feat: [experimental] add difficulty and solutions

Author: hxu296

data-gen/.ipynb_checkpoints/leetcode_scraper-checkpoint.ipynb

@@ -3,7 +3,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "90903826",
+   "id": "abeb0073",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -18,7 +18,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "8acd2701",
+   "id": "c2b6ff54",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -32,7 +32,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "831d782e",
+   "id": "a9bacba8",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -118,10 +118,11 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "cc9ecf71",
+   "id": "8eedcca9",
    "metadata": {},
    "outputs": [],
    "source": [
+    "# note: this cell takes around 1 hour to finish\n",
     "# go through questions in the leetcode question list\n",
     "leetcode_problems = pd.read_csv('../data/leetcode_problems.csv')\n",
     "problems_companies_info = get_problems_companies_info(leetcode_problems)\n",
@@ -131,7 +132,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "d570a4aa",
+   "id": "30c80380",
    "metadata": {},
    "outputs": [],
    "source": []

data-gen/leetcode_scraper.ipynb

@@ -3,7 +3,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "90903826",
+   "id": "abeb0073",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -18,7 +18,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "8acd2701",
+   "id": "c2b6ff54",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -32,7 +32,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "831d782e",
+   "id": "a9bacba8",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -118,10 +118,11 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "cc9ecf71",
+   "id": "8eedcca9",
    "metadata": {},
    "outputs": [],
    "source": [
+    "# note: this cell takes around 1 hour to finish\n",
     "# go through questions in the leetcode question list\n",
     "leetcode_problems = pd.read_csv('../data/leetcode_problems.csv')\n",
     "problems_companies_info = get_problems_companies_info(leetcode_problems)\n",
@@ -131,7 +132,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "d570a4aa",
+   "id": "30c80380",
    "metadata": {},
    "outputs": [],
    "source": []

data-gen/parse_readme.py

@@ -4,15 +4,16 @@
 
 parsed_list = []
 df = md_to_df('../data/leetcode_repo_readme.md')
-
 for _, row in df.iterrows():
     name_and_link = row['Title']
+    difficulty = row['Difficulty']
+    solution = row['Solutions'].replace('..', 'https://github.com/fishercoder1534/Leetcode/blob')
     m = re.search(r'\[(.*)\]\((.*)\)', name_and_link)
     name = m[1]
     link = m[2]
-    parsed_list.append((name, link, ))
+    parsed_list.append((name, link, difficulty, solution, ))
 
-parsed_df = pd.DataFrame(parsed_list, columns=['name', 'link'])
+parsed_df = pd.DataFrame(parsed_list, columns=['name', 'link', 'difficulty', 'solution'])
 parsed_df.to_csv('../data/leetcode_problems.csv', index=False)
 
 

data-gen/transform_data.ipynb

@@ -2,7 +2,7 @@
  "cells": [
   {
    "cell_type": "markdown",
-   "id": "a801754c",
+   "id": "75f28a50",
    "metadata": {},
    "source": [
     "## Generate pivot wider CSV"
@@ -11,7 +11,7 @@
   {
    "cell_type": "code",
    "execution_count": 2,
-   "id": "742dffb8",
+   "id": "5332f87a",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -21,7 +21,7 @@
   {
    "cell_type": "code",
    "execution_count": 53,
-   "id": "ea4a7917",
+   "id": "b4ae5a29",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -33,7 +33,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "9ba1b2b6",
+   "id": "72e54b4c",
    "metadata": {},
    "source": [
     "## Generate CSV for each company"
@@ -42,7 +42,7 @@
   {
    "cell_type": "code",
    "execution_count": 62,
-   "id": "d08c2d89",
+   "id": "daed41c6",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -57,7 +57,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "2a32435c",
+   "id": "9649af13",
    "metadata": {},
    "outputs": [],
    "source": []