docs/code-level-rca.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Using RCAEval to Diagnose Code-Level Faults\n",
    "\n",
    "This notebook demonstrates how RCAEval identifies the root cause service and fine-grained root cause indicators of a code-level fault. In this demonstration, we transform logs to time series and feed them into BARO to diagnose the root cause.\n",
    "\n",
    "**Bug Description**: We injected a code-level fault, namely **Incorrect parameter values** (as described in [D]), into the `cartservice` of the Online Boutique system. Empirical studies show that **Incorrect parameter values** is one of the five most common faults in real-world projects and that injected faults can realistically simulate real software faults [A,B,C,D].\n",
    "\n",
    "Specifically, we triggered this code-level fault by using an extremely large number when calling the function `AddItemAsync` in the file `RedisCartStore.cs` (line 54) to add items to a cart, e.g., attempting to add 100000000000 items.\n",
    "\n",
    "**Symptoms**: After injecting this fault, the `cartservice` became unstable, leading to failures in all its callers. The `front-end` services reported error codes. Resource usage of the `cartservice` spiked, with other services such as `recommendationservice` and `shippingservice` also affected.\n",
    "\n",
    "**Root cause**: The root cause service is the `cartservice`. The fine-grained root cause indicator is the `stack trace` presented below.\n",
    "\n",
    "---\n",
    "```bash\n",
    "info: Grpc.AspNetCore.Server.ServerCallHandler[7]\n",
    "      Error status code 'FailedPrecondition' with detail 'Can't access cart storage. System.OverflowException: Value was either too large or too small for an Int32.\n",
    "         at System.Number.ThrowOverflowException[TInteger]()\n",
    "         at cartservice.cartstore.RedisCartStore.AddItemAsync(String, String, String) in /app/cartstore/RedisCartStore.cs:line 54' raised.   \n",
    "```\n",
    "---\n",
    "**Step-by-Step: Using RCAEval to Diagnose the Root Cause**:\n",
    "\n",
    "Below are the steps taken to diagnose this failure:\n",
    "\n",
    "1. Read available logs.\n",
    "2. Parse logs into time series and visualize them.\n",
    "3. Apply BARO [E].\n",
    "\n",
    "\n",
    "References:\n",
    "- [A] Pham, et al. \"Failure diagnosis for distributed systems using targeted fault injection.\" In:TPDS’16.\n",
    "- [B] Andrews, et al. \"Is mutation an appropriate tool for testing experiments?.\" In: ICSE’05.\n",
    "- [C] Christmansson, et al. \"Generation of an error set that emulates software faults based on field data.\" In: FTCS’96.\n",
    "- [D] Cotroneo, et al. \"How bad can a bug get? an empirical analysis of software failures in the openstack cloud computing platform.\" In: FSE’19.\n",
    "- [E] Pham, et al. \"BARO: Robust root cause analysis for microservices via multivariate bayesian online change point detection.\" In: FSE'24."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd \n",
    "import numpy as np\n",
    "import warnings \n",
    "warnings.filterwarnings(\"ignore\")\n",
    "import matplotlib.pyplot as plt\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Read available logs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "logs = pd.read_csv(\"data/RE3/RE3-OB/cartservice_f1/2/logs.csv\")\n",
    "ta = 1732279831"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Parse logs into time series\n",
    "Parse logs into time series and plot them."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>time</th>\n",
       "      <th>timestamp</th>\n",
       "      <th>container_name</th>\n",
       "      <th>message</th>\n",
       "      <th>pod_name</th>\n",
       "      <th>node_name</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>12:38</td>\n",
       "      <td>1732279111001644531</td>\n",
       "      <td>adservice</td>\n",
       "      <td>received ad request (context_words=[kitchen])</td>\n",
       "      <td>adservice-b9bcc6cc9-rb9qn</td>\n",
       "      <td>ip-192-168-6-79.ap-southeast-2.compute.internal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>12:38</td>\n",
       "      <td>1732279111003879571</td>\n",
       "      <td>frontend</td>\n",
       "      <td>request complete</td>\n",
       "      <td>frontend-588fbcd66b-fjnp2</td>\n",
       "      <td>ip-192-168-83-253.ap-southeast-2.compute.internal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>12:38</td>\n",
       "      <td>1732279111365545236</td>\n",
       "      <td>frontend</td>\n",
       "      <td>request started</td>\n",
       "      <td>frontend-588fbcd66b-fjnp2</td>\n",
