IDMB trailer lookup

imdb_video_list.py

import json
import requests
import time

def get_imdb_videos(imdb_id):
	import re, requests
	API_URL = "https://graphql.prod.api.imdb.a2z.com/"
	HEADERS = {
		'Referer': 'https://www.imdb.com/',
		'Origin': 'https://www.imdb.com',
		'User-Agent': 'Mozilla/5.0'
	}

	def gqlmin(q):
		return re.sub(' {4}', '', q)

	query_subpage = '''
	query TitleVideoGallerySubPage(
		$const: ID!,
		$first: Int!,
		$filter: VideosQueryFilter,
		$sort: VideoSort
	) {
		title(id: $const) {
			titleText { text }
			plot { plotText { plainText } }
			videoStrip(first: $first, filter: $filter, sort: $sort) {
				...VideoGalleryItems
			}
		}
	}
	'''
	query_pagination = '''
	query TitleVideoGalleryPagination(
		$const: ID!,
		$first: Int!,
		$after: ID!,
		$filter: VideosQueryFilter,
		$sort: VideoSort
	) {
		title(id: $const) {
			videoStrip(first: $first, after: $after, filter: $filter, sort: $sort) {
				...VideoGalleryItems
			}
		}
	}
	'''
	fragment = '''
	fragment VideoGalleryItems on VideoConnection {
		pageInfo {
			endCursor
			hasNextPage
		}
		total
		edges {
			node {
				id
				contentType { id }
				name { value }
				runtime { value }
				thumbnail { url }
				primaryTitle {
					series {
						displayableEpisodeNumber {
							displayableSeason {
								season
							}
						}
						series {
							titleText { text }
						}
					}
				}
			}
		}
	}
	'''

	variables = {
		"const": imdb_id,
		"first": 50,
		"filter": {"maturityLevel": "INCLUDE_MATURE","nameConstraints":{},"titleConstraints":{},"types":["TRAILER"]},
		"sort": {"by": "DATE", "order": "DESC"}
	}

	videos = []
	plot_text = ""
	item_title = ""
	total_videos = None

	# First page
	pdata = {
		'operationName': "TitleVideoGallerySubPage",
		'query': gqlmin(query_subpage + fragment),
		'variables': variables
	}
	r = requests.post(API_URL, headers=HEADERS, json=pdata)
	r.raise_for_status()
	json_data = r.json()

	title_data = json_data.get('data', {}).get('title', {})
	plot_text = title_data.get('plot', {}).get('plotText', {}).get('plainText', "")
	item_title = title_data.get('titleText', {}).get('text', "")

	video_data = title_data.get('videoStrip', {})
	total_videos = video_data.get('total')
	videos.extend([edge.get('node', {}) for edge in video_data.get('edges', [])])

	cursor = video_data.get('pageInfo', {}).get('endCursor')
	has_next = video_data.get('pageInfo', {}).get('hasNextPage', False)

	# Pagination loop
	while has_next and cursor:
		variables["after"] = cursor
		pdata = {
			'operationName': "TitleVideoGalleryPagination",
			'query': gqlmin(query_pagination + fragment),
			'variables': variables
		}
		r = requests.post(API_URL, headers=HEADERS, json=pdata)
		r.raise_for_status()
		video_data = r.json().get('data', {}).get('title', {}).get('videoStrip', {})
		videos.extend([edge.get('node', {}) for edge in video_data.get('edges', [])])
		cursor = video_data.get('pageInfo', {}).get('endCursor')
		has_next = video_data.get('pageInfo', {}).get('hasNextPage', False)
		time.sleep(0.3)

