TvTimeToTrakt/TimeToTrackt.py

#!/usr/bin/env python3
import csv
import json
import logging
import os
import re
import sys
import time
from datetime import datetime
from pathlib import Path

import trakt.core
from tinydb import Query, TinyDB
from trakt import init
from trakt.tv import TVShow

# Setup logger
logging.basicConfig(
    format="%(asctime)s [%(levelname)7s] :: %(message)s",
    level=logging.INFO,
    datefmt="%Y-%m-%d %H:%M:%S",
)

# Adjust this value to increase/decrease your requests between episodes.
# Make to remain within the rate limit: https://trakt.docs.apiary.io/#introduction/rate-limiting
DELAY_BETWEEN_EPISODES_IN_SECONDS = 0.75

# Create a database to keep track of completed processes
database = TinyDB("localStorage.json")
syncedEpisodesTable = database.table("SyncedEpisodes")
userMatchedShowsTable = database.table("TvTimeTraktUserMatched")


class Expando(object):
    pass


def isAuthenticated():
    with open(f"{Path.home()}/.pytrakt.json") as f:
        data = json.load(f)
        daysBeforeExpiration = (
            datetime.fromtimestamp(data["OAUTH_EXPIRES_AT"]) - datetime.now()
        ).days
        if daysBeforeExpiration < 1:
            return False
        return True


def getConfiguration():
    configEx = Expando()

    with open("config.json") as f:
        data = json.load(f)

    configEx.TRAKT_USERNAME = data["TRAKT_USERNAME"]
    configEx.CLIENT_ID = data["CLIENT_ID"]
    configEx.CLIENT_SECRET = data["CLIENT_SECRET"]
    configEx.GDPR_WORKSPACE_PATH = data["GDPR_WORKSPACE_PATH"]

    CONFIG_SINGLETON = configEx

    return CONFIG_SINGLETON


config = getConfiguration()

# Return the path to the CSV file contain the watched episode data from TV Time


def getWatchedShowsPath():
    return config.GDPR_WORKSPACE_PATH + "/seen_episode.csv"


def getFollowedShowsPath():
    return config.GDPR_WORKSPACE_PATH + "/followed_tv_show.csv"


def initTraktAuth():
    if isAuthenticated():
        return True
    # Set the method of authentication
    trakt.core.AUTH_METHOD = trakt.core.OAUTH_AUTH
    return init(
        config.TRAKT_USERNAME,
        store=True,
        client_id=config.CLIENT_ID,
        client_secret=config.CLIENT_SECRET,
    )


# With a given title, check if it contains a year (e.g Doctor Who (2005))
# and then return this value, with the title and year removed to improve
# the accuracy of Trakt results.


def getYearFromTitle(title):
    ex = Expando()

    try:
        # Use a regex expression to get the value within the brackets e.g The Americans (2017)
        yearSearch = re.search(r"\(([A-Za-z0-9_]+)\)", title)
        yearValue = yearSearch.group(1)
        # Then, get the title without the year value included
        titleValue = title.split("(")[0].strip()
        # Put this together into an object
        ex.titleWithoutYear = titleValue
        ex.yearValue = int(yearValue)
        return ex
    except Exception:
        # If the above failed, then the title doesn't include a year
        # so return the object as is.
        ex.titleWithoutYear = title
        ex.yearValue = -1
        return ex


# Shows in TV Time are often different to Trakt.TV - in order to improve results and automation,
# calculate how many words are in the title, and return true if more than 50% of the title is a match,
# It seems to improve automation, and reduce manual selection....


def checkTitleNameMatch(tvTimeTitle, traktTitle):
    # If the name is a complete match, then don't bother comparing them!
    if tvTimeTitle == traktTitle:
        return True

    # Split the TvTime title
    tvTimeTitleSplit = tvTimeTitle.split()

    # Create an array of words which are found in the Trakt title
    wordsMatched = []

    # Go through each word of the TV Time title, and check if it's in the Trakt title
    for word in tvTimeTitleSplit:
        if word in traktTitle:
            wordsMatched.append(word)

    # Then calculate what percentage of words matched
    quotient = len(wordsMatched) / len(traktTitle.split())
    percentage = quotient * 100

    # If more than 50% of words in the TV Time title exist in the Trakt title,
    # then return the title as a possibility to use
    return percentage > 50


# Using TV Time data (Name of Show, Season No and Episode) - find the corresponding show
# in Trakt.TV either by automation, or asking the user to confirm.


def getShowByName(name, seasonNo, episodeNo):
    # Parse the TV Show's name for year, if one is present in the string
    titleObj = getYearFromTitle(name)

