Accessing videos in the photo library

[ifuchs]

Sorry for cross-posting but I wasn't sure whether here or the other Pythonista-issues page was the right place,

I am trying to write a script (with help from ChatGPT) to get the latest video file from my Photo library, upload to Nextcloud, and return a shareable link to the file. I have a script which works except it won’t do videos. That is, it works fine for .img files but not .mov. ChatGPT has concluded:
**Pythonista cannot access actual video files from the Photos library.

This is a known limitation: while it can work with image filepaths (.JPG, etc.), Apple restricts access to full video assets via Pythonista’s photos module — unless exported manually first (e.g., via Shortcuts or Share Sheet).**

Any help would be most appreciated (of the human or AI variety).

Here is the script:
`import photos
import requests
import clipboard
import os
from datetime import datetime
from base64 import b64encode
import re
——— Configuration ———

USERNAME = ''
APP_PASSWORD = '' # Replace with your real app password
BASE_WEBDAV = ''
UPLOAD_FOLDER = 'SharedVideos/'
SHARE_API = ''
——— Step 1: Get latest asset ———

images = photos.get_assets(media_type='image')
videos = photos.get_assets(media_type='video')
assets = images + videos
assets.sort(key=lambda a: a.creation_date or datetime.min, reverse=True)

if not assets:
print("❌ No media found.")
exit()

latest = assets[0]
media_type = latest.media_type
timestamp = datetime.now().strftime('%Y-%m-%d_%H-%M-%S')
ext = 'jpg' if media_type == 'image' else 'mov'
filename = f"Upload-{timestamp}.{ext}"

print(f"📸 Selected: {filename} ({media_type})")
——— Step 2: Upload ———

auth_header = 'Basic ' + b64encode(f'{USERNAME}:{APP_PASSWORD}'.encode()).decode()
upload_url = f"{BASE_WEBDAV}{UPLOAD_FOLDER}{filename}"

if media_type == 'image':
data = latest.get_image_data()
print("📤 Uploading image...")
response = requests.put(upload_url, headers={
'Authorization': auth_header
}, data=data)
else:
try:
index = videos.index(latest)
metadata = photos.get_metadata(index)
video_path = metadata.get('filepath', None)

if not video_path or not os.path.exists(video_path):
    print("❌ Video path not found or invalid.")
    exit()

print("📤 Uploading video as stream...")
with open(video_path, 'rb') as f:
    response = requests.put(upload_url, headers={
        'Authorization': auth_header
    }, data=f)

except Exception as e:
print(f"❌ Error during video upload: {e}")
exit()

——— Step 3: Check upload ———

if not (200 <= response.status_code < 300):
print(f"❌ Upload failed with status {response.status_code}")
print(response.text)
exit()

print("✅ Upload successful.")
——— Step 4: Generate share link ———

share_response = requests.post(SHARE_API, headers={
'Authorization': auth_header,
'OCS-APIRequest': 'true',
'Content-Type': 'application/x-www-form-urlencoded'
}, data={
'path': f'{UPLOAD_FOLDER}{filename}',
'shareType': 3
})

if share_response.status_code != 200:
print(f"❌ Share creation failed: {share_response.status_code}")
print(share_response.text)
exit()

match = re.search(r'(.*?)', share_response.text)
if match:
link = match.group(1)
clipboard.set(link)
print(f"🔗 Link copied to clipboard:\n{link}")
else:
print("❌ Could not extract share link.")
print(share_response.text)

[cvpe]

As usual (you know me) I'm not sure I correctly understand your problem. If not, sorry, and forget my comment.
Here after you can find a working script to copy a video from the camera roll to a local file:

import photos
from objc_util import *
import time
import shutil
import os

# Récupère les vidéos de la Pellicule
videos = photos.get_assets(media_type='video')
assets = photos.pick_asset(assets=videos, multi=True)

# Prépare les options pour obtenir l'AVAsset original
options = ObjCClass('PHVideoRequestOptions').new()
options.version = 1  # 1 = PHVideoRequestOptionsVersionOriginal

image_manager = ObjCClass('PHImageManager').defaultManager()

handled_assets = []

