import os
import pathlib
import re
from html import escape
from typing import List, Tuple

from bs4 import BeautifulSoup, NavigableString, Tag

from access import Access
from livecarta_config import LawCartaConfig


def save_image_locally(img_file_path, img_content, book_id):
    folder_path = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
    new_path = pathlib.Path(os.path.join(folder_path, f'json/img_{book_id}/'))
    new_path.mkdir(exist_ok=True)

    new_img_path = new_path / os.path.basename(img_file_path)
    f = open(new_img_path, 'wb+')
    f.write(img_content)
    f.close()

    return new_img_path


def save_image_to_aws(access: Access, img_file_path, img_content: bytes, book_id):
    link = access.send_image_by_bytes(img_file_path, img_content, book_id)
    return link


def update_src_links_in_images(body_tag: Tag,
                               href2img_content: dict,
                               path_to_html,
                               access=None,
                               path2aws_path=None):
    img_tags = body_tag.find_all('img')

    for img in img_tags:
        path_to_img_from_html = img.attrs.get('src')
        html_folder = os.path.dirname(path_to_html)
        path_to_img_from_root = os.path.normpath(os.path.join(html_folder, path_to_img_from_html))

        assert path_to_img_from_root in href2img_content, \
            f'Image {path_to_img_from_html} in file {path_to_html} was not added to manifest.'

        img_content = href2img_content[path_to_img_from_root]
        if access is not None:
            if path_to_img_from_root in path2aws_path:
                new_folder = path2aws_path[path_to_img_from_root]
            else:
                new_folder = save_image_to_aws(access, path_to_img_from_root, img_content, 'book_id')
                path2aws_path[path_to_img_from_root] = new_folder
        else:
            new_folder = save_image_locally(path_to_img_from_root, img_content, 'book_id')

        img.attrs['src'] = str(new_folder)
        if img.attrs.get('width'):
            del img.attrs['width']
        if img.attrs.get('height'):
            del img.attrs['height']

    return path2aws_path


def preprocess_figure():
    pass


def preprocess_table(body_tag: BeautifulSoup):
    tables = body_tag.find_all("table")
    for table in tables:
        tds = table.find_all(re.compile("td|th|tr"))

        border_sizes = []
        for td in tds:
            style = td.get('style')
            width = ''
            if style:
                border_match = re.search(r"border: ?(\d+\.?\d*)(p[tx])", style) or\
                               re.search(r"border-top-width: ?(\d+\.?\d*)(p[tx])", style) or\
                               re.search(r"border-left-width: ?(\d+\.?\d*)(p[tx])", style) or \
                               re.search(r"border-right-width: ?(\d+\.?\d*)(p[tx])", style) or \
                               re.search(r"border-bottom-width: ?(\d+\.?\d*)(p[tx])", style)

                if border_match:
                    size = border_match.group(1)
                    units = border_match.group(2)
                    border_sizes.append(float(size))

                width_match = re.search(r"[^-]width: ?(\d+\.?\d*)(p[tx])", style)

                if width_match:
                    size = width_match.group(1)
                    units = width_match.group(2)
                    width = size+'px'

            width = td.get('width') or width

            if width:
                td.attrs['width'] = width

            if td.attrs.get('style'):
                td.attrs['style'] = td.attrs['style'].replace('border:0;', '')

            if td.attrs.get('style') == '':
                del td.attrs['style']

        if border_sizes:
            border_size = sum(border_sizes) / len(border_sizes)
            table.attrs['border'] = f'{border_size:.2}'
        else:
            table.attrs['border'] = '1'


def _process_lists(body_tag):
    """
    Function to process tags <li>.
    Unwrap <p> tags.
    """
    li_tags = body_tag.find_all("li")

    for il_tag in li_tags:
        if il_tag.p:
            il_tag.attrs.update(il_tag.p.attrs)
            il_tag.p.unwrap()


def insert_span_with_attrs_before_tag(main_tag, tag, id_, class_):
    new_tag = main_tag.new_tag("span")
    new_tag.attrs['id'] = id_ or ''
    new_tag.attrs['class'] = class_ or ''
    new_tag.string = "\xa0"
    tag.insert_before(new_tag)


def clean_headings_content(content: Tag, title: str):
    def _add_span_to_save_ids_for_links(tag_to_be_removed, body_tag):
        if tag_to_be_removed.attrs.get('id'):
            insert_span_with_attrs_before_tag(body_tag,
                                              tag_to_be_removed,
                                              id_=tag_to_be_removed.attrs.get('id'),
                                              class_=tag_to_be_removed.attrs.get('class'))

