根据 csv 生成 SQLAlchemy models 代码

generate-sqalchemy-model-code.py

#!/usr/bin/env python3
# Author    : weaming
# Mail      : garden.yuen@gmail.com
# Created   : 2020-07-19 21:01:50

# Usage:
# csv_json: pip3 install pretty-format-json
# xsv: brew install xsv
# cat esf-nn-rst.csv | xsv stats | csv_json | generate-sqalchemy-model-code.py

import sys
import json


def read_stdin():
    return json.load(sys.stdin)


def try_load_as_json(v):
    if not (str(v).startswith('[') or str(v).startswith('{')):
        return v
    try:
        return json.loads(v)
    except Exception:
        return v


def parse_stat(stat):
    """
    {
       "field": "bieming",
       "type": "Unicode",
       "sum": "",
       "min": "别名：21时代公寓",
       "max": "别名：龙胜小区",
       "min_length": "0",
       "max_length": "141",
       "mean": "",
       "stddev": ""
    }
    """
    stat['min_length'] = int(stat['min_length'])
    stat['max_length'] = int(stat['max_length'])
    stat['min'] = try_load_as_json(stat['min'])
    stat['max'] = try_load_as_json(stat['max'])
    return stat


def gen_field(stat):
    tt = type(stat['max'])
    t = {
        int: 'db.Integer',
        float: 'db.Float',
        dict: 'JSONField',
        list: 'JSONField',
        str: 'db.String({})',
    }[tt]
    if tt is str:
        l = max(stat['max_length'], 64)
        while l % 64 != 0:
            l += 1
        if l >= 1024:
            t = 'db.Text'
        else:
            t = t.format(l)
    nullable = stat['min_length'] == 0 if tt not in [dict, list] else not stat['min']
    return f"{stat['field']} = db.Column({t}, nullable={nullable}, comment='{str(stat['field'])}')  # {stat['field']}"


def main():
    stats = read_stdin()
    for stat in stats:
        stat = parse_stat(stat)
        print(gen_field(stat))


main()