import urllib.request
import os
import re
import json
import jinja2

STATIC_FOLDER_PATH = '.' # without trailing slash
PUBLIC_STATIC_FOLDER_PATH = '.' # without trailing slash
TEMPLATES_DIR = './templates'

# This uses a low quality copy of all the images 
# (using a folder with the name "images-small",
# which stores a copy of all the images generated with:
# $ mogrify -quality 5% -adaptive-resize 25% -remap pattern:gray50 * )
fast = False

def API_request(url, pagename):
	"""
		url = API request url (string)
		data =  { 'query': 
					'pages' : 
						pageid : { 					
							'links' : {
								'?' : '?'
								'title' : 'pagename'
							}
						} 
					}  
				}
	"""
	response = urllib.request.urlopen(url).read()
	data = json.loads(response)

	# Save response as JSON to be able to inspect API call
	json_file = f'{ STATIC_FOLDER_PATH }/{ pagename }.json'
	print('Saving JSON:', json_file)
	with open(json_file, 'w') as out:
		out.write(json.dumps(data, indent=4))
		out.close()

	return data

def download_media(html, images, wiki):
	"""
		html = string (HTML)
		images = list of filenames (str)
	"""
	# check if 'images/' already exists
	if not os.path.exists(f'{ STATIC_FOLDER_PATH }/images'):
		os.makedirs(f'{ STATIC_FOLDER_PATH }/images')

	# download media files
	for filename in images:
		filename = filename.replace(' ', '_') # safe filenames

		# check if the image is already downloaded
		# if not, then download the file
		if not os.path.isfile(f'{ STATIC_FOLDER_PATH }/images/{ filename }'):

			# first we search for the full filename of the image
			url = f'{ wiki }/api.php?action=query&list=allimages&aifrom={ filename }&format=json'
			response = urllib.request.urlopen(url).read()
			data = json.loads(response)

			# we select the first search result
			# (assuming that this is the image we are looking for)
			image = data['query']['allimages'][0] 

			# then we download the image
			image_url = image['url']
			image_filename = image['name']
			print('Downloading:', image_filename)
			image_response = urllib.request.urlopen(image_url).read()

			# and we save it as a file
			image_path = f'{ STATIC_FOLDER_PATH }/images/{ image_filename }'
			out = open(image_path, 'wb') 
			out.write(image_response)
			out.close()

			import time
			time.sleep(3) # do not overload the server

		# replace src link
		image_path = f'{ PUBLIC_STATIC_FOLDER_PATH }/images/{ filename }' # here the images need to link to the / of the domain, for flask :/// confusing! this breaks the whole idea to still be able to make a local copy of the file
		matches = re.findall(rf'src="/images/.*?px-{ filename }"', html) # for debugging
		if matches:
			html = re.sub(rf'src="/images/.*?px-{ filename }"', f'src="{ image_path }"', html)
		else:
			matches = re.findall(rf'src="/images/.*?{ filename }"', html) # for debugging
			html = re.sub(rf'src="/images/.*?{ filename }"', f'src="{ image_path }"', html) 
		# print(f'{filename}: {matches}\n------') # for debugging: each image should have the correct match!

	return html

def add_item_inventory_links(html):
	"""
		html = string (HTML)
	"""
	# Find all references in the text to the item index
	pattern = r'Item \d\d\d'
	matches = re.findall(pattern, html)
	index = {}
	new_html = ''
	from nltk.tokenize import sent_tokenize
	for line in sent_tokenize(html):
		for match in matches:
			if match in line:
				number = match.replace('Item ', '').strip()
				if not number in index:
					index[number] = []
					count = 1
				else:
					count = index[number][-1] + 1
				index[number].append(count)
				item_id = f'ii-{ number }-{ index[number][-1] }'
				line = line.replace(match, f'Item <a id="{ item_id }" href="#Item_Index">{ number }</a>')

		# the line is pushed back to the new_html
		new_html += line + ' '
		
	# Also add a <span> around the index nr to style it
	matches = re.findall(r'<li>\d\d\d', new_html)
	for match in matches:
		new_html = new_html.replace(match, f'<li><span class="item_nr">{ match }</span>')

