metadata >> People (bulk refresh posters)

[crunchyemby 9]

Hello All

 

First of all, Emby is turning out to be a pretty amazing application... Very glad I found it. I have attempted to find the answer to my problem, before posting, without a clear answer. So, here goes:

 

Due to inconsistencies and mis-matched identification, I have disabled all metadata services and authored a script to publish my own "People" data to the emby metadata location ( /var/db/emby-server/metadata/People/*/* ). The problem is, the updated images "poster.jpg" are not being picked up on a "Scheduled tasks >> Refresh people" ... Rather, I have to manually "Refresh", from the context menu, for each Person. Obviously this is not going to work for the long-run.

 

I have found a workaround, if I delete these two files: "library.db" and "refreshinfo.db" ... I can rebuild the complete database and start over. Also a pain in the ... but better than manually refreshing through the GUI.

 

Any recommendations?

 

Cheers!

[crunchyemby 9]

In case anybody else is having this problem, here's a snippet to get you started... I'm still trying to track down some "enhanced image caching" that occurs at times... but I have tested that the functionality works as expected.

 

Cheers!

 

 

BEGIN PYTHON SCRIPT:

import requests
import json
import hashlib

#config
emby_server_ip = 'xxx.xxx.xxx.xxx' #EXAMPLE: 192.168.0.100'
emby_server_port = 'xxxx' #EXAMPLE: '8096'
emby_username = 'USERNAME'
password_plain_text = 'PASSWORD'
refresh_name_list = [] #EXAMPLE: ['John Doe', 'Mary Palmer']

def authenticate_user():
    global user_id
    global user_api_key
    password_sha1 = hashlib.sha1(password_plain_text).hexdigest()
    password_md5 = hashlib.md5(password_plain_text).hexdigest()
    request_url = "http://%s:%s/Users/AuthenticateByName" % (emby_server_ip, emby_server_port)
    request_params = {'Username': emby_username, 'Password': password_sha1, 'PasswordMd5': password_md5, 'Authorization': 'MediaBrowser', 'Client': 'Python', 'Device': 'PythonScript', 'DeviceId': '12345678910987654321', 'Version': '1.0.0.0', 'format': 'json'}
    r = requests.post(request_url, params=request_params)
    print 'Authenticating. Status: %s' % (r.status_code)
    j = json.loads(r.text)
    user_id = j['User']['Id']
    user_api_key = j['AccessToken']
    print 'user_id: %s' % (user_id)
    print 'user_api_key: %s' % (user_api_key)
    return user_id, user_api_key

def refresh_primary_image_for_person(full_name):
    print '\nProcessing refresh: %s' % (full_name)
    search_name = full_name.strip().replace(' ', '+')
    request_url = "http://%s:%s/Persons/%s" % (emby_server_ip, emby_server_port, search_name)
    request_params = {'api_key': user_api_key, 'format': 'json'}
    r = requests.get(request_url, params=request_params)
    print 'Fetch primary image id, response: (%s, %s)' % (r.status_code, r.reason)
    j = json.loads(r.text)
    try:
        primary_image_id = j['ImageTags']['Primary']
        print 'primary_image_id: %s' % (primary_image_id)
    except:
        print 'WARNING: primary image not found for: %s' % (full_name)
        return
    request_url = 'http://%s:%s/emby/Items/%s/Refresh?Recursive=true&ImageRefreshMode=FullRefresh&MetadataRefreshMode=FullRefresh&ReplaceAllImages=false&ReplaceAllMetadata=true' % (emby_server_ip, emby_server_port, primary_image_id)
    params={'UserId': user_id, 'X-MediaBrowser-Token': user_api_key, 'Authorization': 'MediaBrowser', 'Client': 'Python', 'Device': 'PythonScript', 'DeviceId': '12345678910987654321', 'Version': '1.0.0.0', 'Recursive': 'true', 'ImageRefreshMode': 'FullRefresh', 'MetadataRefreshMode': 'FullRefresh', 'ReplaceAllImages': 'true', 'ReplaceAllMetadata': 'true'}
    r = requests.post(request_url, params=request_params)
    print 'Refresh primary image. Status: (%s, %s)' % (r.status_code, r.reason)
    if str(r.status_code) == '204':
        print 'STATUS: %s\'s primary image was refreshed.' % (full_name)
    return

user_id, user_api_key = authenticate_user()
for name in refresh_name_list:
    refresh_primary_image_for_person(name)

Edited December 18, 2015 by crunchyemby

[crunchyemby 9]

I'm pulling my hair out, tracking down the cached images...

 

Things I've tried:

• Cleared the browser cache.
• [DISABLED] Manage Server >> Advanced >> Developer Options >> Enable web client response caching
• [DISABLED] Manage Server >> Advanced >> Developer Options >> Enable web client resource minification
• Looked through the SQL database, found the People table... confirmed that there is not a cache path.
• Found a json data object in the SQL table 'TypedBaseItems', containing the path to the Person's directory... confirmed that there is not a reference to a cached resource.
• Searched emby-server/* file system for any reference to the image ID or tag=ID, found in the image url... no result found.

OBSERVATION:

From my black box testing, there seems to be inconsistent caching rules, depending on where the images are used... for example, the Primary image on this page [server:port/web/itemdetails.html?id=PERSONID] does not adhere to the same rules as the Person's Primary Image on [server:port/web/itemdetails.html?id=MOVIEID] page. 

POSSIBLE EXPLANATION:

One possible explanation is the optimization (ImageMagick) of the source images are getting stored somewhere (yet to be discovered by me) and not being refreshed... or have a TTL (time-to-live) set on the asset, which makes testing very difficult.

 

ANYBODY:

Please, if anybody can point me where to look for TTL/caching rules, cached assets with a method for cross reference, other (e.g., Fix my api call above. Maybe I'm missing a critical header in my request?)...