{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "ac3a07af-2b66-41c8-8548-6f951460aedb",
   "metadata": {},
   "outputs": [],
   "source": [
    "from datasets import load_dataset\n",
    "\n",
    "ACCESS_TOKEN = \n",
    "NUM_SAMPLES = 500000\n",
    "\n",
    "dataset = load_dataset(\"uonlp/CulturaX\",\n",
    "                 \"ur\",\n",
    "                 split=f\"train[:{NUM_SAMPLES}]\",\n",
    "                 token = ACCESS_TOKEN\n",
    "                 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "b6515d96-1129-4aac-a670-796fee9302db",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Dataset({\n",
       "    features: ['text', 'timestamp', 'url', 'source'],\n",
       "    num_rows: 500000\n",
       "})"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "13831730-9fc1-4d89-b4fd-060ce0a976cb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Dataset({\n",
       "    features: ['text'],\n",
       "    num_rows: 500000\n",
       "})"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# remove columns other than text\n",
    "\n",
    "dataset = dataset.remove_columns([col for col in dataset.column_names if col != 'text'])\n",
    "dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "69466306-5190-4581-82fc-c5839bf15a80",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "500000"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(dataset)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "25249452-d545-45ce-8c47-a6ee4b20eee1",
   "metadata": {},
   "source": [
    "Curiously, I found out that number of rows counted using \"wc -l file.csv\" in a Linux terminal gives number of lines, not number of rows. See comment in https://stackoverflow.com/questions/32913151/is-it-possible-to-get-the-number-of-rows-in-a-csv-file-without-opening-it"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "68b6087b-0c27-4a9f-bc6c-1317a87c3f3f",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████████| 500000/500000 [00:31<00:00, 16019.97it/s]\n"
     ]
    }
   ],
   "source": [
    "from tqdm import tqdm\n",
    "\n",
    "for idx in tqdm(range(NUM_SAMPLES)):\n",
    "    with open(f'data/culturaX_ur_500k/ur_sample_{idx}.txt', 'w') as file:\n",
    "        file.write(dataset[idx][\"text\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a3bd917f-a70f-421d-9680-33ee676f193b",
   "metadata": {},
   "source": [
    "### Bengali"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "7c6a1722-7f4f-436b-a8b3-612f24483ee5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3d90843e65214727bf8ccf27c76caac9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Resolving data files:   0%|          | 0/18 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from datasets import load_dataset\n",
    "\n",
    "ACCESS_TOKEN = \n",
    "NUM_SAMPLES = 500000\n",
    "\n",
    "dataset = load_dataset(\"uonlp/CulturaX\",\n",
    "                 \"bn\",\n",
    "                 split=f\"train[:{NUM_SAMPLES}]\",\n",
    "                 token = ACCESS_TOKEN\n",
    "                 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "1e431914-1402-4ac7-94ed-d2427da318c8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Dataset({\n",
       "    features: ['text', 'timestamp', 'url', 'source'],\n",
       "    num_rows: 500000\n",
       "})"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "b2a55b87-a0e8-40ea-bef7-1349652337b7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Dataset({\n",
       "    features: ['text'],\n",
       "    num_rows: 500000\n",
       "})"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# remove columns other than text\n",
    "\n",
    "dataset = dataset.remove_columns([col for col in dataset.column_names if col != 'text'])\n",
    "dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "2f1b64c9-f568-42b2-81f5-fb6de504bcfc",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 500000/500000 [00:25<00:00, 19827.65it/s]\n"
     ]
    }
   ],
   "source": [
    "from tqdm import tqdm\n",
    "\n",
    "for idx in tqdm(range(NUM_SAMPLES)):\n",
    "    with open(f'data/culturaX_bn_500k/bn_sample_{idx}.txt', 'w') as file:\n",
    "        file.write(dataset[idx][\"text\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cc1b8c26-22da-4a5b-ba1e-1557a191c218",
   "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": "3.11.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}