Insert not null information to EA generated HTML Documentation for Data model using DDL

InsertNotNull.py

# -*- coding: utf-8 -*-
import sys
import re
import glob
import os
from HTMLParser import HTMLParser

class FileIndicator:
  def __init__(self):
    self.hasNotNull = False
    self.hasTable = False
# create a subclass and override the handler methods
class MyHTMLParser(HTMLParser):
  def handle_starttag(self, tag, attrs):
    #print "Encountered a start tag:", tag
    for attr in attrs:
      if attr[0]=="class" and attr[1]=="TableRow":
        self.hasFoundColumn = True
  def handle_endtag(self, tag):
    pass
    #print "Encountered an end tag :", tag
  def handle_data(self, data):
    #print "Encountered some data  :", data
    if self.get_starttag_text()=="<title>" and len(data.strip())>0:
      self.hasFoundTitle = True
      self.title = data
    elif self.get_starttag_text()=="<strong>" and len(data.strip())>0:
      self.columnName = data

# main
DDLRootDir = "B-TableStructure"
EARootDir = "DataModel-html/EARoot/"
# ["table"]["column"] = true or false
nullable = {} 

# Iterate *.sql under DDLRootDir to find table definition
for dirpath, dirnames, filenames in os.walk(DDLRootDir):
  for filename in filenames:
    if not filename.endswith(".sql"):
      continue
    sqlFile = open(os.path.join(dirpath, filename))
    # Find [CREATE TABLE XXX] syntax in the file
    state = "FindingTable"
    for line in sqlFile:
      words = line.strip().split()
      if state in "FindingTable":
        if len(words)>2 and words[0]=="CREATE" and words[1]=="TABLE":
          tableName = words[2]
          nullable[tableName] = {}
          state = "FindingColumn"
      elif state in "FindingColumn":
        # Assumption: We assume that there will be only one column definition per line
        # The syntax ends with a semicolon or no syntax of [ColumnName type]
        if ";" in line or len(words)<2:
          state = "FindingTable"
          continue
        columnName = words[0]
        nullable[tableName][columnName] = bool(not "NOT NULL" in line)
    sqlFile.close()

parser = MyHTMLParser()
parser.hasFoundTitle = False
parser.hasFoundColumn = False
# Iterate *.htm, *.html under EARootDir to find table definition
for dirpath, dirnames, filenames in os.walk(EARootDir):
  for filename in filenames:
    if not (filename.endswith(".htm") or filename.endswith(".html")):
      continue
    filePath = os.path.join(dirpath, filename)
    htmlFile = open(filePath, "r")
    lines = htmlFile.readlines()
    htmlFile.close()
    htmlFileIndicator = FileIndicator()
    state = "FindingTitle"
    for n, line in enumerate(lines):
      parser.feed(line)
      if state in "FindingTitle":
        if parser.hasFoundTitle:
          htmlFileIndicator.hasTable = "::" in parser.title
          tableName = parser.title.replace("::", ".")
          parser.hasFoundTitle = False
          state = "FindingColumn" if htmlFileIndicator.hasTable else "Exit"
      elif state in "FindingColumn":
        if parser.hasFoundColumn:
          # Assumption: CoulmnBody would be at the next line
          state = "FindingColumnBody"
          parser.hasFoundColumn = False
      elif state in "FindingColumnBody":
        state = "FindingColumn"
        columnName = parser.columnName
        if columnName in nullable[tableName]:
          if not nullable[tableName][columnName]:
            print "%s in %s is NOT NULL" % (tableName, columnName)
            lines[n] = lines[n].rstrip() + "<i>, NOT NULL</i>\n"
            htmlFileIndicator.hasNotNull = True
      elif state in "Exit":
        break
    if htmlFileIndicator.hasNotNull:
      htmlFile = open(filePath, "w")
      htmlFile.writelines(lines)
      htmlFile.close()

raw_input("Pause...")
sys.exit(0)