{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.13","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"nvidiaTeslaT4","dataSources":[{"sourceId":70367,"databundleVersionId":9188054,"sourceType":"competition"}],"dockerImageVersionId":30747,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# Data Preparation\n\n![data.jpg](attachment:08b1c039-1e43-4b50-87fe-48444478348b.jpg)","metadata":{},"attachments":{"08b1c039-1e43-4b50-87fe-48444478348b.jpg":{"image/jpeg":"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"}}},{"cell_type":"code","source":"%%writefile preprocess.py\n\nimport pandas as pd\nimport numpy as np\nfrom tqdm import tqdm\nimport multiprocessing as mp\nfrom astropy.stats import sigma_clip\nimport torch\nimport torch.nn.functional as F\nimport os\n\n\nROOT = \"/kaggle/input/ariel-data-challenge-2024/\"\nVERSION = \"v26\"\nMODE = \"train\"\n\nBINNING = 1\n\nsensor_sizes_dict = {\n    \"AIRS-CH0\": [[11250, 32, 356], [32, 356]],\n    \"FGS1\": [[135000, 32, 32], [32, 32]],\n}  # input, mask\n\n# 16 center pixels, rest is just noise\ncl = 8\ncr = 24\n\n\ndef get_gain_offset(planet_id, sensor, mode):\n    \"\"\"\n    Get the gain and offset for a given planet and sensor\n    \"\"\"\n    gain_offset_csv = pd.read_csv(f\"{ROOT}/{mode}_adc_info.csv\")\n    planet_gain_offset = gain_offset_csv[gain_offset_csv[\"planet_id\"] == planet_id]\n\n    gain = planet_gain_offset[sensor + \"_adc_gain\"].values[0]\n    offset = planet_gain_offset[sensor + \"_adc_offset\"].values[0]\n    return gain, offset\n\n\ndef read_data(planet_id, sensor, mode):\n    \"\"\"\n    Read the data for a given planet and sensor\n    \"\"\"\n    # get all noise correction frames and signal\n    signal = pd.read_parquet(\n        f\"{ROOT}/{mode}/{planet_id}/{sensor}_signal.parquet\",\n        engine=\"pyarrow\",\n    )\n    dark_frame = pd.read_parquet(\n        f\"{ROOT}/{mode}/{planet_id}/{sensor}_calibration/dark.parquet\",\n        engine=\"pyarrow\",\n    )\n    dead_frame = pd.read_parquet(\n        f\"{ROOT}/{mode}/{planet_id}/{sensor}_calibration/dead.parquet\",\n        engine=\"pyarrow\",\n    )\n    linear_corr_frame = pd.read_parquet(\n        f\"{ROOT}/{mode}/{planet_id}/{sensor}_calibration/linear_corr.parquet\",\n        engine=\"pyarrow\",\n    )\n    flat_frame = pd.read_parquet(\n        f\"{ROOT}/{mode}/{planet_id}/{sensor}_calibration/flat.parquet\",\n        engine=\"pyarrow\",\n    )\n    # read_frame = pd.read_parquet(\n    #     f\"{ROOT}/{mode}/{planet_id}/{sensor}_calibration/read.parquet\",\n    #     engine=\"pyarrow\",\n    # )\n\n    # reshape to sensor shape and cast to float64\n    signal = signal.values.astype(np.float64).reshape(sensor_sizes_dict[sensor][0])[\n        :, cl:cr, :\n    ]\n    dark_frame = dark_frame.values.astype(np.float64).reshape(\n        sensor_sizes_dict[sensor][1]\n    )[cl:cr, :]\n    dead_frame = dead_frame.values.reshape(sensor_sizes_dict[sensor][1])[cl:cr, :]\n    flat_frame = flat_frame.values.astype(np.float64).reshape(\n        sensor_sizes_dict[sensor][1]\n    )[cl:cr, :]\n    # read_frame = read_frame.values.reshape(sensor_sizes_dict[sensor][1])\n    linear_corr = linear_corr_frame.values.astype(np.float64).reshape(\n        [6] + sensor_sizes_dict[sensor][1]\n    )[:, cl:cr, :]\n\n    return (\n        signal,\n        dark_frame,\n        dead_frame,\n        linear_corr,\n        flat_frame,\n        # read_frame,\n    )\n\n\ndef ADC_convert(signal, gain, offset):\n    \"\"\"\n    Step 1: Analog-to-Digital Conversion (ADC) correction\n\n    The Analog-to-Digital Conversion (adc) is performed by the detector to convert the\n    pixel voltage into an integer number. We revert this operation by using the gain\n    and offset for the calibration files 'train_adc_info.csv'.