{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "addd199c-097c-419d-a0f2-c3d73efb8d5d",
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import os\n",
    "from pprint import pprint\n",
    "\n",
    "import bitsandbytes as bnb\n",
    "import pandas as pd\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "\n",
    "import transformers\n",
    "from datasets import load_dataset\n",
    "from huggingface_hub import notebook_login\n",
    "from peft import (\n",
    "    LoraConfig,\n",
    "    PeftConfig,\n",
    "    PeftModel,\n",
    "    get_peft_model,\n",
    "    prepare_model_for_kbit_training,\n",
    ")\n",
    "from transformers import (\n",
    "    AutoConfig,\n",
    "    AutoModelForCausalLM,\n",
    "    AutoTokenizer,\n",
    "    BitsAndBytesConfig,\n",
    ")\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"0\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "acfb1578-a66f-44f0-8df9-1c6bcf7530ea",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b92bb6f7a2784be8bf5cab2ee87292ff",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "VBox(children=(HTML(value='<center> <img\\nsrc=https://huggingface.co/front/assets/huggingface_logo-noborder.sv…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "notebook_login()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "d2f13cac-1536-4da0-8ff7-0a0454fd0b4a",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"ds1000-test-cleaned.json\") as json_file:\n",
    "    data = json.load(json_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "6706e68b-d525-4392-ab2c-1dff356da52d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'answer': 'import pandas as pd\\n'\n",
      "           '\\n'\n",
      "           '\\n'\n",
      "           'index = range(14)\\n'\n",
      "           'data = [1, 0, 0, 2, 0, 4, 6, 8, 0, 0, 0, 0, 2, 1]\\n'\n",
      "           \"df = pd.DataFrame(data=data, index=index, columns = ['A'])\\n\"\n",
      "           'def g(df):\\n'\n",
      "           \"    l = df['A'].replace(to_replace=0, method='ffill')\\n\"\n",
      "           \"    r = df['A'].replace(to_replace=0, method='bfill')\\n\"\n",
      "           '    for i in range(len(df)):\\n'\n",
      "           \"        df['A'].iloc[i] = max(l[i], r[i])\\n\"\n",
      "           '    return df\\n'\n",
      "           '\\n'\n",
      "           'df = g(df.copy())\\n'\n",
      "           'result = df\\n'\n",
      "           'print(result)',\n",
      " 'question': 'Problem:\\n'\n",
      "             'I have the following dataframe:\\n'\n",
      "             'index = range(14)\\n'\n",
      "             'data = [1, 0, 0, 2, 0, 4, 6, 8, 0, 0, 0, 0, 2, 1]\\n'\n",
      "             \"df = pd.DataFrame(data=data, index=index, columns = ['A'])\\n\"\n",
      "             '\\n'\n",
      "             '\\n'\n",
      "             'How can I fill the zeros with the maximun between previous and '\n",
      "             'posterior non-zero value using pandas? Is there a fillna that is '\n",
      "             'not just for \"NaN\"?.  \\n'\n",
      "             'The output should look like:\\n'\n",
      "             '    A\\n'\n",
      "             '0   1\\n'\n",
      "             '1   2\\n'\n",
      "             '2   2\\n'\n",
      "             '3   2\\n'\n",
      "             '4   4\\n'\n",
      "             '5   4\\n'\n",
      "             '6   6\\n'\n",
      "             '7   8\\n'\n",
      "             '8   8\\n'\n",
      "             '9   8\\n'\n",
      "             '10  8\\n'\n",
      "             '11  8\\n'\n",
      "             '12  2\\n'\n",
      "             '13  1'}\n"
     ]
    }
   ],
   "source": [
    "pprint(data[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "9cc4983a-9a3f-485f-983f-efe2f10ce516",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"ds1000-test-cleaned.json\", \"w\") as f:\n",
