docxtpl这是一个比较冷门的 Python 库，而且库中的函数也不多，才几个，但功能却很专一。如果说 docx 库的强大在于生成我们想要的 Word 文档，那么 Docxtpl 库的存在就是将这些文档按固定格式输出。

今天在编写数据库结构文档，想到如此重复的工作，自然有系统办法实现，就找到了这个库。

docxtpl是基于python-docx和jinja2开发出来的库。他通过对docx文档模版加载，使用类似jinja2网页模版开发的语法对其进行修改。

第一步，设计文档模板，按照jinjia2的结构要求

第二步，编写获取元数据字典的脚本

import jinja2
from docxtpl import DocxTemplate
from docxtpl import InlineImage
from docx.shared import Mm
import psycopg2
import psycopg2.extras

tablesql = """
SELECT row_number() over(ORDER BY table_name) as table_no,
       table_name as table_english_name,
       obj_description (oid, 'pg_class') as table_chinese_name,
       table_catalog as table_catalog,
       table_schema as table_schema,
       pg_size_pretty (pg_relation_size (table_name)) as pg_size_pretty
  FROM information_schema.tables t1,
       pg_class t2
 WHERE table_schema = 'public'
   AND t1.table_name = t2.relname
 ORDER BY 2"""

tablecolumnsql = """
SELECT t1.relname as table_english_name,
          base.column_name column_english_name,
          col_description (t1.oid,t2.attnum) column_chinese_name,
          base.udt_name column_data_type,
          t2.attnum as column_no,
          COALESCE(character_maximum_length, numeric_precision, datetime_precision) column_length,
          (case when t2.attnotnull = true then 'Y' else 'N' end) as column_null_flag,
          (case when (select count(pg_constraint.*)
                           from pg_constraint
                          inner join pg_class on pg_constraint.conrelid = pg_class.oid
                          inner join pg_attribute on pg_attribute.attrelid = pg_class.oid and pg_attribute.attnum = any(pg_constraint.conkey)
                          inner join pg_type on pg_type.oid = pg_attribute.atttypid
                          where pg_class.relname = t1.relname
                            and pg_constraint.contype = 'p'
                            and pg_attribute.attname =t2.attname) > 0 then 'Y'
                  else 'N' end) as column_primarykey_flag                 
  FROM information_schema.COLUMNS base,
       pg_class t1,
       pg_attribute t2
 WHERE t1.relname = base."table_name"
   AND t2.attname = base."column_name"
    AND base.table_schema='public'
   AND t1.oid = t2.attrelid
   AND t2.attnum > 0
order by 1,5      
"""

第三步，构造数据库链接，获取数据

def gettablecolumninfo():
    tablelist = []
    tablecolumnlist = []
    hostname = '*.*.*.*'
    port = '5432'
    username = 'postgres'
    password = '*******'
    database = '*****db'
    connection = psycopg2.connect(host=hostname, port=port, user=username, password=password, dbname=database)
    cur = connection.cursor(cursor_factory=psycopg2.extras.DictCursor)

    cur.execute(tablesql)
    results = cur.fetchall()
    for result in results:
        tablelist.append(result)

    cur.execute(tablecolumnsql)
    results = cur.fetchall()
    for result in results:
        tablecolumnlist.append(result)
    cur.close()

    connection.close()
    return tablelist, tablecolumnlist

第四步，将数据结构转换为嵌套方式，以便前端容易实现

def trandataformat(tablelist, tablecolumnlist):
    tablecolumnlist_out = []
    for table in tablelist:
        tabledict = {}
        tabledict['table_no'] = table[0]
        tabledict['table_english_name'] = table[1]
        tabledict['table_chinese_name'] = table[2]
        columnlist = []
        for column in tablecolumnlist:
            if tabledict['table_english_name'] == column[0]:
                columndict = {}
                columndict['table_english_name'] = column[0]
                columndict['column_english_name'] = column[1]
                columndict['column_chinese_name'] = column[2]
                columndict['column_data_type'] = column[3]
                columndict['column_no'] = column[4]
                columndict['column_length'] = column[5]
                columndict['column_pk_flag'] = column[7]
                columnlist.append(columndict)
        tabledict['columns'] = columnlist
        tablecolumnlist_out.append(tabledict)
    return tablecolumnlist_out

第五步，进行转换，将模板和数据输出到文件中

tablelist, tablecolumnlist = gettablecolumninfo()
tablecolumnlist_out = trandataformat(tablelist, tablecolumnlist)

project_name = 'XXXX系统'
project_unit_name = 'XXXX中心'
project_description = 'XXXXXXXXXXXX'
db_type = 'PostgreSQL'
db_name = 'XXXdb'

templatefile = r'D:\JetBrains\PycharmProjects\solutionprj\表设计说明书模板.docx'
tpl = DocxTemplate(templatefile)
context = {"project_name": project_name,
           "project_unit_name": project_unit_name,
           "project_description": project_description,
           "db_type": db_type,
           "db_name": db_name,
           "tablecolumnlist": tablecolumnlist_out
           }
jinja_env = jinja2.Environment()
tpl.render(context, jinja_env)
tpl.save("D:\JetBrains\PycharmProjects\solutionprj\XXXX系统表结构.docx")

以下是最后生成的样例

欢迎关注公众号：python与大数据分析