    # Create a boolean to indicate if the title contains a year,
    # this is used later on to improve the accuracy of picking
    # from search results
    doesTitleIncludeYear = titleObj.yearValue != -1

    # If the title contains a year, then replace the local variable with the stripped version
    if doesTitleIncludeYear:
        name = titleObj.titleWithoutYear

    # Request the Trakt API for search results, using the name
    tvSearch = TVShow.search(name)

    # Create an array of shows which have been matched
    showsWithSameName = []

    # Go through each result from the search
    for show in tvSearch:
        # Check if the title is a match, based on our conditions (e.g over 50% of words match)
        if checkTitleNameMatch(name, show.title):
            # If the title included the year of broadcast, then we can be more picky in the results
            # to look for a show with a broadcast year that matches
            if doesTitleIncludeYear:
                # If the show title is a 1:1 match, with the same broadcast year, then bingo!
                if (name == show.title) and (show.year == titleObj.yearValue):
                    # Clear previous results, and only use this one
                    showsWithSameName = []
                    showsWithSameName.append(show)
                    break

                # Otherwise, only add the show if the broadcast year matches
                if show.year == titleObj.yearValue:
                    showsWithSameName.append(show)
            # If the program doesn't have the broadcast year, then add all the results
            else:
                showsWithSameName.append(show)

    # Sweep through the results once more for 1:1 title name matches,
    # then if the list contains one entry with a 1:1 match, then clear the array
    # and only use this one!
    completeMatchNames = []
    for nameFromSearch in showsWithSameName:
        if nameFromSearch.title == name:
            completeMatchNames.append(nameFromSearch)

    if len(completeMatchNames) == 1:
        showsWithSameName = completeMatchNames

    # If the search contains multiple results, then we need to confirm with the user which show
    # the script should use, or access the local database to see if the user has already provided
    # a manual selection
    if len(showsWithSameName) > 1:

        # Query the local database for existing selection
        userMatchedQuery = Query()
        queryResult = userMatchedShowsTable.search(userMatchedQuery.ShowName == name)

        # If the local database already contains an entry for a manual selection
        # then don't bother prompting the user to select it again!
        if len(queryResult) == 1:
            # Get the first result from the query
            firstMatch = queryResult[0]
            # Get the value contains the selection index
            firstMatchSelectedIndex = int(firstMatch.get("UserSelectedIndex"))
            # Check if the user previously requested to skip the show
            skipShow = firstMatch.get("SkipShow")
            # If the user did not skip, but provided an index selection, get the
            # matching show
            if not skipShow:
                return showsWithSameName[firstMatchSelectedIndex]
            # Otherwise, return None, which will trigger the script to skip
            # and move onto the next show
            else:
                return None
        # If the user has not provided a manual selection already in the process
        # then prompt the user to make a selection
        else:
            print(
                f"INFO - MANUAL INPUT REQUIRED: The TV Time data for Show '{name}' (Season {seasonNo}, Episode {episodeNo}) has {len(showsWithSameName)} matching Trakt shows with the same name."
            )

            # Output each show for manual selection
            for idx, item in enumerate(showsWithSameName):
                # Display the show's title, broadcast year, amount of seasons and a link to the Trakt page.
                # This will provide the user with enough information to make a selection.
                print(
                    f"    ({idx + 1}) {item.title} - {item.year} - {len(item.seasons)} Season(s) - More Info: https://trakt.tv/{item.ext}"
                )

            while True:
                try:
                    # Get the user's selection, either a numerical input, or a string 'SKIP' value
                    indexSelected = input(
                        "Please make a selection from above (or enter SKIP):"
                    )

                    if indexSelected != "SKIP":
                        # Since the value isn't 'skip', check that the result is numerical
                        indexSelected = int(indexSelected) - 1
                        # Exit the selection loop
                        break
                    # Otherwise, exit the loop
                    else:
                        break
                # Still allow the user to provide the exit input, and kill the program
                except KeyboardInterrupt:
                    sys.exit("Cancel requested...")
                # Otherwise, the user has entered an invalid value, warn the user to try again
                except Exception:
                    logging.error(
                        f"Sorry! Please select a value between 0 to {len(showsWithSameName)}"
                    )