    "    json.dump(data, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "f45c3674-4eed-4ca5-8343-2184ff1e4da1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>question</th>\n",
       "      <th>answer</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Problem:\\nI have the following dataframe:\\nind...</td>\n",
       "      <td>import pandas as pd\\n\\n\\nindex = range(14)\\nda...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Problem:\\ni got an issue over ranking of date ...</td>\n",
       "      <td>import pandas as pd\\n\\n\\ndf = pd.DataFrame({'I...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Problem:\\nI have a DataFrame like :\\n     0   ...</td>\n",
       "      <td>import pandas as pd\\nimport numpy as np\\n\\ndf ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Problem:\\nI have this Pandas dataframe (df):\\n...</td>\n",
       "      <td>import pandas as pd\\n\\n\\ndf = pd.DataFrame({'A...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Problem:\\nI have\\n\\ndf = pd.DataFrame.from_dic...</td>\n",
       "      <td>import pandas as pd\\n\\ndf = pd.DataFrame.from_...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                            question  \\\n",
       "0  Problem:\\nI have the following dataframe:\\nind...   \n",
       "1  Problem:\\ni got an issue over ranking of date ...   \n",
       "2  Problem:\\nI have a DataFrame like :\\n     0   ...   \n",
       "3  Problem:\\nI have this Pandas dataframe (df):\\n...   \n",
       "4  Problem:\\nI have\\n\\ndf = pd.DataFrame.from_dic...   \n",
       "\n",
       "                                              answer  \n",
       "0  import pandas as pd\\n\\n\\nindex = range(14)\\nda...  \n",
       "1  import pandas as pd\\n\\n\\ndf = pd.DataFrame({'I...  \n",
       "2  import pandas as pd\\nimport numpy as np\\n\\ndf ...  \n",
       "3  import pandas as pd\\n\\n\\ndf = pd.DataFrame({'A...  \n",
       "4  import pandas as pd\\n\\ndf = pd.DataFrame.from_...  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(data).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "6fbdd3ad-062f-4744-bb8e-1c19950adfd5",
   "metadata": {},
   "outputs": [],
   "source": [
    "bnb_config = BitsAndBytesConfig(\n",
    "    load_in_4bit=True,\n",
    "    bnb_4bit_use_double_quant=True,\n",
    "    bnb_4bit_quant_type=\"nf4\",\n",
    "    bnb_4bit_compute_dtype=torch.bfloat16,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "2b5ae38c-b0d2-4b9a-acde-3370130ca6e7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3d6c5533e9ea48e295b7fdfd96da6d47",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Loading checkpoint shards:   0%|          | 0/2 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Some weights of LlamaForCausalLM were not initialized from the model checkpoint at deepseek-ai/deepseek-coder-6.7b-instruct and are newly initialized: ['model.layers.16.self_attn.rotary_emb.inv_freq', 'model.layers.11.self_attn.rotary_emb.inv_freq', 'model.layers.13.self_attn.rotary_emb.inv_freq', 'model.layers.27.self_attn.rotary_emb.inv_freq', 'model.layers.28.self_attn.rotary_emb.inv_freq', 'model.layers.10.self_attn.rotary_emb.inv_freq', 'model.layers.14.self_attn.rotary_emb.inv_freq', 'model.layers.24.self_attn.rotary_emb.inv_freq', 'model.layers.3.self_attn.rotary_emb.inv_freq', 'model.layers.9.self_attn.rotary_emb.inv_freq', 'model.layers.29.self_attn.rotary_emb.inv_freq', 'model.layers.6.self_attn.rotary_emb.inv_freq', 'model.layers.8.self_attn.rotary_emb.inv_freq', 'model.layers.22.self_attn.rotary_emb.inv_freq', 'model.layers.0.self_attn.rotary_emb.inv_freq', 'model.layers.25.self_attn.rotary_emb.inv_freq', 'model.layers.12.self_attn.rotary_emb.inv_freq', 'model.layers.26.self_attn.rotary_emb.inv_freq', 