\n    \"\"\"\n\n    return signal / gain + offset\n\n\ndef mask_hot_dead(signal, dead, dark):\n    \"\"\"\n    Step 2: Mask hot/dead pixel\n\n    The dead pixels map is a map of the pixels that do not respond to light and, thus,\n    can't be accounted for any calculation. In all these frames the dead pixels are\n    masked using python masked arrays. The bad pixels are thus masked but left\n    uncorrected. Some methods can be used to correct bad-pixels but this task,\n    if needed, is left to the participants.\n    \"\"\"\n\n    hot = sigma_clip(dark, sigma=5, maxiters=5).mask\n    hot = np.tile(hot, (signal.shape[0], 1, 1))\n    dead = np.tile(dead, (signal.shape[0], 1, 1))\n\n    # Set values to np.nan where dead or hot pixels are found\n    signal[dead] = np.nan\n    signal[hot] = np.nan\n    return signal\n\n\ndef apply_linear_corr(c, signal):\n    \"\"\"\n    Step 3: linearity Correction\n\n    The non-linearity of the pixels' response can be explained as capacitive leakage\n    on the readout electronics of each pixel during the integration time. The number\n    of electrons in the well is proportional to the number of photons that hit the\n    pixel, with a quantum efficiency coefficient. However, the response of the pixel\n    is not linear with the number of electrons in the well. This effect can be\n    described by a polynomial function of the number of electrons actually in the well.\n    The data is provided with calibration files linear_corr.parquet that are the\n    coefficients of the inverse polynomial function and can be used to correct this\n    non-linearity effect.\n    Using horner's method to evaluate the polynomial\n    \"\"\"\n    assert c.shape[0] == 6  # Ensure the polynomial is of degree 5\n\n    return (\n        (((c[5] * signal + c[4]) * signal + c[3]) * signal + c[2]) * signal + c[1]\n    ) * signal + c[0]\n\n\ndef clean_dark(signal, dark, dt):\n    \"\"\"\n    Step 4: dark current subtraction\n\n    The data provided include calibration for dark current estimation, which can be\n    used to pre-process the observations. Dark current represents a constant signal\n    that accumulates in each pixel during the integration time, independent of the\n    incoming light. To obtain the corrected image, the following conventional approach\n    is applied: The data provided include calibration files such as dark frames or\n    dead pixels' maps. They can be used to pre-process the observations. The dark frame\n    is a map of the detector response to a very short exposure time, to correct for the\n    dark current of the detector.\n\n    image - (dark * dt)\n\n    The corrected image is conventionally obtained via the following: where the dark\n    current map is first corrected for the dead pixel.