            # If the user entered 'SKIP', then exit from the loop with no selection, which
            # will trigger the program to move onto the next episode
            if indexSelected == "SKIP":
                # Record that the user has skipped the TV Show for import, so that
                # manual input isn't required everytime
                userMatchedShowsTable.insert(
                    {"ShowName": name, "UserSelectedIndex": 0, "SkipShow": True}
                )

                return None
            # Otherwise, return the selection which the user made from the list
            else:
                selectedShow = showsWithSameName[int(indexSelected)]

                userMatchedShowsTable.insert(
                    {
                        "ShowName": name,
                        "UserSelectedIndex": indexSelected,
                        "SkipShow": False,
                    }
                )

                return selectedShow

    else:
        if len(showsWithSameName) > 0:
            # If the search returned only one result, then awesome!
            # Return the show, so the import automation can continue.
            return showsWithSameName[0]
        else:
            return None


# Since the Trakt.Py starts the indexing of seasons in the array from 0 (e.g Season 1 in Index 0), then
# subtract the TV Time numerical value by 1 so it starts from 0 as well. However, when a TV series includes
# a 'special' season, Trakt.Py will place this as the first season in the array - so, don't subtract, since
# this will match TV Time's existing value.


def parseSeasonNo(seasonNo, traktShowObj):
    # Parse the season number into a numerical value
    seasonNo = int(seasonNo)

    # Then get the Season Number from the first item in the array
    firstSeasonNo = traktShowObj.seasons[0].number

    # If the season number is 0, then the Trakt show contains a "special" season
    if firstSeasonNo == 0:
        # No need to modify the value, as the TV Time value will match Trakt
        return seasonNo
    # Otherwise, if the Trakt seasons start with no specials, then return the seasonNo,
    # but subtracted by one (e.g Season 1 in TV Time, will be 0)
    else:
        # Only subtract if the TV Time season number is greater than 0.
        if seasonNo != 0:
            return seasonNo - 1
        # Otherwise, the TV Time season is a special! Then you don't need to change the starting position
        else:
            return seasonNo


def processWatchedShows():
    # Total amount of rows which have been processed in the CSV file
    rowsCount = 0
    # Total amount of rows in the CSV file
    errorStreak = 0
    # Open the CSV file within the GDPR exported data
    with open(getWatchedShowsPath(), newline="") as csvfile:
        # Create the CSV reader, which will break up the fields using the delimiter ','
        showsReader = csv.DictReader(csvfile, delimiter=",")
        # Get the total amount of rows in the CSV file,
        rowsTotal = len(list(showsReader))
        # Move position to the beginning of the file
        csvfile.seek(0, 0)
        # Loop through each line/record of the CSV file
        # Ignore the header row
        next(showsReader, None)
        for rowsCount, row in enumerate(showsReader):
            # Get the name of the TV show
            tvShowName = row["tv_show_name"]
            # Get the TV Time Episode Id
            tvShowEpisodeId = row["episode_id"]
            # Get the TV Time Season Number
            tvShowSeasonNo = row["episode_season_number"]
            # Get the TV Time Episode Number
            tvShowEpisodeNo = row["episode_number"]
            # Get the date which the show was marked 'watched' in TV Time
            tvShowDateWatched = row["updated_at"]
            # Parse the watched date value into a Python type
            tvShowDateWatchedConverted = datetime.strptime(
                tvShowDateWatched, "%Y-%m-%d %H:%M:%S"
            )

            # Query the local database for previous entries indicating that
            # the episode has already been imported in the past. Which will
            # ease pressure on TV Time's API server during a retry of the import
            # process, and just save time overall without needing to create network requests
            episodeCompletedQuery = Query()
            queryResult = syncedEpisodesTable.search(
                episodeCompletedQuery.episodeId == tvShowEpisodeId
            )