       "      <td>ip-192-168-83-253.ap-southeast-2.compute.internal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>12:38</td>\n",
       "      <td>1732279111365568125</td>\n",
       "      <td>frontend</td>\n",
       "      <td>setting currency</td>\n",
       "      <td>frontend-588fbcd66b-fjnp2</td>\n",
       "      <td>ip-192-168-83-253.ap-southeast-2.compute.internal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>12:38</td>\n",
       "      <td>1732279111365587350</td>\n",
       "      <td>frontend</td>\n",
       "      <td>request complete</td>\n",
       "      <td>frontend-588fbcd66b-fjnp2</td>\n",
       "      <td>ip-192-168-83-253.ap-southeast-2.compute.internal</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    time            timestamp container_name   \n",
       "0  12:38  1732279111001644531      adservice  \\\n",
       "1  12:38  1732279111003879571       frontend   \n",
       "2  12:38  1732279111365545236       frontend   \n",
       "3  12:38  1732279111365568125       frontend   \n",
       "4  12:38  1732279111365587350       frontend   \n",
       "\n",
       "                                         message                   pod_name   \n",
       "0  received ad request (context_words=[kitchen])  adservice-b9bcc6cc9-rb9qn  \\\n",
       "1                               request complete  frontend-588fbcd66b-fjnp2   \n",
       "2                                request started  frontend-588fbcd66b-fjnp2   \n",
       "3                               setting currency  frontend-588fbcd66b-fjnp2   \n",
       "4                               request complete  frontend-588fbcd66b-fjnp2   \n",
       "\n",
       "                                           node_name  \n",
       "0    ip-192-168-6-79.ap-southeast-2.compute.internal  \n",
       "1  ip-192-168-83-253.ap-southeast-2.compute.internal  \n",
       "2  ip-192-168-83-253.ap-southeast-2.compute.internal  \n",
       "3  ip-192-168-83-253.ap-southeast-2.compute.internal  \n",
       "4  ip-192-168-83-253.ap-southeast-2.compute.internal  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logs.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# separate normal logs and stack traces, stack traces has more than 1 line\n",
    "normal_logs = logs[logs['message'].apply(lambda x: len(x.split(\"\\n\")) == 1)]\n",
    "stack_traces = logs[logs['message'].apply(lambda x: len(x.split(\"\\n\")) > 1)]\n",
    "\n",
    "df_dict = {}\n",
    "for svc in logs['container_name'].unique():\n",
    "    svc_logs = logs[logs['container_name'] == svc]\n",
    "    svc_normal_logs = svc_logs[svc_logs['message'].apply(lambda x: len(x.split(\"\\n\")) == 1)]\n",
    "    svc_stack_traces = svc_logs[svc_logs['message'].apply(lambda x: len(x.split(\"\\n\")) > 1)]\n",
    "    \n",
    "    svc_normal_logs['timestamp'] = pd.to_datetime(svc_normal_logs['timestamp'])\n",
    "    svc_stack_traces['timestamp'] = pd.to_datetime(svc_stack_traces['timestamp'])\n",
    "    svc_normal_logs['timestamp'] = svc_normal_logs['timestamp'].dt.floor('15s')\n",
    "    svc_stack_traces['timestamp'] = svc_stack_traces['timestamp'].dt.floor('15s')\n",
    "    \n",
    "    svc_normal_logs = svc_normal_logs['timestamp'].value_counts().sort_index()\n",
    "    svc_stack_traces = svc_stack_traces['timestamp'].value_counts().sort_index()\n",
    "    svc_normal_logs = svc_normal_logs.resample('15s').sum().fillna(0)\n",
    "    svc_stack_traces = svc_stack_traces.resample('15s').sum().fillna(0)\n",
    "    \n",
    "    svc_normal_logs = pd.DataFrame(svc_normal_logs)\n",
    "    svc_stack_traces = pd.DataFrame(svc_stack_traces)\n",
    "    \n",
    "    df_dict[f\"{svc}_normal\"] = svc_normal_logs\n",
    "    df_dict[f\"{svc}_stack\"] = svc_stack_traces\n",
    "\n",
    "# merge df_dict into 1 df by time\n",
    "logs_df = pd.DataFrame({'time': df_dict['cartservice_normal'].index})\n",
    "logs_df = logs_df.set_index('time')\n",
    "for key in df_dict.keys():\n",
    "    df = df_dict[key]\n",
    "    df.columns = [key]\n",
    "    logs_df = logs_df.join(df, how='outer')\n",
    "# fill na \n",
    "logs_df = logs_df.fillna(0)\n",
    "\n",
    "# plot, set legend outside of the plot\n",
    "# import matplotlib.pyplot as plt\n",
    "# fig, ax = plt.subplots()\n",
    "# logs_df.plot(ax=ax)\n",
    "# ax.legend(loc='center left', bbox_to_anchor=(1, 0.5))\n",
    "# plt.show()\n",
    "# logs_df[\"cartservice_stack\"].plot()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 4. Apply BARO"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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>time</th>\n",
       "      <th>adservice_normal</th>\n",
