Ich frage die #bahn #bahnBubble und #linux Bubble nach neuen Erkenntnissen: Mein nettes kleines Reise-Notebook (Manjaro Linux), auf dem auch #docker wegen #paperless_ngx läuft, bekommt im #ice kein WLAN. Ich hatte vor 1/2 Jahr komplett Docker deaktiviert, das hat aber auch nichts genützt.
Man findet dazu auch Threads im Internet - hatte irgend ein Linux-Nutzer in letzter Zeit WLAN-Probleme im ICE und konnte diese lösen?
Leute, wie geil ist denn bitte #paperless_ngx?
Indem ich ein kleines Label (s. Foto) auf ein Dokument klebe, wird automatisch eine Seriennnummer vergeben, mit der man ganz leicht digitales und physisches Dokument im Ordner wiederfindet. Zusätzlich wird das Label als Dokumententrenner beim Import verwendet. D.h.: Label jeweils auf die erste Seite eines (mehrseitigen) Dokuments, alles auf einen Stapel, der am Stück gescannt wird und #paperless teilt den Stapel in einzelne Dokumente. Awesome!
#paperless_ngx: Nachtrag zur Session gestern. Heinz hatte das Buch "Digital Cleaning" von @_DigitalWriter_ empfohlen. Dem Autor könnt ihr auch hier folgen.
Going live right now! Building out a #paperless_ngx dashboard using #FSharp and #FalcoFramework! DOing some mysterious, #TTRPG-themed additions tonight!
Wanna know what they are?
See you on-air!
EDIT: We had some technical difficulties. . .again. We're back live!
Tonight at roughly 9 PM Eastern, we're going to build some more views for my #paperless_ngx dashboard using #FSharp and Falco!
I'm thinking we lean into the theme of the _contents_ of my documents, which are almost entirely #TTRPG resources.
What do I mean by that? Come and find out!
How to design an API which manages documents, especially ones which may be very long:
1. Do NOT include the entire OCR'd text for the document in your response when someone asks for a list of documents. That would be insane and lead to very long load times.
2. DO. . .see 1.
/facepalm
Tonight at (roughly) 8:30 PM Eastern, I'll be streaming some #FSharp development!
We'll be fleshing out my #paperless_ngx dashboard application.
Edit: running late putting kiddo to bed. Should be live by 9!
EDIT: We're live! Just before 9, heck yeah!
One thing in trying to figure out.... Why does #paperless_ngx use so much disk space on its "application" filesystem? I only have roughly 20GB of content, but the root filesystem for the container has 70GB of ocrmypdf crud in the /tmp directory.
Auf das automatische Drehen von Dokumenten in #paperless_ngx kann man sich scheinbar nicht verlassen -.-
Aber man kann es halt nachher manuell drehen und dann nochmal durch die OCR jagen, dann passt alles...
Im Log sieht man, dass es zwar erkannt wurde, dass die Seite falschherum ist, aber das ist angeblich nicht wichtig genug...
[ocrmypdf._pipeline] page is facing ⇩, confidence 7.94 - confidence too low to rotate
Not a bad night for #FSharp and #paperless_ngx!
Thanks #Falco for the awesome DSL!
Tonight, I'll be streaming some #FSharp development! I'm going to develop a dashboard for #paperless_ngx in #Falco, an F#-native libary for building #AspNet services.
Hop in around 8:30 PM Eastern: https://twitch.tv/b4ux1t3
(update: running a few minutes late, kiddos having trouble tonight)
(update2: Apparently when I upgraded mint. . .it ditched the proprietary nvidia driver, meaning I can't do hardware encoding. . .I'm fixing this right now, I'll be streaming momentarily)
FINAL UPDATE: We're LIVE!!!!!
Woke up to a mess of server errors.
Turns out, the drive had filled itself up while processing everything last night. Got everything moved to a new disk, brought it back up, and we're off to the races. We'll have to recover some stuff, it looks like, but I gotta say: I'm impressed with #Paperless_NGX #PaperlessNGX and its ability to recover from an albeit very common failure case, which is nonetheless not often handled well.
Tonight's stream was massively successful. I got all 18 GB of files recovered and renamed (less a few that weren't properly renamed).
We're at 130/4000 items processed... Looks like I gotta let it run for a while. 
Is anyone familiar enough with #PaperlessNGX #paperless_ngx to tell me if I can find a dashboard for the consumption directory? Even if it's just, like, a celery dashboard, I just wanna get am idea for what it's doing and where.
#