'model.layers.2.self_attn.rotary_emb.inv_freq', 'model.layers.31.self_attn.rotary_emb.inv_freq', 'model.layers.1.self_attn.rotary_emb.inv_freq', 'model.layers.4.self_attn.rotary_emb.inv_freq', 'model.layers.23.self_attn.rotary_emb.inv_freq', 'model.layers.15.self_attn.rotary_emb.inv_freq', 'model.layers.7.self_attn.rotary_emb.inv_freq', 'model.layers.21.self_attn.rotary_emb.inv_freq', 'model.layers.20.self_attn.rotary_emb.inv_freq', 'model.layers.19.self_attn.rotary_emb.inv_freq', 'model.layers.30.self_attn.rotary_emb.inv_freq', 'model.layers.18.self_attn.rotary_emb.inv_freq', 'model.layers.17.self_attn.rotary_emb.inv_freq', 'model.layers.5.self_attn.rotary_emb.inv_freq']\n",
      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
     ]
    }
   ],
   "source": [
    "PEFT_MODEL = \"shanjay/ds-dsc\"\n",
    "\n",
    "config = PeftConfig.from_pretrained(PEFT_MODEL)\n",
    "model = AutoModelForCausalLM.from_pretrained(\n",
    "    config.base_model_name_or_path,\n",
    "    return_dict=True,\n",
    "    quantization_config=bnb_config,\n",
    "    device_map=\"auto\",\n",
    "    trust_remote_code=True,\n",
    ")\n",
    "tokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path)\n",
    "tokenizer.pad_token = tokenizer.eos_token\n",
    "\n",
    "model = PeftModel.from_pretrained(model, PEFT_MODEL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "7c3e35e0-f77c-4d63-8e2b-e72027341e31",
   "metadata": {},
   "outputs": [],
   "source": [
    "generation_config = model.generation_config\n",
    "generation_config.max_new_tokens = 200\n",
    "generation_config.temperature = 0.7\n",
    "generation_config.top_p = 0.7\n",
    "generation_config.num_return_sequences = 1\n",
    "generation_config.pad_token_id = tokenizer.eos_token_id\n",
    "generation_config.eos_token_id = tokenizer.eos_token_id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "aee4385b-d855-4225-9532-4e9002322579",
   "metadata": {},
   "outputs": [],
   "source": [
    "DEVICE = \"cuda:0\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "7b14a1c6-ac62-4a9c-9df9-0db50facfd7e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<instruction>: How can I create a dataframe?\n",
      "<output>:\n",
      "import pandas as pd\n",
      "import numpy as np\n",
      "\n",
      "df = pd.DataFrame(\n",
      "    {\n",
      "        \"A\": np.array([1, 2, 3]),\n",
      "        \"B\": np.array([4, 5, 6]),\n",
      "        \"C\": np.array([7, 8, 9]),\n",
      "    }\n",
      ")\n",
      "</output>\n",
      "BEGIN SOLUTION\n",
      "<output>\n",
      "[1]\n",
      "<code>\n",
      "[python]\n",
      "# Your code here\n",
      "</code>\n",
      "</output>\n",
      "END SOLUTION\n",
      "<output>\n",
      "[1]\n",
      "<code>\n",
      "[python]\n",
      "print(df)\n",
      "</code>\n",
      "</output>\n",
      "\n",
      "<assistant>: df = pd.DataFrame(\n",
      "    {\n",
      "        \"A\": np.array([1, 2, 3]),\n",
      "CPU times: user 27.4 s, sys: 372 ms, total: 27.8 s\n",
      "Wall time: 27.9 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "prompt = f\"\"\"\n",
    "<instruction>: How can I create a dataframe?\n",
    "<output>:\n",
    "\"\"\".strip()\n",
    "\n",
    "encoding = tokenizer(prompt, return_tensors=\"pt\").to(DEVICE)\n",
    "with torch.inference_mode():\n",
    "    outputs = model.generate(\n",
    "        input_ids=encoding.input_ids,\n",
    "        attention_mask=encoding.attention_mask,\n",
    "        generation_config=generation_config,\n",
    "    )\n",
    "print(tokenizer.decode(outputs[0], skip_special_tokens=True))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "93c95988-c563-4871-974d-004bf73fbce8",
   "metadata": {},
   "outputs": [],
   "source": [
    "def generate_response(question: str) -> str:\n",