       "      <th>adservice_stack</th>\n",
       "      <th>frontend_normal</th>\n",
       "      <th>frontend_stack</th>\n",
       "      <th>currencyservice_normal</th>\n",
       "      <th>currencyservice_stack</th>\n",
       "      <th>cartservice_normal</th>\n",
       "      <th>cartservice_stack</th>\n",
       "      <th>recommendationservice_normal</th>\n",
       "      <th>...</th>\n",
       "      <th>shippingservice_normal</th>\n",
       "      <th>shippingservice_stack</th>\n",
       "      <th>checkoutservice_normal</th>\n",
       "      <th>checkoutservice_stack</th>\n",
       "      <th>paymentservice_normal</th>\n",
       "      <th>paymentservice_stack</th>\n",
       "      <th>emailservice_normal</th>\n",
       "      <th>emailservice_stack</th>\n",
       "      <th>redis_normal</th>\n",
       "      <th>redis_stack</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1732279110</td>\n",
       "      <td>35</td>\n",
       "      <td>0.0</td>\n",
       "      <td>192</td>\n",
       "      <td>0.0</td>\n",
       "      <td>223</td>\n",
       "      <td>0.0</td>\n",
       "      <td>55</td>\n",
       "      <td>0.0</td>\n",
       "      <td>37</td>\n",
       "      <td>...</td>\n",
       "      <td>34.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1732279125</td>\n",
       "      <td>51</td>\n",
       "      <td>0.0</td>\n",
       "      <td>262</td>\n",
       "      <td>0.0</td>\n",
       "      <td>264</td>\n",
       "      <td>0.0</td>\n",
       "      <td>75</td>\n",
       "      <td>0.0</td>\n",
       "      <td>53</td>\n",
       "      <td>...</td>\n",
       "      <td>30.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1732279140</td>\n",
       "      <td>43</td>\n",
       "      <td>0.0</td>\n",
       "      <td>258</td>\n",
       "      <td>0.0</td>\n",
       "      <td>259</td>\n",
       "      <td>0.0</td>\n",
       "      <td>79</td>\n",
       "      <td>0.0</td>\n",
       "      <td>59</td>\n",
       "      <td>...</td>\n",
       "      <td>56.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1732279155</td>\n",
       "      <td>44</td>\n",
       "      <td>0.0</td>\n",
       "      <td>243</td>\n",
       "      <td>0.0</td>\n",
       "      <td>246</td>\n",
       "      <td>0.0</td>\n",
       "      <td>76</td>\n",
       "      <td>0.0</td>\n",
       "      <td>59</td>\n",
       "      <td>...</td>\n",
       "      <td>48.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1732279170</td>\n",
       "      <td>51</td>\n",
       "      <td>0.0</td>\n",
       "      <td>303</td>\n",
       "      <td>0.0</td>\n",
       "      <td>305</td>\n",
       "      <td>0.0</td>\n",
       "      <td>93</td>\n",
       "      <td>0.0</td>\n",
       "      <td>66</td>\n",
       "      <td>...</td>\n",
       "      <td>60.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         time  adservice_normal  adservice_stack  frontend_normal   \n",
       "0  1732279110                35              0.0              192  \\\n",
       "1  1732279125                51              0.0              262   \n",
       "2  1732279140                43              0.0              258   \n",
       "3  1732279155                44              0.0              243   \n",
       "4  1732279170                51              0.0              303   \n",
       "\n",
       "   frontend_stack  currencyservice_normal  currencyservice_stack   \n",
       "0             0.0                     223                    0.0  \\\n",
       "1             0.0                     264                    0.0   \n",
       "2             0.0                     259                    0.0   \n",
       "3             0.0                     246                    0.0   \n",
       "4             0.0                     305                    0.0   \n",
       "\n",
       "   cartservice_normal  cartservice_stack  recommendationservice_normal  ...   \n",
       "0                  55                0.0                            37  ...  \\\n",
       "1                  75                0.0                            53  ...   \n",
       "2                  79                0.0                            59  ...   \n",
       "3                  76                0.0                            59  ...   \n",
       "4                  93                0.0                            66  ...   \n",
       "\n",
       "   shippingservice_normal  shippingservice_stack  checkoutservice_normal   \n",
       "0                    34.0                    0.0                     9.0  \\\n",
       "1                    30.0                    0.0                     3.0   \n",