	# Match old output: inject plot, total, and item_title
	for idx, v in enumerate(videos):
		v["plot"] = plot_text
		v["total"] = total_videos
		v["item_title"] = item_title
		videos[idx] = v

	return videos

def time_format(seconds: int) -> str:
	if seconds is not None:
		seconds = int(seconds)
		d = seconds // (3600 * 24)
		h = seconds // 3600 % 24
		m = seconds % 3600 // 60
		s = seconds % 3600 % 60
		if d > 0:
			return '{:02d}D {:02d}H {:02d}m {:02d}s'.format(d, h, m, s)
		elif h > 0:
			return '{:02d}H {:02d}m {:02d}s'.format(h, m, s)
		elif m > 0:
			return '{:02d}m {:02d}s'.format(m, s)
		elif s > 0:
			return '{:02d}s'.format(s)
	return '-'

import re

def extract_season_number(title):
	# Match "Season" or "Series" followed by optional spaces, optional punctuation, and digits
	pattern = r"(:?.*(?:Season|Series))(?:\s*\d*)"
	match = re.search(pattern, title, re.IGNORECASE)
	try: extract_season_number = int(match.group(0).replace(match.group(1),'').strip())
	except: extract_season_number = None
	return extract_season_number



def find_best_trailer(trailer_list, season_number=None):
	if len(trailer_list) == 0:
		return None
	best_match = None
	best_score = -1
	fallback_thumbnail = None
	trailer_list = sorted(trailer_list, key=lambda x: x['runtime']['value'], reverse=True)

	match_list = []
	new_trailer_list = []
	season_list = []
	official_flag = False
	theatrical_list = ['theatrical','full','final']
	theatrical_flag = False
	titleText = None

	for trailer in trailer_list:
		if trailer['contentType']['id'] == 'amzn1.imdb.video.contenttype.trailer':
			curr_dict = {}
			if trailer['primaryTitle'].get('series',{}) != {}:
				try: season = int(trailer['primaryTitle']['series']['displayableEpisodeNumber']['displayableSeason']['season'])
				except: season = None
				#print(trailer)
				curr_dict['id'] =  trailer['id']
				curr_dict['vid_url'] =  'https://www.imdb.com/video/%s/?ref_=ttvg_vi_1' % (str(trailer['id']))
				curr_dict['season'] = season
				curr_dict['title'] = trailer['name']['value']
				if season:
					titleText = trailer['primaryTitle']['series']['series']['titleText']['text']
				if not season:
					season = extract_season_number(curr_dict['title'])
					if season:
						curr_dict['season'] = season

				if  any(word in str(curr_dict['title']).lower() for word in theatrical_list):
					curr_dict['theatrical'] = True
					theatrical_flag = True
				else:
					curr_dict['theatrical'] = False

				if 'official' in str(curr_dict['title']).lower():
					curr_dict['official'] = True
					official_flag = True
					if season:
						official_flag = False
						curr_dict['official'] = False
				else:
					curr_dict['official'] = False
				if season and not season in season_list:
					season_list.append(season)
				curr_dict['thumbnail'] = trailer['thumbnail']['url']
				curr_dict['runtime'] = trailer['runtime']['value']
				curr_dict['time'] = time_format(trailer['runtime']['value'])
				#print(curr_dict['title'])
				new_trailer_list.append(curr_dict)
	
	if season_number and season_number in season_list:
		season_match = True
	elif season_list != []:
		if season_number:
			for i in reversed(sorted(season_list)):
				if i <= season_number:
					break
			season_match = i
		else:
			season_match = False
	else:
		season_match = False
	
	if type(season_match) == type(season_number):
		if season_match > season_number:
			season_match = False

	offical_trailer = None
	season_trailer = None
	if season_match == True and type(season_match) == type(True):
		for trailer in new_trailer_list:
			if trailer['season'] == season_number:
				season_trailer = trailer
				break
	elif season_match == False:
		season_trailer = new_trailer_list[0]
	else:
		for trailer in new_trailer_list:
			if trailer['season'] == season_match:
				season_trailer = trailer
				break

	if theatrical_flag == True:
		for trailer in new_trailer_list:
			if trailer['theatrical']:
				offical_trailer = trailer
				break
	elif official_flag == True:
		for trailer in new_trailer_list:
			if trailer['official'] and not 'teaser' in str(trailer['title']).lower():
				offical_trailer = trailer
				break
		if not offical_trailer:
			for trailer in new_trailer_list:
				if trailer['official']:
					offical_trailer = trailer
					break

	elif titleText:
		for trailer in new_trailer_list:
			if trailer['title'] == titleText:
				offical_trailer = trailer
				break