# Fonction appelée de manière asynchrone pour chaque vidéo
def handleAsset(_cmd, av_asset_ptr, audioMix_ptr, info_ptr):
    av_asset = ObjCInstance(av_asset_ptr)
    
    # Vérifie qu'on a un AVURLAsset
    if av_asset.isKindOfClass_(ObjCClass('AVURLAsset')):
        url = av_asset.URL()
        path = str(url.path())
        print(f"Ajout de: {path}")
        handled_assets.append(path)
    else:
        print("⚠️ Type d’AVAsset non pris en charge")

# Crée un bloc Objective-C pour le callback
handlerblock = ObjCBlock(handleAsset, argtypes=[c_void_p]*4)

# Lance la récupération des vidéos (asynchrone)
for A in assets:
    image_manager.requestAVAssetForVideo_options_resultHandler_(A, options, handlerblock)

# Attend que tous les assets aient été traités
while len(handled_assets) < len(assets):
    time.sleep(1)

# Copie chaque fichier vidéo dans le répertoire courant
for path in handled_assets:
    try:
        local = os.path.basename(path)
        print(f"📥 Copie de {path} → {local}")
        shutil.copy(path, local)
    except Exception as e:
        print(f"❌ Erreur de copie: {e}")```

[ifuchs]

@cvpe Thanks for the reply and the script. After adding "import photos", I got it to run but this is not exactly what I need. Your script does indeed open the file picker and allow me to select a file (including a video) and download it to Pythonista's file space. However, What I am looking to do is to have the script find the last video in the library with no manual intervention and then upload that file to Nextcloud via webdav and then get a share link to the file using the Nextcloud Web API. All of this is in the script I posted above but it fails for some reason that I don't fully understand.

[cvpe]

I understand your problem but you said that Apple did not allow to pick a video file, manually excepted, and my little script shows it is possible.

I just take a video, thus last item in my Photos, and this works

images = photos.get_assets(media_type='image')
videos = photos.get_assets(media_type='video')
assets = images + videos
assets.sort(key=lambda a: a.creation_date or datetime.min, reverse=True)

if not assets:
    print("❌ No media found.")
    exit()

latest = assets[0]
assets = [latest]

[cvpe]

I think that

index = videos.index(latest)
metadata = photos.get_metadata(index)

Is not correct. Index is the index of your asset in only the videos but you access the index-th of all photos

[cvpe]

Perhaps, you could try my code to have a copy of the video in a local temporary file, read this file in binary mode in your "data" for upload, then remove the local temporary file, just to see if it works.

[ifuchs]

I've tried but I can't seem to get it to work. I can get it to upload the latest .jpg but not a video for some reason. According to ChatGPt:
Pythonista’s photos.get_assets() does not reliably return recent videos, even though they’re in the library.
• Videos are skipped even if they are the most recent files.
• photos.get_assets(media_type='all') still returns only images in some cases.
• Attempts to export or get metadata for those videos fail silently or crash.
I may go back to doing this manually as I have now tried dozens of failing variations.

[ifuchs]

This is a better formatted version of the script that works except the uploaded video is corrupt.

import requests
import clipboard
from datetime import datetime
from base64 import b64encode
import re

USERNAME = 'X'
APP_PASSWORD = 'X'  # Replace with your real app password
BASE_WEBDAV = 'X'
UPLOAD_FOLDER = 'SharedVideos/'
SHARE_API = 'X'

# Get all image and video assets
images = photos.get_assets(media_type='image')
videos = photos.get_assets(media_type='video')
assets = images + videos
assets.sort(key=lambda a: a.creation_date or datetime.min, reverse=True)

if not assets:
    print("❌ No media found.")
    exit()

latest = assets[0]
media_type = latest.media_type
timestamp = datetime.now().strftime('%Y-%m-%d_%H-%M-%S')
ext = 'jpg' if media_type == 'image' else 'mov'
filename = f"Upload-{timestamp}.{ext}"

print(f"📸 Selected: {filename} ({media_type})")

auth_header = 'Basic ' + b64encode(f'{USERNAME}:{APP_PASSWORD}'.encode()).decode()
upload_url = f"{BASE_WEBDAV}{UPLOAD_FOLDER}{filename}"

try:
    # Note: This is not ideal for videos, but avoids crash
    data = latest.get_image_data()
    print("📤 Uploading media...")
    response = requests.put(upload_url, headers={
        'Authorization': auth_header
    }, data=data)

except Exception as e:
    print(f"❌ Error during upload: {e}")
    exit()

if not (200 <= response.status_code < 300):
    print(f"❌ Upload failed with status {response.status_code}")
    print(response.text)
    exit()

print("✅ Upload successful.")