       "2                    56.0                    0.0                     9.0   \n",
       "3                    48.0                    0.0                     6.0   \n",
       "4                    60.0                    0.0                    12.0   \n",
       "\n",
       "   checkoutservice_stack  paymentservice_normal  paymentservice_stack   \n",
       "0                    0.0                    6.0                   0.0  \\\n",
       "1                    0.0                    2.0                   0.0   \n",
       "2                    0.0                    6.0                   0.0   \n",
       "3                    0.0                    4.0                   0.0   \n",
       "4                    0.0                    8.0                   0.0   \n",
       "\n",
       "   emailservice_normal  emailservice_stack  redis_normal  redis_stack  \n",
       "0                  3.0                 0.0           0.0          0.0  \n",
       "1                  1.0                 0.0           0.0          0.0  \n",
       "2                  3.0                 0.0           0.0          0.0  \n",
       "3                  2.0                 0.0           5.0          0.0  \n",
       "4                  4.0                 0.0           0.0          0.0  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from RCAEval.e2e import baro\n",
    "\n",
    "logs_df.index = logs_df.index.astype(int) // 1000000000\n",
    "\n",
    "# convert index to 'time' col \n",
    "time_col = logs_df.index\n",
    "logs_df = logs_df.reset_index()\n",
    "logs_df.insert(0, \"time\", time_col)\n",
    "# remove index col\n",
    "logs_df = logs_df.drop(columns=[\"index\"])\n",
    "logs_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['cartservice_stack',\n",
       " 'redis_normal',\n",
       " 'time',\n",
       " 'checkoutservice_normal',\n",
       " 'paymentservice_normal',\n",
       " 'emailservice_normal',\n",
       " 'paymentservice_stack',\n",
       " 'adservice_normal',\n",
       " 'cartservice_normal',\n",
       " 'frontend_normal',\n",
       " 'shippingservice_normal',\n",
       " 'adservice_stack',\n",
       " 'frontend_stack',\n",
       " 'currencyservice_normal',\n",
       " 'currencyservice_stack',\n",
       " 'recommendationservice_normal',\n",
       " 'recommendationservice_stack',\n",
       " 'shippingservice_stack',\n",
       " 'checkoutservice_stack',\n",
       " 'emailservice_stack',\n",
       " 'redis_stack']"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "baro(logs_df, ta)['ranks']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# show `cartservice_stack`\n",
    "logs_df[\"cartservice_stack\"].plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "info: Grpc.AspNetCore.Server.ServerCallHandler[7]\n",
      "      Error status code 'FailedPrecondition' with detail 'Can't access cart storage. System.OverflowException: Value was either too large or too small for an Int32.\n",
      "         at System.Number.ThrowOverflowException[TInteger]()\n",
      "         at System.Number.ThrowOverflowOrFormatException[TChar,TInteger](ParsingStatus, ReadOnlySpan`1)\n",
      "         at System.Number.ParseBinaryInteger[TChar,TInteger](ReadOnlySpan`1, NumberStyles, NumberFormatInfo)\n",
      "         at System.Int32.Parse(ReadOnlySpan`1 s, NumberStyles style, IFormatProvider provider)\n",
      "         at System.Int32.Parse(String s, NumberStyles style, IFormatProvider provider)\n",
      "         at System.Int32.Parse(String s)\n",
      "         at cartservice.cartstore.RedisCartStore.AddItemAsync(String, String, String) in /app/cartstore/RedisCartStore.cs:line 54' raised.\n"
     ]
    }
   ],
   "source": [
    "# config pandas to display full text\n",
    "import pandas as pd\n",
    "pd.set_option('display.max_colwidth', None)\n",
    "\n",
    "cart_logs = logs[logs['container_name'] == \"cartservice\"]\n",
    "abnormal_logs = cart_logs[cart_logs['timestamp']//1000000000 >= ta]\n",
    "stack_traces = abnormal_logs[abnormal_logs['message'].apply(lambda x: len(x.split(\"\\n\")) > 1)]\n",
    "\n",
    "# display the stack trace \n",
    "print(stack_traces['message'].iloc[10])\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Summary: \n",
    "\n",
    "**BARO can rank `cartservice_stack` as the fine-grained root cause, indicating `cartservice` is the root cause services, and the stack traces is the root cause indicator**\n",
    "\n",
    "## Future works:\n",
    "\n",
    "Using LLM to diagnose the line of code, faulty function from the stack traces could be very interesting. "
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}