            # If the query returned no results, then continue to import it into Trakt
            if len(queryResult) == 0:
                # Create a repeating loop, which will break on success, but repeats on failures
                while True:
                    # If more than 10 errors occurred in one streak, whilst trying to import the episode
                    # then give up, and move onto the next episode, but warn the user.
                    if errorStreak > 10:
                        logging.warning(
                            "An error occurred 10 times in a row... skipping episode..."
                        )
                        break
                    try:
                        # Sleep for a second between each process, before going onto the next watched episode.
                        # This is required to remain within the API rate limit, and use the API server fairly.
                        # Other developers share the service, for free - so be considerate of your usage.
                        time.sleep(DELAY_BETWEEN_EPISODES_IN_SECONDS)
                        # Search Trakt for the TV show matching TV Time's title value
                        traktShowObj = getShowByName(
                            tvShowName, tvShowSeasonNo, tvShowEpisodeNo
                        )
                        # If the method returned 'None', then this is an indication to skip the episode, and
                        # move onto the next one
                        if traktShowObj is None:
                            break
                        # Show the progress of the import on-screen
                        logging.info(
                            f"({rowsCount+1}/{rowsTotal}) - Processing '{tvShowName}' Season {tvShowSeasonNo} / Episode {tvShowEpisodeNo}"
                        )
                        # Get the season from the Trakt API
                        season = traktShowObj.seasons[
                            parseSeasonNo(tvShowSeasonNo, traktShowObj)
                        ]
                        # Get the episode from the season
                        episode = season.episodes[int(tvShowEpisodeNo) - 1]
                        # Mark the episode as watched!
                        episode.mark_as_seen(tvShowDateWatchedConverted)
                        # Add the episode to the local database as imported, so it can be skipped,
                        # if the process is repeated
                        syncedEpisodesTable.insert({"episodeId": tvShowEpisodeId})
                        # Clear the error streak on completing the method without errors
                        errorStreak = 0
                        break
                    # Catch errors which occur because of an incorrect array index. This occurs when
                    # an incorrect Trakt show has been selected, with season/episodes which don't match TV Time.
                    # It can also occur due to a bug in Trakt Py, whereby some seasons contain an empty array of episodes.
                    except IndexError:
                        tvShowSlug = traktShowObj.to_json()["shows"][0]["ids"]["ids"][
                            "slug"
                        ]
                        logging.warning(
                            f"({rowsCount}/{rowsTotal}) - {tvShowName} Season {tvShowSeasonNo}, Episode {tvShowEpisodeNo} does not exist in Trakt! (https://trakt.tv/shows/{tvShowSlug}/seasons/{tvShowSeasonNo}/episodes/{tvShowEpisodeNo})"
                        )
                        break
                    # Catch any errors which are raised because a show could not be found in Trakt
                    except trakt.errors.NotFoundException:
                        logging.warning(
                            f"({rowsCount}/{rowsTotal}) - {tvShowName} Season {tvShowSeasonNo}, Episode {tvShowEpisodeNo} does not exist (search) in Trakt!"
                        )
                        break
                    # Catch errors because of the program breaching the Trakt API rate limit
                    except trakt.errors.RateLimitException:
                        logging.warning(
                            "The program is running too quickly and has hit Trakt's API rate limit! Please increase the delay between "
                            + "episdoes via the variable 'DELAY_BETWEEN_EPISODES_IN_SECONDS'. The program will now wait 60 seconds before "
                            + "trying again."
                        )
                        time.sleep(60)

                        # Mark the exception in the error streak
                        errorStreak += 1
                    # Catch a JSON decode error - this can be raised when the API server is down and produces a HTML page, instead of JSON
                    except json.decoder.JSONDecodeError:
                        logging.warning(
                            f"({rowsCount}/{rowsTotal}) - A JSON decode error occuring whilst processing {tvShowName} "
                            + f"Season {tvShowSeasonNo}, Episode {tvShowEpisodeNo}! This might occur when the server is down and has produced "
                            + "a HTML document instead of JSON. The script will wait 60 seconds before trying again."
                        )

                        # Wait 60 seconds
                        time.sleep(60)

                        # Mark the exception in the error streak
                        errorStreak += 1
                    # Catch a CTRL + C keyboard input, and exits the program
                    except KeyboardInterrupt:
                        sys.exit("Cancel requested...")
            # Skip the episode
            else:
                logging.info(
                    f"({rowsCount}/{rowsTotal}) - Already imported, skipping '{tvShowName}' Season {tvShowSeasonNo} / Episode {tvShowEpisodeNo}."
                )


def start():
    # Create the initial authentication with Trakt, before starting the process
    if initTraktAuth():
        # Display a menu selection
        print(">> What do you want to do?")
        print("    1) Import Watch History from TV Time")

        while True:
            try:
                menuSelection = input("Enter your menu selection: ")
                menuSelection = 1 if not menuSelection else int(menuSelection)
                break
            except ValueError:
                logging.warning("Invalid input. Please enter a numerical number.")
        # Start the process which is required
        if menuSelection == 1:
            # Invoke the method which will import episodes which have been watched
            # from TV Time into Trakt
            processWatchedShows()
        else:
            logging.warning("Sorry - that's an unknown menu selection")
    else:
        logging.error(
            "ERROR: Unable to complete authentication to Trakt - please try again."
        )


if __name__ == "__main__":
    # Check that the user has created the config file
    if os.path.exists("config.json"):
        # Check that the user has provided the GDPR path
        if os.path.isdir(config.GDPR_WORKSPACE_PATH):
            start()
        else:
            logging.error(
                "Oops! The TV Time GDPR folder '"
                + config.GDPR_WORKSPACE_PATH
                + "' does not exist on the local system. Please check it, and try again."
            )
    else:
        logging.error(
            "The 'config.json' file cannot be found - have you created it yet?"
        )