	# import json
	# print(json.dumps(index, indent=4))
	
	return new_html

def tweaking(html):
	"""
		html = string (HTML)
	"""
	html = html.replace('<a href="#X,_y,_z_(4_filmstills)"', '<a href="#x,_y,_z_(4_filmstills)"') # change the anchor link in the TOC to lowercase
	html = html.replace('<a href="#Rehearsal_as_the_%E2%80%98Other%E2%80%99_to_Hypercomputation"', '<a href="#Rehearsal_as_the_‘Other’_to_Hypercomputation"') # change the anchor link in the TOC to lowercase
	html = html.replace('<a href="#We_hardly_encounter_anything_that_didn%E2%80%99t_really_matter"', '<a href="#We_hardly_encounter_anything_that_didn’t_really_matter"') # change the anchor link in the TOC to lowercase
	html = re.sub(r'''<h3><span class="mw-headline" id="References.*?">References</span><span class="mw-editsection"><span class="mw-editsection-bracket"></span></span></h3>
<ul>''', '''<h3 class="references"><span class="mw-headline" id="References">References</span><span class="mw-editsection"><span class="mw-editsection-bracket"></span></span></h3>
<ul class="references">''', html) # add id="references" to h3 and ul, so the elements can be selected with CSS
	html = html.replace('src="./images/Userinfo.jpg"', 'src="./images/Userinfo.svg"')
	
	return html

def clean_up(html):
	"""
		html = string (HTML)
	"""
	html = re.sub(r'\[.*edit.*\]', '', html) # remove the [edit]
	html = re.sub(r'href="/index.php\?title=', 'href="#', html) # remove the internal wiki links
	html = re.sub(r'&#91;(?=\d)', '', html) # remove left footnote bracket [
	html = re.sub(r'(?<=\d)&#93;', '', html) # remove right footnote bracket ]
	return html

def fast_loader(html):
	"""
		html = string (HTML)
	"""
	if fast == True:
		html = html.replace('/images/', '/images-small/')
		print('--- rendered in FAST mode ---')

	return html

def parse_page(pagename, wiki):
	"""
		pagename = string
		html = string (HTML)
	"""
	parse = f'{ wiki }/api.php?action=parse&page={ pagename }&pst=True&format=json'
	data = API_request(parse, pagename)
	# print(json.dumps(data, indent=4))
	if 'parse' in data:
		html = data['parse']['text']['*']
		images = data['parse']['images']
		html = download_media(html, images, wiki)
		html = clean_up(html)
		html = add_item_inventory_links(html)
		html = tweaking(html)
		html = fast_loader(html)
	else: 
		html = None

	return html

def save(html, pagename):
	"""
		html = string (HTML)
		pagename = string
	"""
	if __name__ == "__main__":
		# command-line

		# save final page that will be used with PagedJS
		template_file = open(f'{ STATIC_FOLDER_PATH }/{ TEMPLATES_DIR }/template.html').read()
		template = jinja2.Template(template_file)
		html = template.render(publication_unfolded=html, title=pagename)
		
		html_file = f'{ STATIC_FOLDER_PATH }/{ pagename }.html'
		print('Saving HTML:', html_file)
		with open(html_file, 'w') as out:
			out.write(html)
			out.close()

		# save extra html page for debugging (CLI only)
		template_file = open(f'{ STATIC_FOLDER_PATH }/{ TEMPLATES_DIR }/template.inspect.html').read()
		template = jinja2.Template(template_file)
		html = template.render(publication_unfolded=html, title=pagename)

		html_file = f'{ STATIC_FOLDER_PATH }/{ pagename }.inspect.html'
		print('Saving HTML:', html_file)
		with open(html_file, 'w') as out:
			out.write(html)
			out.close()

	else:
		# Flask application 

		with open(f'{ STATIC_FOLDER_PATH }/Unfolded.html', 'w') as out:
			out.write(html) # save the html to a file (without <head>)

def update_material_now(pagename, wiki):
	"""
		pagename = string
		publication_unfolded = string (HTML)
	"""
	publication_unfolded = parse_page(pagename, wiki)

	return publication_unfolded

# ---

if __name__ == "__main__":

	wiki = 'https://volumetricregimes.xyz' # remove tail slash '/'
	pagename = 'Unfolded'
	
	publication_unfolded = update_material_now(pagename, wiki) # download the latest version of the page
	save(publication_unfolded, pagename) # save the page to file