	if offical_trailer and official_flag:
		if season_match == False or season_trailer == None:
			season_trailer = offical_trailer
	elif official_flag == False and offical_trailer:
		if season_match == False:
			season_trailer = offical_trailer
	#print(new_trailer_list)
	#print(titleText)
	return season_trailer

def extract_imdb_mp4_url(video_id):
	url = f"https://www.imdb.com/video/{video_id}?ref_=ttvg_vi_26"
	headers = {'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_11_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/50.0.2661.102 Safari/537.36'}

	response = requests.get(url, headers=headers)
	if response.status_code != 200:
		raise Exception(f"Failed to fetch page: {response.status_code}")

	html = response.text
	PlaybackURL = ('[' + html.split('"playbackURLs":[')[1].split('}]')[0] + '}]')
	url = None
	for i in eval(PlaybackURL):
		if i['videoMimeType'] == 'MP4':
			return i['url'], i
		else:
			if not url:
				url = i['url']
				video = i
		#print(i['videoDefinition'])
		#print(i['videoMimeType'])
	return url, video



all_videos = get_imdb_videos(imdb_id='tt4532368')
#print(all_videos)

best_trailer = find_best_trailer(all_videos, season_number=None)

if best_trailer:
	print(best_trailer)
	print(best_trailer['title'])


	video_url, video = extract_imdb_mp4_url(best_trailer['id'])
	print("MP4 URL:", video_url)
	print(video)
exit()


#print(json.dumps(all_videos[:3], indent=2))  # Show first 3 videos
print(f"Total videos fetched: {len(all_videos)}")
print(all_videos)
for i in all_videos:
	if 'contenttype.trailer' in str(i) and 'season' in str(i['name']['value']).lower():
		print(i['name']['value'],' - ' ,time_format(i['runtime']['value']))

I stay away from AI coding and I am not a programmer by profession either. Just trial and error in Python at hobbyist level to get things going :)

Ideally i would like to find out the nameConstraints or titleConstraints to filter only trailers to be returned rather than all videos, but havent been able to figure that out yet, may be fellow kiwi @matthuisman has some ideas

Yeah I don't do too much Ai coding either but as I don't know JavaScript at all I never would have been able to figure out how to make a compliant query without it.
But I started out trial and error in the terminal too.
Although I now do programming adjacent stuff for work. We don't really use AI as there are GDPR data issues involved but the copilot thing is new and is our own instance I believe so we've been curious about the capabilities and trialing it a bit in our area.
But I mostly do SQL on random data so ai isn't much help as half the thing is figuring out what you are looking at.

But when I am doing random things it is pretty handy to now have access to an interactive stack overflow on steroids.
Does it hallucinate occasionally, yes. But can it parse badly documented/undocumented code and give good info back, yes surprisingly it can. And getting working example code relevant to your problem when you go googling is often half the battle so it's definitely a valuable tool.

FYI the video properties returned contain contentType which has a trailer/clip info like "amzn1.imdb.video.contenttype.trailer" which might be what you need?

Otherwise maybe the query has those as variables inputs?
They might be called nameSearchConstraints?
That difference (IE search) was an error I saw myself (did you mean...)

Yes I already tried contenttype as a filter key and it came with incorrect parameter response, so yeah that is exactly what I am targeting and yet to figure out

Think this should be what you need:

"filter": {"maturityLevel": "INCLUDE_MATURE","nameConstraints":{},"titleConstraints":{},"types":["TRAILER"]},

import json
import requests
import time

def get_imdb_videos(imdb_id):
	import re, requests
	API_URL = "https://graphql.prod.api.imdb.a2z.com/"
	HEADERS = {
		'Referer': 'https://www.imdb.com/',
		'Origin': 'https://www.imdb.com',
		'User-Agent': 'Mozilla/5.0'
	}

	def gqlmin(q):
		return re.sub(' {4}', '', q)