\n    \"\"\"\n\n    dark = torch.tile(dark, (signal.shape[0], 1, 1))\n    signal -= dark * dt[:, None, None]\n    return signal\n\n\ndef get_cds(signal):\n    \"\"\"\n    Step 5: Get Correlated Double Sampling (CDS)\n\n    The science frames are alternating between the start of the exposure and the end of\n    the exposure. The lecture scheme is a ramp with a double sampling, called\n    Correlated Double Sampling (CDS), the detector is read twice, once at the start\n    of the exposure and once at the end of the exposure. The final CDS is the\n    difference (End of exposure) - (Start of exposure).\n    \"\"\"\n\n    return torch.subtract(signal[1::2, :, :], signal[::2, :, :])\n\n\ndef bin_obs(signal, binning):\n    \"\"\"\n    Step 5.1: Bin Observations\n\n    The data provided are binned in the time dimension. The binning is performed by\n    summing the signal over the time dimension.\n    \"\"\"\n\n    assert signal.shape[0] % binning == 0  # Ensure the binning is possible\n\n    # cds_transposed = signal.transpose(0, 2, 1)\n    cds_binned = torch.zeros(\n        (\n            signal.shape[0] // binning,\n            signal.shape[1],\n            signal.shape[2],\n        ),\n        device=\"cuda:0\",\n    )\n    for i in range(signal.shape[0] // binning):\n        cds_binned[i, :, :] = torch.sum(\n            signal[i * binning : (i + 1) * binning, :, :], axis=0\n        )\n    return cds_binned\n\n\ndef correct_flat_field(flat, signal):\n    \"\"\"\n    Step 6: Flat Field Correction\n\n    The flat field is a map of the detector response to uniform illumination, to\n    correct for the pixel-to-pixel variations of the detector, for example the\n    different quantum efficiencies of each pixel.\n    \"\"\"\n\n    return signal / flat\n\n\ndef nan_interpolation(tensor):\n    # Assume tensor is of shape (batch, height, width)\n    nan_mask = torch.isnan(tensor)\n\n    # Replace NaNs with zero temporarily\n    tensor_filled = torch.where(\n        nan_mask, torch.tensor(0.0, device=tensor.device), tensor\n    )\n\n    # Create a binary mask (0 where NaNs were and 1 elsewhere)\n    ones = torch.ones_like(tensor, device=tensor.device)\n    weight = torch.where(nan_mask, torch.tensor(0.0, device=tensor.device), ones)\n\n    # Perform interpolation by convolving with a kernel\n    # using bilinear interpolation\n    kernel = torch.ones(1, 1, 1, 3, device=tensor.device, dtype=tensor.dtype)\n\n    # Apply padding to the tensor and weight to prevent boundary issues\n    tensor_padded = F.pad(\n        tensor_filled.unsqueeze(1), (1, 1, 0, 0), mode=\"replicate\"\n    ).squeeze(1)\n    weight_padded = F.pad(weight.unsqueeze(1), (1, 1, 0, 0), mode=\"replicate\").squeeze(\n        1\n    )\n\n    # Convolve the filled tensor and the weight mask\n    tensor_conv = F.conv2d(tensor_padded.unsqueeze(1), kernel, stride=1)\n    weight_conv = F.conv2d(weight_padded.unsqueeze(1), kernel, stride=1)\n\n    # Compute interpolated values (normalized by weights)\n    interpolated_tensor = tensor_conv / weight_conv\n\n    # Apply the interpolated values only to the positions of NaNs\n    result = torch.where(nan_mask, interpolated_tensor.squeeze(1), tensor)\n\n    return result\n\n\ndef process_planet(planet_id):\n    \"\"\"\n    Process a single planet's data\n    \"\"\"\n    axis_info = pd.read_parquet(ROOT + \"axis_info.parquet\")\n    dt_airs = axis_info[\"AIRS-CH0-integration_time\"].dropna().values\n\n    for sensor in [\"FGS1\", \"AIRS-CH0\"]:\n        # load all data for this planet and sensor\n        signal, dark_frame, dead_frame, linear_corr, flat_frame = read_data(\n            planet_id, sensor, mode=MODE\n        )\n        gain, offset = get_gain_offset(planet_id, sensor, mode=MODE)\n\n        # Step 1: ADC correction\n        signal = ADC_convert(signal, gain, offset)\n\n        # Step 2: Mask hot/dead pixel\n        signal = mask_hot_dead(signal, dead_frame, dark_frame)\n\n        # clip at 0\n        signal = signal.clip(0)\n\n        # Step 3: linearity Correction\n        signal = apply_linear_corr(\n            torch.tensor(linear_corr).to(\"cuda:0\"), torch.tensor(signal).to(\"cuda:0\")\n        )\n\n        # Step 4: dark current subtraction\n        if sensor == \"FGS1\":\n            dt = torch.ones(len(signal), device=\"cuda:0\") * 0.1\n            dt[1::2] += 4.5\n        elif sensor == \"AIRS-CH0\":\n            dt = torch.tensor(dt_airs).to(\"cuda:0\")\n            dt[1::2] += 0.1\n\n        signal = clean_dark(signal, torch.tensor(dark_frame).to(\"cuda:0\"), dt)\n\n        # Step 5: Get Correlated Double Sampling (CDS)\n        signal = get_cds(signal)\n\n        # Step 5.1: Bin Observations\n        if sensor == \"FGS1\":\n            signal = bin_obs(signal, binning=BINNING * 12)\n        elif sensor == \"AIRS-CH0\":\n            signal = bin_obs(signal, binning=BINNING)\n\n        # Step 6: Flat Field Correction\n        signal = correct_flat_field(torch.tensor(flat_frame).to(\"cuda:0\"), signal)\n\n        # Step 7: Interpolate NaNs (twice!)\n        signal = nan_interpolation(signal)\n        signal = nan_interpolation(signal)\n\n        # Step 7.1: Clip at 0 again\n        signal = signal.clip(0)\n\n        # Step 8: Sum over spatial axis\n        if sensor == \"FGS1\":\n            signal = torch.nanmean(signal, axis=[1, 2]).cpu().numpy()\n        elif sensor == \"AIRS-CH0\":\n            signal = torch.nanmean(signal, axis=1).cpu().numpy()\n\n        # save the processed signal\n        np.save(\n            str(planet_id) + \"_\" + sensor + f\"_signal_{VERSION}.npy\",\n            signal.astype(np.float64),\n        )\n\n\nif __name__ == \"__main__\":\n    adc_info = pd.read_csv(f\"{ROOT}/{MODE}_adc_info.csv\", index_col=\"planet_id\")\n    planet_ids = adc_info.index.tolist()\n\n    with mp.Pool(processes=4) as pool:\n        list(tqdm(pool.imap(process_planet, planet_ids), total=len(planet_ids)))\n\n    signal_train = []\n\n    for planet_id in planet_ids:\n        f_raw = np.load(str(planet_id) + f\"_FGS1_signal_{VERSION}.npy\")\n        a_raw = np.load(str(planet_id) + f\"_AIRS-CH0_signal_{VERSION}.npy\")\n\n        # flip a_raw\n        signal = np.concatenate([f_raw[:, None], a_raw], axis=1)\n        signal_train.append(signal)\n\n        # free space from disk\n        os.remove(str(planet_id) + f\"_FGS1_signal_{VERSION}.npy\")\n        os.remove(str(planet_id) + f\"_AIRS-CH0_signal_{VERSION}.npy\")\n\n    signal_train = np.array(signal_train)\n    np.save(f\"{MODE}_signal_{VERSION}.npy\", signal_train, allow_pickle=False)\n\n    print(\"Processing complete!\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-11-07T09:48:34.004581Z","iopub.execute_input":"2024-11-07T09:48:34.004947Z","iopub.status.idle":"2024-11-07T09:48:34.031966Z","shell.execute_reply.started":"2024-11-07T09:48:34.004909Z","shell.execute_reply":"2024-11-07T09:48:34.030632Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"!python preprocess.py","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-11-07T09:48:34.091513Z","iopub.execute_input":"2024-11-07T09:48:34.092191Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null}]}