# Create public share link
share_response = requests.post(SHARE_API, headers={
    'Authorization': auth_header,
    'OCS-APIRequest': 'true',
    'Content-Type': 'application/x-www-form-urlencoded'
}, data={
    'path': f'{UPLOAD_FOLDER}{filename}',
    'shareType': 3
})

if share_response.status_code != 200:
    print(f"❌ Share creation failed: {share_response.status_code}")
    print(share_response.text)
    exit()

match = re.search(r'<url>(.*?)</url>', share_response.text)
if match:
    link = match.group(1)
    clipboard.set(link)
    print(f"🔗 Link copied to clipboard:\n{link}")
else:
    print("❌ Could not extract share link.")
    print(share_response.text)```

[cvpe]

Sincerely, I can't understand because I can get my last video and copy it as a local file, and I can read it without any problem. This perhaps is your upload the problem.

[ifuchs]

All I know is that when I run the above script, it gets the file from the Photos library, uploads it to Nextcloud, and when I try to view it it is unviewable.

[ifuchs]

can you please give me a version of your script that does not use the file picker but gets the last video in your library. If I could see how that works, I could perhaps get my script to do what I need. I don't want a local copy but if it is necessary to be able to do a correct upload then that's fine. As long as there is no manual intervention.

[cvpe]

import photos
from objc_util import *
import time
import shutil
import os

# get images and videos from camera roll
#images = photos.get_assets(media_type='image')
videos = photos.get_assets(media_type='video')
#assets = photos.pick_asset(assets=videos, multi=True)
#assets = images + videos
assets = videos						# use only videos
assets.sort(key=lambda a: a.creation_date or datetime.min, reverse=True)


if not assets:
    print("❌ No media found.")
    exit()

latest = assets[0]				# get latest video

assets = [latest]

# Prépare options to get the original AVAsset
options = ObjCClass('PHVideoRequestOptions').new()
options.version = 1  # 1 = PHVideoRequestOptionsVersionOriginal

image_manager = ObjCClass('PHImageManager').defaultManager()

handled_assets = []

# Function called in a  asynchroneous way for each video
def handleAsset(_cmd, av_asset_ptr, audioMix_ptr, info_ptr):
    av_asset = ObjCInstance(av_asset_ptr)
    
    # check it is a AVURLAsset
    if av_asset.isKindOfClass_(ObjCClass('AVURLAsset')):
        url = av_asset.URL()
        path = str(url.path())
        print(f"Add: {path}")
        handled_assets.append(path)
    else:
        print("⚠️ AVAsset type not supported")

# Create Objective-C block for callback
handlerblock = ObjCBlock(handleAsset, argtypes=[c_void_p]*4)

# Get vidéos (asynchroneous)
for A in assets:
    image_manager.requestAVAssetForVideo_options_resultHandler_(A, options, handlerblock)

# Wait for all vidos have been processed
while len(handled_assets) < len(assets):
    time.sleep(1)

# Copy each vidéo in current folder
for path in handled_assets:
    try:
        local = os.path.basename(path)
        print(f"📥 copy {path} → {local}")
        shutil.copy(path, local)
    except Exception as e:
        print(f"❌ Copy error: {e}")

# you could now read this local file as binary in data variable for tour upload

now you code

print("📤 Uploading video as stream...")
with open(video_path, 'rb') as f:
    response = requests.put(upload_url, headers={
        'Authorization': auth_header
    }, data=f)

should work with path of local copy

[ifuchs]

You get the genius award! That worked perfectly. I combined your code with mine to upload and get the Share URL back from Nextcloud. Thanks very much!