	query_subpage = '''
	query TitleVideoGallerySubPage(
		$const: ID!,
		$first: Int!,
		$filter: VideosQueryFilter,
		$sort: VideoSort
	) {
		title(id: $const) {
			titleText { text }
			plot { plotText { plainText } }
			videoStrip(first: $first, filter: $filter, sort: $sort) {
				...VideoGalleryItems
			}
		}
	}
	'''
	query_pagination = '''
	query TitleVideoGalleryPagination(
		$const: ID!,
		$first: Int!,
		$after: ID!,
		$filter: VideosQueryFilter,
		$sort: VideoSort
	) {
		title(id: $const) {
			videoStrip(first: $first, after: $after, filter: $filter, sort: $sort) {
				...VideoGalleryItems
			}
		}
	}
	'''
	fragment = '''
	fragment VideoGalleryItems on VideoConnection {
		pageInfo {
			endCursor
			hasNextPage
		}
		total
		edges {
			node {
				id
				contentType { id }
				name { value }
				runtime { value }
				thumbnail { url }
				primaryTitle {
					series {
						displayableEpisodeNumber {
							displayableSeason {
								season
							}
						}
						series {
							titleText { text }
						}
					}
				}
			}
		}
	}
	'''

	variables = {
		"const": imdb_id,
		"first": 50,
		"filter": {"maturityLevel": "INCLUDE_MATURE","nameConstraints":{},"titleConstraints":{},"types":["TRAILER"]},
		"sort": {"by": "DATE", "order": "DESC"}
	}

	videos = []
	plot_text = ""
	item_title = ""
	total_videos = None

	# First page
	pdata = {
		'operationName': "TitleVideoGallerySubPage",
		'query': gqlmin(query_subpage + fragment),
		'variables': variables
	}
	r = requests.post(API_URL, headers=HEADERS, json=pdata)
	r.raise_for_status()
	json_data = r.json()

	title_data = json_data.get('data', {}).get('title', {})
	plot_text = title_data.get('plot', {}).get('plotText', {}).get('plainText', "")
	item_title = title_data.get('titleText', {}).get('text', "")

	video_data = title_data.get('videoStrip', {})
	total_videos = video_data.get('total')
	videos.extend([edge.get('node', {}) for edge in video_data.get('edges', [])])

	cursor = video_data.get('pageInfo', {}).get('endCursor')
	has_next = video_data.get('pageInfo', {}).get('hasNextPage', False)

	# Pagination loop
	while has_next and cursor:
		variables["after"] = cursor
		pdata = {
			'operationName': "TitleVideoGalleryPagination",
			'query': gqlmin(query_pagination + fragment),
			'variables': variables
		}
		r = requests.post(API_URL, headers=HEADERS, json=pdata)
		r.raise_for_status()
		video_data = r.json().get('data', {}).get('title', {}).get('videoStrip', {})
		videos.extend([edge.get('node', {}) for edge in video_data.get('edges', [])])
		cursor = video_data.get('pageInfo', {}).get('endCursor')
		has_next = video_data.get('pageInfo', {}).get('hasNextPage', False)
		time.sleep(0.3)

	# Match old output: inject plot, total, and item_title
	for idx, v in enumerate(videos):
		v["plot"] = plot_text
		v["total"] = total_videos
		v["item_title"] = item_title
		videos[idx] = v

	return videos




all_videos = get_imdb_videos(imdb_id='tt11280740')
print(all_videos)
exit()

however other than "types", nameConstraints are "nameConstraints":{allNameIds":["nm0004395","nm3138882"]} "

And titleconstraints:

nameConstraints: {
                    allNameIds: r.nameIds?.sort( (e, t) => e.localeCompare(t))
                },
                titleConstraints: {
                    anyTitleIds: r.titleIds?.sort( (e, t) => e.localeCompare(t))

ie anyTitleIds being - tt11280740 eg IMDB ids. So not actually "video clip name" or "video clip title"

found a small issue when a trailer contains Final but is for final season eg. Beter Call Sault "Better Call Saul: A Look At The Final Season" It gets picked up as theatrical trailer To stop this, i just did a 2nd check that it doesnt contain season

@Gujal00, @matthuisman

checkout Gujal00/Kodi-Official@c556e4e

I've provided working API lookups for all the pages currently scraped:

VideoPlayback => https://www.imdb.com/video/vi1020905497/?ref_=ttvg_vi_1
CalendarPage => https://www.imdb.com/calendar/?region=US&type=MOVIE&ref_=rlm
movies_near_you => https://www.imdb.com/showtimes/