        for sub_tag in tag_to_be_removed.find_all():
            if sub_tag.attrs.get('id'):
                insert_span_with_attrs_before_tag(body_tag,
                                                  tag_to_be_removed,
                                                  id_=sub_tag.attrs['id'],
                                                  class_=sub_tag.attrs.get('class'))

    title = title.lower()
    for child in content.contents:
        if isinstance(child, NavigableString):
            text = child
        else:
            text = child.text
        if text and re.sub(r'([\n\t\xa0])', '', text):
            text = re.sub(r'([\n\t\xa0])', ' ', text)
            text = re.sub(r' +', ' ', text).strip()
            text = text.lower()
            if title == text:
                _add_span_to_save_ids_for_links(child, content)
                child.extract()
            elif (title in text) and (child.name in ['h1', 'h2', 'h3']):
                _add_span_to_save_ids_for_links(child, content)
                child.extract()
            break


def _preprocessing_headings(body_tag):
    """
    Function to convert all lower level headings to p tags
    """
    pattern = f'^h[{LawCartaConfig.SUPPORTED_LEVELS + 1}-9]$'
    header_tags = body_tag.find_all(re.compile(pattern))
    for tag in header_tags:
        tag.name = 'p'


def clean_title_from_numbering(title: str):
    """
    Function to remove digits  from headers.
    """
    title = re.sub(r'^(\s+)+', '', title)
    title = re.sub(r'^(?:\.?\d+\.? ?)+', '', title)
    # title = re.sub(r'^(?:\.?[MDCLXVIclxvi]+\.? ?)+ ', '', title)  # delete chapter numbering from the title
    title = re.sub(r'^(?:[A-Za-z]\. ?)+', '', title)
    return title


def replace_with_livecarta_anchor_tag(anchor, i):
    new_tag = BeautifulSoup(features='lxml').new_tag('sup')
    new_tag['class'] = 'footnote-element'
    new_tag['data-id'] = i + 1
    new_tag['id'] = f'footnote-{i + 1}'
    new_tag.string = '*'
    anchor.replace_with(new_tag)


def preprocess_footnotes(source_html_tag: Tag, href2soup_html: dict = None, noteref_attr_name='epub:type') -> List[str]:
    """
    This function should be earlier that adding fonts in pipeline.

    <p>Here is an example footnote<sup><a epub:type="noteref" href="#n1">1</a></sup></p>
    <aside epub:type="footnote" id="n1"><p>With a footnote here.</p></aside>

    """
    footnotes = []
    noterefs_tags = source_html_tag.find_all(attrs={noteref_attr_name: 'noteref'})
    bad_noterefs_tags = set([tag for tag in noterefs_tags if not tag.attrs.get('href')])
    noterefs_tags = [tag for tag in noterefs_tags if tag not in bad_noterefs_tags]
    [tag.decompose() for tag in bad_noterefs_tags]

    def parse_a_tag_href(s: str):
        assert '#' in s, f'Error. Unexpected href: {s} in a tag. Href must contain an id.'
        f, id_ = s.split('#')
        return f, id_

    def verify_footnote_tag(tags: list):
        assert len(tags) <= 1, f'Error, Multiple id: {href}.\n{tags}'
        if len(tags) == 0:
            anchored_tags = list(target_html_tag.find_all(id=element_id))
            if len(anchored_tags):
                print(f'Warning. Href for tag is detected as footnote:\n{noteref_tag}')
                return anchored_tags
            else:
                assert 0, f'Error, No element with id: {href} found.'

        return tags

    def get_footnote_tags2str(t):
        unicode_string = ''
        for child in t.children:
            if type(child) is NavigableString:
                unicode_string += str(child)
            else:
                unicode_string += child.decode_contents()

        return unicode_string.strip()

    for i, noteref_tag in enumerate(noterefs_tags):
        href = noteref_tag.attrs['href']
        file, element_id = parse_a_tag_href(href)
        if not file:
            target_html_tag = source_html_tag
        else:
            target_html_tag = href2soup_html.get(file)
            if not target_html_tag:
                print(f'Error while footnotes processing. For {noteref_tag} invalid path: {file}.')
                continue

        possible_footnote = 'note|footnote|endnote|rearenote'
        expected_footnote_tags = list(target_html_tag.find_all(id=element_id,
                                                               attrs={'epub:type': re.compile(possible_footnote)}))

        expected_footnote_tags = verify_footnote_tag(expected_footnote_tags)
        footnote_tag = expected_footnote_tags[0]
        replace_with_livecarta_anchor_tag(noteref_tag, i)
        content = get_footnote_tags2str(footnote_tag)

