from html.parser import HTMLParser

STATE_NONE = 0
STATE_PRE  = 1
STATE_CODE = 2

class CodeParser(HTMLParser):
    def __init__(self, **kwargs):
        super().__init__()
    
        # This shall be a list of dictionnaries, which 
        # look like this: 
        # [
        #   {'lang': 'ly', 'content': '...'},
        #   {'lang': 'ly', 'content': '...'},
        #   {'lang': 'tex', 'content': '...'},
        #   ...
        # ]
        self.blocks = []

        # Yes, this thing acts kinda like a state machine.
        self.state = STATE_NONE
        self.language = ''
        self.data = ''

    def handle_starttag(self, tag, attrs):
        if self.state == STATE_NONE and tag == "pre":
            self.state = STATE_PRE
        elif self.state == STATE_PRE and tag == "code":
            self.state = STATE_CODE
            for val in attrs:
                v = val[1]
                if val[0] == "class":
                    if v == "language-lilypond":
                        self.language = 'ly'
                    if v == "language-tex" or v == "language-latex":
                        self.language = 'tex'

            

    def handle_endtag(self, tag):
        if self.state == STATE_CODE and tag == "pre":
            self.blocks.append({'lang': self.language, 'content': self.data})

            self.state = STATE_NONE
            self.data = ''
            self.language = ''
            
    def handle_data(self, data):
        if self.state == STATE_CODE:
            self.data += data