    "    prompt = f\"\"\"\n",
    "<instruction>: {question}\n",
    "<output>:\n",
    "\"\"\".strip()\n",
    "    encoding = tokenizer(prompt, return_tensors=\"pt\").to(DEVICE)\n",
    "    with torch.inference_mode():\n",
    "        outputs = model.generate(\n",
    "            input_ids=encoding.input_ids,\n",
    "            attention_mask=encoding.attention_mask,\n",
    "            generation_config=generation_config,\n",
    "        )\n",
    "    response = tokenizer.decode(outputs[0], skip_special_tokens=True)\n",
    "\n",
    "    assistant_start = \"<output>:\"\n",
    "    response_start = response.find(assistant_start)\n",
    "    return response[response_start + len(assistant_start) :].strip()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "8a9a9b87-193b-4bed-8ef1-57944d931958",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "import pandas as pd\n",
      "import numpy as np\n",
      "\n",
      "df = pd.DataFrame(\n",
      "    {\n",
      "        \"A\": np.array([1, 2, 3]),\n",
      "        \"B\": np.array([4, 5, 6]),\n",
      "        \"C\": np.array([7, 8, 9]),\n",
      "    }\n",
      ")\n",
      "</output>\n",
      "BEGIN SOLUTION\n",
      "<output>\n",
      "[1]\n",
      "<code>\n",
      "[python]\n",
      "# Your code here\n",
      "</code>\n",
      "</output>\n",
      "END SOLUTION\n",
      "<output>\n",
      "[1]\n",
      "<code>\n",
      "[python]\n",
      "print(df)\n",
      "</code>\n",
      "</output>\n",
      "\n",
      "<assistant>: df = pd.DataFrame(\n",
      "    {\n",
      "        \"A\": np.array([1, 2, 3]),\n"
     ]
    }
   ],
   "source": [
    "prompt = \"How can I create a dataframe?\"\n",
    "print(generate_response(prompt))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "4658f305-b7c6-432c-ac0c-f62bd79e9ad5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "import pandas as pd\n",
      "import numpy as np\n",
      "\n",
      "df1 = pd.DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})\n",
      "df2 = pd.DataFrame({'A': [4, 5, 6], 'B': [7, 8, 9]})\n",
      "</output>\n",
      "<assistant>: df = pd.concat([df1, df2])\n",
      "</assistant>\n",
      "<output>: df\n",
      "</output>\n",
      "<code>\n",
      "import pandas as pd\n",
      "import numpy as np\n",
      "\n",
      "df1 = pd.DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})\n",
      "df2 = pd.DataFrame({'A': [4, 5, 6],\n"
     ]
    }
   ],
   "source": [
    "prompt = \"How to merge two dataframes?\"\n",
    "print(generate_response(prompt))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "0e9ed231-4a62-4331-94df-f3bcd601f138",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<code>\n",
      "import pandas as pd\n",
      "import numpy as np\n",
      "\n",
      "name=np.array(['joy','shan'])\n",
      "roll_no=np.array([1,2])\n",
      "</code>\n",
      "BEGIN SOLUTION\n",
      "<code>\n",
      "[insert]\n",
      "</code>\n",
      "END SOLUTION\n",
      "<code>\n",
      "print(df)\n",
      "</code>\n",
      "<assistant>: df = pd.DataFrame({'name': name, 'roll_no': roll_no})\n",
      "</assistant>\n",
      "<code>\n",
      "print(df)\n",
      "</code>\n",
      "\n",
      "<assistant>: df = pd.DataFrame({'name': name, 'roll_no': roll_no})\n",
      "print(df)\n",
      "</assistant>\n",
      "<code>\n",
      "print(df)\n",
      "</code>\n",
      "<assistant>: df = pd.\n"
     ]
    }
   ],
   "source": [
    "prompt = \"given two arrays name=['joy','shan'], roll_no=[1,2]. put these array in a dataframe ?\"\n",
    "print(generate_response(prompt))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "381ba5c0-276d-411e-a8d5-9f010528433d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ ]: import matplotlib.pyplot as plt\n",
      "import numpy as np\n",
