information (180 circuits with 8192 samples). Notice how the too big initial noise `noise_iter0=0.07` forces algorithm to think that gates are in fact less noisy then it thought, which is reflected in the graphs: gates norms are decreasing with iterations. Try setting `noise_iter0=0.05` and re-running the notebook to achieve somewhat better final results."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "55cd1840",
   "metadata": {},
   "source": [
    "### Post-procesing, saving and loading data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "307f1dba",
   "metadata": {},
   "source": [
    "You may have a need to conduct experiments on different devices. Or just conveniently save the results somewhere in order to access them later. Fortunately, there's a file `solver/file_management.py` dedicated to such tasks. The code there should be simple and easy to read.\n",
    "\n",
    "The most important information after an algorithm has done its job now resides in the member `estimated_gates_dict` of a `Solver` class. However, we need to convert it to complex channel form for better readability."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "ea31ccca",
   "metadata": {},
   "outputs": [],
   "source": [
    "dict_result = get_complex_channel_form(QC_t.estimated_gates_dict)\n",
    "# print(dict_result)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "643b3977",
   "metadata": {},
   "source": [
    "These are numerical estimations on gates of a quantum processor. Single-qubit gates or SPAM operations have $n$ versions - one for each qubit. Two-qubit gates have $n (n-1)$ versions - one for each pair of qubits. We can save a dict containing this information into a .pickle file, and then extract these tensors somewhere else."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "57d60dfa",
   "metadata": {},
   "outputs": [],
   "source": [
    "fm.save_tensors_to_file(dict_result, ['tomo-tutorial'], prefix_args=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e6a85df6",
   "metadata": {},
   "source": [
    "The folder `data\\tomo-tutorial` should now appear, with a .pickle file containing the results of the experiment. Of course, you can instantly load them back:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "c5d2cb3b",
   "metadata": {},
   "outputs": [],
   "source": [
    "reuploaded_dict = fm.import_tensors_from_file(['tomo-tutorial'], prefix_args=1)\n",
    "assert reuploaded_dict.keys() == dict_result.keys()\n",
    "for key in reuploaded_dict:\n",
    "    assert tf.abs(tf.linalg.norm(reuploaded_dict[key] - dict_result[key])) < 1e-10"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "322c531f",
   "metadata": {},
   "source": [
    "Sometimes it is desirable to work with Choi matrices which are familiar to many people instead of some quantum channels denoted in some article. There are ways to convert the gate set into Choi matrices:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "a49f4a9d",
   "metadata": {},
   "outputs": [],
   "source": [
    "choi1 = util.choi_swap_1qchannel(dict_result['Rx'][0])\n",
    "choi2 = util.choi_swap_2qchannel(dict_result['CX'][0])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d155376",
   "metadata": {},
   "source": [
    "It may be useful to save not only the final results as tensors, but also the quantum circuits or generated bit strings. This becomes of critical importance when working with external sources of bit strings (e. g. running only the optimizer with pre-generated data). Below are some examples for working with such information."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "fdbe4cef",
   "metadata": {},
   "outputs": [],
   "source": [
    "fm.save_bitstrings_to_file(QC_t.samples_compressed, ['tomo-tutorial', 'comp'], prefix_args=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "3bfb5ca6",
   "metadata": {},
   "outputs": [],
   "source": [
    "fm.save_ncon_list_to_file(list(QC_t.tn_templates.values()), ['tomo-tutorial'], prefix_args=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "dad9aa17",
   "metadata": {},
   "outputs": [],
   "source": [
    "fm.save_ncon_list_to_file(circs + [[]], ['tomo-tutorial', 'human-form'], prefix_args=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f2738d0c",
   "metadata": {},
   "source": [
    "Congratulations, we hope you gained a bit of insight about this humble project and maybe even learned something useful."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8b1a472c",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}