#
# htmlTableParser.py
#
# Example of parsing a simple HTML table into a list of rows, and optionally into a little database
#
# Copyright 2019, Paul McGuire
#

import pyparsing as pp
import urllib.request


# define basic HTML tags, and compose into a Table
table, table_end = pp.makeHTMLTags('table')
thead, thead_end = pp.makeHTMLTags('thead')
tbody, tbody_end = pp.makeHTMLTags('tbody')
tr, tr_end = pp.makeHTMLTags('tr')
th, th_end = pp.makeHTMLTags('th')
td, td_end = pp.makeHTMLTags('td')
a, a_end = pp.makeHTMLTags('a')

# method to strip HTML tags from a string - will be used to clean up content of table cells
strip_html = (pp.anyOpenTag | pp.anyCloseTag).suppress().transformString

# expression for parsing <a href="url">text</a> links, returning a (text, url) tuple
link = pp.Group(a + a.tag_body('text') + a_end.suppress())
link.addParseAction(lambda t: (t[0].text, t[0].href))

# method to create table rows of header and data tags
def table_row(start_tag, end_tag):
    body = start_tag.tag_body
    body.addParseAction(pp.tokenMap(str.strip),
                        pp.tokenMap(strip_html))
    row = pp.Group(tr.suppress()
                   + pp.ZeroOrMore(start_tag.suppress()
                                   + body
                                   + end_tag.suppress())
                   + tr_end.suppress())
    return row

th_row = table_row(th, th_end)
td_row = table_row(td, td_end)

# define expression for overall table - may vary slightly for different pages
html_table = table + tbody + pp.Optional(th_row('headers')) + pp.ZeroOrMore(td_row)('rows') + tbody_end + table_end


# read in a web page containing an interesting HTML table
with urllib.request.urlopen("https://en.wikipedia.org/wiki/List_of_tz_database_time_zones") as page:
    page_html = page.read().decode()

tz_table = html_table.searchString(page_html)[0]

# convert rows to dicts
rows = [dict(zip(tz_table.headers, row)) for row in tz_table.rows]

# make a dict keyed by TZ database name
tz_db = {row['TZ database name']: row for row in rows}

from pprint import pprint
pprint(tz_db['America/Chicago'])