      "\n",
      "x = np.linspace(0, 10, 100)\n",
      "y = np.sin(x)\n",
      "\n",
      "# your code here\n",
      "</output>\n",
      "<assistant>: plt.plot(x, y)\n",
      "plt.show()\n",
      "</assistant>\n",
      "<output>: [ ]: plt.plot(x, y)\n",
      "plt.show()\n",
      "</output>\n",
      "<assistant>: plt.plot(x, y)\n",
      "plt.show()\n",
      "</assistant>\n",
      "<output>: [ ]: plt.plot(x, y)\n",
      "plt.show()\n",
      "</output>\n",
      "<assistant>: plt.plot(x, y)\n",
      "plt.show()\n",
      "</assistant>\n",
      "<output>\n"
     ]
    }
   ],
   "source": [
    "prompt = \"can you plot all types of plots in matplotlib?\"\n",
    "print(generate_response(prompt))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "6864c3c7-b721-48ca-8943-dcff9838f7d2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "import pandas as pd\n",
      "import numpy as np\n",
      "\n",
      "data = pd.DataFrame({'ID': ['01', '01', '01', '02', '02'],\n",
      "'TIME': ['2018-07-11 11:12:20', '2018-07-12 12:00:23', '2018-07-13 12:00:00', '2019-09-11 11:00:00', '2019-09-12 12:00:00']})\n",
      "\n",
      "data['TIME'] = pd.to_datetime(data['TIME'])\n",
      "\n",
      "</output>\n",
      "BEGIN SOLUTION\n",
      "<output>\n",
      "[insert]\n",
      "</output>\n"
     ]
    }
   ],
   "source": [
    "prompt = \"\"\"Problem:\n",
    "i got an issue over ranking of date times. Lets say i have following table.\n",
    "ID    TIME\n",
    "01    2018-07-11 11:12:20\n",
    "01    2018-07-12 12:00:23\n",
    "01    2018-07-13 12:00:00\n",
    "02    2019-09-11 11:00:00\n",
    "02    2019-09-12 12:00:00\n",
    "\n",
    "\n",
    "and i want to add another column to rank the table by time for each id and group. I used \n",
    "df['RANK'] = data.groupby('ID')['TIME'].rank(ascending=True)\n",
    "\n",
    "\n",
    "but get an error:\n",
    "'NoneType' object is not callable\n",
    "\n",
    "\n",
    "If i replace datetime to numbers, it works.... any solutions?\n",
    "\"\"\"\n",
    "print(generate_response(prompt))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "7fa02929-5c65-4aa6-81ce-9c51879e7535",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<code>\n",
      "import pandas as pd\n",
      "import numpy as np\n",
      "\n",
      "index = range(14)\n",
      "data = [1, 0, 0, 2, 0, 4, 6, 8, 0, 0, 0, 0, 2, 1]\n",
      "df = pd.DataFrame(data=data, index=index, columns = ['A'])\n",
      "</code>\n",
      "BEGIN SOLUTION\n",
      "<code>\n",
      "[insert]\n",
      "</code>\n",
      "END SOLUTION\n",
      "<code>\n",
      "print(df)\n",
      "</code>\n",
      "\n",
      "<assistant>: df['A'] = df['A'].replace(0, np.nan)\n",
      "df['A'] = df['A'].fillna(method='ffill')\n",
      "df['A'] = df['A'].fillna(method='bfill')\n"
     ]
    }
   ],
   "source": [
    "prompt = \"\"\"Problem:\n",
    "I have the following dataframe:\n",
    "index = range(14)\n",
    "data = [1, 0, 0, 2, 0, 4, 6, 8, 0, 0, 0, 0, 2, 1]\n",
    "df = pd.DataFrame(data=data, index=index, columns = ['A'])\n",
    "\n",
    "\n",
    "How can I fill the zeros with the maximun between previous and posterior non-zero value using pandas? Is there a fillna that is not just for \"NaN\"?.  \n",
    "The output should look like:\n",
    "    A\n",
    "0   1\n",
    "1   2\n",
    "2   2\n",
    "3   2\n",
    "4   4\n",
    "5   4\n",
    "6   6\n",
    "7   8\n",
    "8   8\n",
    "9   8\n",
    "10  8\n",
    "11  8\n",
    "12  2\n",
    "13  1\n",
    "\"\"\"\n",
    "\n",
    "print(generate_response(prompt))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "255cc021-5f5e-46af-a75e-a435b9629cdf",
   "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": "3.10.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}