        footnote_tag.decompose()
        footnotes.append(content)

    return footnotes


def unwrap_structural_tags(body_tag):

    def _add_span_to_save_ids_for_links(tag_to_be_removed):
        if tag_to_be_removed.attrs.get('id'):
            insert_span_with_attrs_before_tag(main_tag=body_tag, tag=tag_to_be_removed,
                                              id_=tag_to_be_removed.attrs['id'],
                                              class_=tag_to_be_removed.attrs.get('class'))

    structural_tags_names = [
        'div', 'section', 'article', 'main', 'body', 'html', 'aside', 'canvas', 'data',
        'figure', 'footer', 'iframe', 'span', 'p'
    ]
    # should be before other tags processing, not to remove converter empty tags with id
    # for s in body_tag.find_all("span"):
    #     if (s.attrs.get('epub:type') == 'pagebreak') or s.attrs.get('id'):
    #         continue
    #     if s.contents:
    #         is_not_struct_tag = [child.name not in structural_tags_names for child in s.contents]
    #         if all(is_not_struct_tag):
    #             continue
    #
    #     _add_span_to_save_ids_for_links(s)
    #     s.unwrap()

    for div in body_tag.find_all("div"):
        if div.contents:
            is_not_struct_tag = [child.name not in structural_tags_names for child in div.contents]
            if all(is_not_struct_tag):
                div.name = 'p'
                continue

        _add_span_to_save_ids_for_links(div)
        div.unwrap()

    for s in body_tag.find_all("section"):
        if s.attrs.get('class'):
            class_ = s.attrs['class'] if not isinstance(s.attrs['class'], list) else s.attrs['class'][0]
            if s.parent.name == 'aside':
                if not s.parent.attrs.get('class'):
                    s.parent.attrs['class'] = class_
        _add_span_to_save_ids_for_links(s)
        s.unwrap()

    for s in body_tag.find_all("article"):
        _add_span_to_save_ids_for_links(s)
        s.unwrap()

    for s in body_tag.find_all("figure"):
        s.name = 'p'
        s.attrs['style'] = "text-align: center;"

    for s in body_tag.find_all("figcaption"):
        _add_span_to_save_ids_for_links(s)
        s.unwrap()

    for s in body_tag.find_all("aside"):
        s.name = 'blockquote'

    for s in body_tag.find_all("main"):
        _add_span_to_save_ids_for_links(s)
        s.unwrap()

    for s in body_tag.find_all("body"):
        _add_span_to_save_ids_for_links(s)
        s.unwrap()

    for s in body_tag.find_all("html"):
        _add_span_to_save_ids_for_links(s)
        s.unwrap()

    for s in body_tag.find_all("header"):
        s.name = 'span'

    # check marks for chapter starting are on the same 1 level
    marks = body_tag.find_all(attrs={'class': 'converter-chapter-mark'})
    parents_marks_are_body = [x.parent == body_tag for x in marks]

    # fix marks to be on 1 level
    if not all(parents_marks_are_body):
        for x in marks:
            while x.parent != body_tag:
                x.parent.unwrap()  # todo warning! could reflect on formatting/internal links in some cases

    parents_marks_are_body = [x.parent == body_tag for x in marks]
    assert all(parents_marks_are_body), 'Anchor for chapter is deeper than 2 level. Chapters can not be parsed.'

    _preprocessing_headings(body_tag)

    for node in body_tag:
        if isinstance(node, NavigableString):
            content = str(node)
            content = re.sub(r'([\n\t\xa0])', ' ', content)
            content = content.strip()
            if content:
                tag = body_tag.new_tag('p')
                tag.append(str(node))
                node.replace_with(tag)

    return body_tag


def get_tags_between_chapter_marks(first_id, href, html_soup):
    marked_tags = html_soup.find(attrs={'id': first_id, 'class': 'converter-chapter-mark'})
    if marked_tags:
        next_tag = marked_tags.next_sibling
        tags = []
        while next_tag:
            # TODO: why we hve there NavString

            if not isinstance(next_tag, NavigableString) and\
                    (next_tag.attrs.get('class') == 'converter-chapter-mark'):
                break
            tags.append(next_tag)
            next_tag = next_tag.next_sibling

        tags = [tag.extract() for tag in tags]
        html_soup.smooth()

    else:
        assert 0, f'Warning: no match for {first_id, href}'

