Dear fellow selfhosters,

If you use immich and have a digital camera that shoots JPG+RAW, you must have noticed the duplicate images taking up your screen space. I recently found out that immich has a neat feature called stacking where you can group images in the timeline. I wrote a very simple Python script to search and stack the JPG and RAW images in my instance and thought I would share it with the community. Make sure you edit the search parameters and API key and also read the whole script before running it.

For advanced immich stacking use this github.com/tenekev/immich-auto…

NOTE: I did not know this project existed before I wrote the script :)

Happy Holidays..

```import json
import requests
from pathlib import Path
from collections import defaultdict

--- Configuration & Constants ---

API_KEY = "API_KEY"
BASE_URL = "https://immich.local.website.tld/"
RAW_FILE_EXT = ".RAF"
HEADERS = {
"Content-Type": "application/json",
"Accept": "application/json",
"x-api-key": API_KEY
}
STACKS_URL = f"{BASE_URL}/api/stacks"
SEARCH_URL = f"{BASE_URL}/api/search/metadata"
ASSETS_URL = f"{BASE_URL}/api/assets" # For checking if an asset is already stacked

---------------------------------

1. CREATE SEARCH PAYLOAD

---------------------------------

def create_search_payload(page: int) -> str:
"""
Build the JSON payload to send with the search request.
Modify search settings for your camera
"""
payload = {
"make": "FUJIFILM",
"size": 1000,
"page": page,
"model": "X-S20",
"takenAfter": "2024-12-20T00:00:00.000Z"
}
return json.dumps(payload)

---------------------------------

2. FETCH SEARCH RESULTS

---------------------------------

def fetch_search_results(page: int) -> dict:
"""
Send a POST request to the search metadata endpoint
and return the parsed JSON response.
"""
payload = create_search_payload(page)
response = requests.request("POST", SEARCH_URL, headers=HEADERS, data=payload)
response.raise_for_status() # raises an exception if the request fails
return response.json()

---------------------------------

3. PROCESS SEARCH RESULTS

---------------------------------

def process_search_results(search_results: dict, assets: defaultdict) -> None:
"""
Parse the items in the search results and store them in the assets dict.
The key is the file stem (without suffix), and the value is a list of items.
"""
for item in search_results["assets"]["items"]:
original_file_name = Path(item["originalFileName"])
assets[original_file_name.stem].append(item)

---------------------------------

4a. HELPER: Check if a single asset is already stacked

---------------------------------

def is_asset_stacked(asset_id: str) -> bool:
"""
Perform a GET request on /api/assets/:id to determine if
that asset is already part of a stack.

Returns True if 'stack' is present (and not None) in the response.
"""
url = f"{ASSETS_URL}/{asset_id}"
response = requests.get(url, headers=HEADERS)
response.raise_for_status()
data = response.json()

# If the 'stack' key exists and is not None, the asset is stacked
return bool(data.get("stack"))

---------------------------------

4b. STACK IMAGES

---------------------------------

def stack_images(image: str, items: list) -> None:
"""
For each image group (stem), determine if it should be stacked.
1) Check if any item in the group is already stacked. If yes, skip.
2) Order/reverse items if needed based on suffix. To ensure the first item is a JPG, which will be the primary image in the immich stack
3) If the group meets the criteria, send a POST request to stack them.
"""
ids = [item["id"] for item in items]
name_suffixes = [Path(item["originalFileName"]).suffix.upper() for item in items]

# Skip stacking if any asset is already stacked
if any(is_asset_stacked(asset_id) for asset_id in ids):
    print(f"Skipping '{image}' because one or more assets are already stacked.")
    return

# If the first suffix is RAW_FILE_EXT, reverse the order
if name_suffixes and name_suffixes[0] == RAW_FILE_EXT:
    ids.reverse()
    name_suffixes.reverse()

# Ensure there's at least one .RAF if the group includes a .RAF
if RAW_FILE_EXT in name_suffixes:
    assert name_suffixes.count(RAW_FILE_EXT) >= 1

# Stack if more than one file and the first is .JPG
if len(name_suffixes) > 1 and name_suffixes[0] == ".JPG":
    payload = json.dumps({"assetIds": ids})
    response = requests.request("POST", STACKS_URL, headers=HEADERS, data=payload)
    print(f"{response.status_code}: {image} - Stacked {len(ids)} images")

---------------------------------

5. MAIN LOGIC

---------------------------------

def main():
assets = defaultdict(list)
page = 1

# Paginate until no nextPage
while True:
    search_results = fetch_search_results(page)
    items_on_page = search_results["assets"]["items"]
    print(f"Page {page} - Retrieved {len(items_on_page)} items")

    # Store items by grouping them by file stem
    process_search_results(search_results, assets)

    next_page = search_results["assets"]["nextPage"]
    page += 1
    if next_page is None:
        break

# Process each group to optionally stack images
for image, items in assets.items():
    stack_images(image, items)

Fediversum ägs inte av nåt storföretagsmonopol. Men till skillnad från andra sociala medier är det helt gratis att använda. Där finns ingen reklam, inga riskkapitalister som investerare, ingen spårning av användarna, inget kontrollerande företag och ett välkomnade och öppet samtalsklimat. Men där finns företag, aktieägare, reklam på en del servrar (instanser) och jobbiga människor samt programvaror som inte alltid är de mest användarvänliga.

Anders_S

2024-12-25 12:07:17

Fediversum ägs inte av nåt storföretagsmonopol. Men till skillnad från andra sociala medier är det helt gratis att använda. Där finns ingen reklam, inga riskkapitalister som investerare, ingen spårning av användarna, inget kontrollerande företag och ett välkomnade och öppet samtalsklimat. Men där finns företag, aktieägare, reklam på en del servrar (instanser) och jobbiga människor samt programvaror som inte alltid är de mest användarvänliga.

blog.zaramis.se/2024/12/25/fed…

diggita, nata nel 2007 ha recentemente abbracciato il fediverso scegliendo lemmy come progetto che gestirà da ora in futuro il proprio sito di giornalismo partecipativo.

Diggita era nata nel 2007 con il progetto open source pligg che però dopo varie vicissitudini e fork smise lo sviluppo circa circa 10 anni fa. Il sito è quindi stato sviluppato internamente dallo staff per i susseguenti 10 anni fino in questi gioni, quando si è scelto di dare un taglio netto al passato usando un progetto più moderno e siluppato da una community molto attiva.

Lemmy infatti non solo è un progetto maturo che viene usato da oltre 500 siti ma ha anche una decina di app sviluppate da terzi e sopratutto è un progetto che ha ricevuto 50mila euro di fondi europei per migliorare, rendendo quindi di fatto un ottimo progetto su cui basare il futuro diggita.

Diggita nasce cercando di rendere sempre più usato lemmy in italia e collaborando con lo staff di feddit.it la prima istanza italiana per renderlo sempre più una valida alternativa a reddit.