mondei1 / docsort

Program to scan your documents, search for them and find them physically in your chaos.

whoami

My name is Nicolas and I would describe myself as a hobby full-stack developer (backend and frontend stuff).

• 😠 The MEAN-stack is my passion
• 🐧 Linux Mint [...] Pop!_OS Arch Fedora 🎩 and Windows 🪟
• 🔭 Currently working on "never getting finished"-projects
• 🎓 Still trying to get the hang out of C++ but I could never betray my beloved Rust
• 🍿 My favorite movies/series are Mr. Robot, Chappie, Rick and Morty, Final Space, Lackadaisy and Helluva Boss
• 🎮 My favorite games are Ori and the Will of the Wisps, Portal 2 and Minecraft
• 💬 You have a problem or a cool idea for a crypto project that will make a lot of money (please don't, I dare you) then don't hesitate to contact me:
  • If you want to E-Mail me, just look at my commits to get my E-Mail address and here's my GPG key (you know, bots and stuff)
  • Telegram: @mondei1
  • Discord: mondei1
  • Mastodon: @[email protected]
  • Threema: 53P48HNT

Hope this short summary could tell you a bit about me.

docsort's Issues

replace SYNC functions with ASYNC

Before relase we should replace most of the SYNC functions (fs.writeFileSync etc) with async functions (await).
If the original functions don't support this yet, i will just write small wrapper-functions for the used functions.

put title in filename

Putting the title into the filename (<filename>_<uid>_<primaryNumber>.>secondaryNumber>.<fileExtension>) offers the ability to make the document files itself human readable. This would be maybe a good idea if the storage is taken elsewhere without the software or on datacrash within the database.
But in doing this, the file must be renamed every time the user changes the title of a document. But from the expected use-case, this should be not often, if any.

Replace isNullOrUndefined()

isNullOrUndefined is deprecated (https://nodejs.org/api/util.html#util_util_isnullorundefined_object).
Instead we should use obj == null, which checks for null AND undefined.

utils in each separate file

For the future we should put the separate functions from lib/utils into their own files.
Makes it easier to read and easier to understand changes in git.

set "fileExtension" and "mimeType" by beforeSafe-hook

As the title says, we should set things, that are generated automatically, by hooks.

Decide on decent hash-iteration

For the final release we should decide on a decent but not over-extensive hash-iteration.

decide on dynamic ot static typed typescript

As the title says.

cleanup app.ts

Most functions in app.ts can/should be extracted into separate functions.

Backup technique

DocSort should offer the ability to backup all documents inclusive database into a cloud or on a external drive. There are some question open:

• Should be the backup encrypted?
• Which compression? (zip, gzip, xz, ..)
• Which file structure? Like this?

backup_24.02.2019.tar.gz
├── backup_meta.json
├── database.sqlite
└── documents
    ├── invoice_foobar_xxxxxxx.2.0.pdf
    └── reminder_to_pay_bill_xxxx.1.0.pdf

• Which meta data to save?
  • Creation time of the backup (unix timestamp)
  • More?
• Which cloud services we should support first?

License for DocSort

A very important point is to choose the right license for a project. Because this is a open source project we can choose GPLv2/3, MIT and Apache 2.0.

I personally prefer the GPLv3 license. Currently the project is unlicensed and this is not good for a open source project, this is why we have to resolve this issue till release.