Hosuke

通則

凡例

• 每 session 之初，先以 /model 設心力於極
• 繁難之事，善用 ultrathink 與 ultraplan
• 勿充百萬之境——至二三分滿則當約（compact）或交接（handoff）
• 時察 /insights，據之以更新 CLAUDE.md

校讎之則（必遵於諸項）

凡將 git commit，必循此序：

Hosuke

import json
import requests
import pandas as pd

def get_coinpaprika_data():
    # Get data from coinpaprika
    url = "https://api.coinpaprika.com/v1/coins/"
    response = requests.get(url)
    return response.json()

Hosuke

import pandas as pd

if __name__ == '__main__':

    # Reload the CSV file due to code execution state reset
    file_path = '01HKPMHCEM0TFYD6KNPM66XHTT.csv'
    df = pd.read_csv(file_path)

    # Construct the SQL query in the desired format with dynamic field names
    sql_query = "SELECT "

Hosuke

import pandas as pd
import yaml

# 读取CSV文件
df = pd.read_csv('dex_base_trades_seed.csv')

# 初始化schema字典
schema = {
    "version": 2,
    "seeds": []

Hosuke

import os
import re

# Get the current working directory
directory = os.getcwd()

# Iterate over all files in the directory and its subdirectories
for dirpath, dirnames, filenames in os.walk(directory):
    for filename in filenames:
        # Process only files with a .sql extension

Hosuke

import os
import re

# 获取当前工作目录
directory = os.getcwd()

# 定义日期的正则表达式
date_pattern = r"'((19|20)\d\d[-](0[1-9]|1[012])[-](0[1-9]|[12][0-9]|3[01]))'"

# 遍历目录及其所有子目录下的所有文件

Hosuke

import os
import re

# 获取当前工作目录
directory = os.getcwd()

# 遍历目录及其所有子目录下的所有文件
for dirpath, dirnames, filenames in os.walk(directory):
    for filename in filenames:
        # 只处理.sql文件

Hosuke

import os
import re

# 获取当前工作目录
directory = os.getcwd()

# 遍历目录及其所有子目录下的所有文件
for dirpath, dirnames, filenames in os.walk(directory):
    for filename in filenames:
        # 只处理.sql文件

Hosuke

import json

def get_model_dependencies(manifest_file, model_name):
    with open(manifest_file, 'r') as f:
        manifest = json.load(f)

    # Build the adjacency list
    adj_list = {}
    for full_model_name, model in manifest['nodes'].items():
        if not full_model_name.startswith('model.'):

Hosuke

import json

def get_model_dependencies(manifest_file, model_name):
    with open(manifest_file, 'r') as f:
        manifest = json.load(f)

    def get_dependencies(full_model_name, indent=0):
        if full_model_name.startswith('model.'):
            model = manifest['nodes'].get(full_model_name)
            if model is None: