Daily LeetCode

Generates bilingual (EN + ZH) LeetCode solution notes and commits them with historical dates from 2018-06-04 to 2026-05-25, producing 8 years of git history across 2913 problems.

Each problem gets its own Markdown file with:

• English original (assembled directly from JSON)
• Chinese translation (Ollama, 1 call)
• Step-by-step Python solution for beginners (Ollama, 1 call)

---

Data Source

• 题目 JSON：leetcode_problems.json，来源 neenza/leetcode-problems。仓库不内置该 JSON 文件，请大家自行下载 merged_problems.json，并按项目需要保存为 leetcode_problems.json。

Each problem JSON file contains the following fields:

• title: The name of the problem (e.g., "Container With Most Water").
• problem_id: The internal problem ID (string).
• frontend_id: The LeetCode frontend ID (string).
• difficulty: The difficulty level (Easy, Medium, or Hard).
• problem_slug: The URL-friendly name (e.g., container-with-most-water).
• topics: Array of topic tags (e.g., Array, Two Pointers).
• description: The full problem statement, usually in Markdown format.
• examples: Array of example objects, each with example_num, example_text, and images.
• constraints: Array of constraints or limits for the problem.
• follow_ups: Array of follow-up questions (if any).
• hints: Array of hints for solving the problem.
• code_snippets: Object containing starter code for various languages (e.g., python, cpp, java, etc.).
• solutions: HTML string containing editorial content for some problems.
• Some fields (like solutions, images, follow_ups) may be missing for certain problems.

---

Prerequisites

• Python 3.10+
• Ollama running locally with gpt-oss:120b pulled

ollama pull gpt-oss:120b
ollama serve

---

Installation

pip install -r requirements.txt

---

Usage

# Process the next unfinished problem (auto-resume)
python main.py

# Process all remaining problems with a progress bar
python main.py --all

# Force a specific problem by index (0-based)
python main.py --index 0

# Dry run — print info without calling Ollama or writing files
python main.py --dry-run

---

Output

solutions/
├── 2018/
│   ├── 20180604_1-two-sum.md
│   └── 20180605_2-add-two-numbers.md
├── ...
└── 2026/
    └── 20260525_3640-trionic-array-ii.md

Each commit is dated to the corresponding day with a randomised time between 08:00 and 23:30:

solve #1: Two Sum          2018-06-04T14:37:22
solve #2: Add Two Numbers  2018-06-05T09:12:55
...

---

Environment Variables

Variable	Default	Description
OLLAMA_MODEL	gpt-oss:120b	Ollama model name
OLLAMA_URL	http://10.54.79.119:11434/api/generate	Ollama API endpoint (old: localhost)
OLLAMA_TIMEOUT	300	Request timeout in seconds

---

Docker

# Build
docker build -t daily-leetcode .

# Run (Ollama must be running on the host)
docker run --add-host=host.docker.internal:host-gateway \
  -v $(pwd)/solutions:/app/solutions \
  daily-leetcode

On Linux add --add-host=host.docker.internal:host-gateway so the container can reach the host's Ollama.
On macOS / Windows host.docker.internal resolves automatically.

---

Project Structure

├── config.py               Global config (env-overridable)
├── llm.py                  Ollama API wrapper
├── prompts.py              Translation & solution prompt builders
├── committer.py            git add + commit with historical dates
├── main.py                 Entry point
├── leetcode_problems.json  Source data (2913 problems, read-only)
└── solutions/              Generated Markdown files