    return tags


def wrap_span_with_table(main_tag, old_tag):
    table = main_tag.new_tag("table")
    table.attrs['border'] = '1px solid #ccc;'
    table.attrs['style'] = 'width:100%;'
    tbody = main_tag.new_tag("tbody")
    tr = main_tag.new_tag("tr")
    td = main_tag.new_tag("td")
    td.attrs['bgcolor'] = '#f5f5f5'
    td.attrs['border-radius'] = '4px'
    old_tag.wrap(td)
    td.wrap(tr)
    tr.wrap(tbody)
    tbody.wrap(table)
    return table


def wrap_block_with_table(main_tag, old_tag, color=None):
    table = main_tag.new_tag("table")
    table.attrs['border'] = '1px solid'
    table.attrs['align'] = 'center'
    table.attrs['style'] = 'width:95%;'
    tbody = main_tag.new_tag("tbody")
    tr = main_tag.new_tag("tr")
    td = main_tag.new_tag("td")
    td.attrs['border-radius'] = '8px'
    if color:
        td.attrs['bgcolor'] = color
    old_tag.wrap(td)
    td.wrap(tr)
    tr.wrap(tbody)
    tbody.wrap(table)
    table.insert_after(BeautifulSoup(features='lxml').new_tag("br"))
    return table


def preprocess_block_tags(chapter_tag):
    for block in chapter_tag.find_all("blockquote"):
        if block.attrs.get('class') in ['feature1', 'feature2', 'feature3', 'feature4']:
            hrs = block.find_all("p", attrs={"class": re.compile(".+ hr")})
            for hr in hrs:
                hr.extract()

            h = block.find(re.compile("h[1-9]"))
            if h:
                h.name = "p"
                h.insert_before(BeautifulSoup(features='lxml').new_tag("br"))

            color = '#DDDDDD' if block.attrs.get('class') == 'feature1' else None
            color = '#EEEEEE' if block.attrs.get('class') == 'feature2' else color
            wrap_block_with_table(chapter_tag, block, color)
            block.insert_after(BeautifulSoup(features='lxml').new_tag("br"))
            block.unwrap()


def preprocess_pre_tags(chapter_tag):
    for pre in chapter_tag.find_all("pre"):
        new_tag = BeautifulSoup(features='lxml').new_tag("span")
        new_tag.attrs = pre.attrs.copy()

        for child in pre.children:
            if isinstance(child, NavigableString):
                text = pre.text
                text = text.replace('\t', "\xa0 \xa0 \xa0 ")
                text = text.replace(' ', "\xa0 ")
                elements = re.split('\r\n|\n|\r', text)
                for i in elements:
                    new_tag.append(NavigableString(i))
                    new_tag.append(BeautifulSoup(features='lxml').new_tag('br'))
            else:
                new_tag.append(child.extract())

        new_tag.attrs['style'] = "font-family: courier new,courier,monospace; " \
                                 "font-size: 14px; white-space: pre-wrap;"
        pre.insert_before(new_tag)
        pre.extract()
        table = wrap_span_with_table(chapter_tag, new_tag)
        p_for_br = chapter_tag.new_tag("p")
        p_for_br.string = "\xa0"
        table.insert_after(p_for_br)


def preprocess_code_tags(chapter_tag):
    for code in chapter_tag.find_all(re.compile("code|kdb|var")):
        code.name = 'span'
        code.attrs['style'] = 'color:#c7254e; font-size: 14px; font-family: courier new,courier,monospace;'


def prepare_title_and_content(title, chapter_tag: BeautifulSoup, remove_title_from_chapter) -> Tuple[str, str]:
    title_str = BeautifulSoup(title, features='lxml').string
    title_str = re.sub(r'([\n\t\xa0])', ' ', title_str)
    title_str = re.sub(r' +', ' ', title_str).rstrip()
    # 0. cleaning \n
    to_remove = []
    for child in chapter_tag.contents:
        if isinstance(child, NavigableString):
            s = re.sub(r'([\n\t])', '', child.string)
            if s == '':
                to_remove.append(child)

    [x.extract() for x in to_remove]
    # 1. heading removal
    if remove_title_from_chapter:
        clean_headings_content(chapter_tag, title_str)
    _process_lists(chapter_tag)
    preprocess_table(chapter_tag)
    preprocess_code_tags(chapter_tag)
    preprocess_pre_tags(chapter_tag)
    preprocess_block_tags(chapter_tag)
    # 2. class removal
    for tag in chapter_tag.find_all(recursive=True):
        if hasattr(tag, 'attrs') and tag.attrs.get('class') and (tag.attrs.get('class') not in ['link-anchor']):
            del tag.attrs['class']

    # content_str = re.sub(r'([\n\t\xa0])', ' ', str(content_tag))
    title_str = clean_title_from_numbering(title_str)
    return title_str, str(chapter_tag)