get MIME types for source files #3

Closed hoijui closed 1 year ago

In GitLab by @moedn on Sep 15, 2021, 15:20

based on the discussion here

we need a CSV file with the following columns:

file extension
open/closed format?
binary or plain text?

→ for all (popular) CAD formats; which is everything in [this]() list plus:

FreeCAD
openSCAD
QCAD
LibreCAD
CadQuery

In GitLab by @moedn on Sep 21, 2021, 11:59

File is here: https://gitlab.com/OSEGermany/osh-tool/-/blob/b9228fe2666b8e7ad5b9e944c12ea7f0d1ef0dbf/Supporting%20docs/Extension_formats_list.csv

In GitLab by @moedn on Sep 21, 2021, 12:30

@damamitesh any help needed regarding format types? (open/proprietary)

In GitLab by @damamitesh on Sep 21, 2021, 13:00

Hi @moedn I have uploaded the csv file for the same here There are some unknown cells. Updated the same to Robin

In GitLab by @amerezoji1362 on Sep 22, 2021, 11:40

Inkscape is one of the popular open-source software that the makers use it. It can add to the list of CAD formats

In GitLab by @amerezoji1362 on Sep 22, 2021, 14:02

This link will be useful for the identification of file formats.

https://transmagic.com/cad-formats/

In GitLab by @moedn on Sep 22, 2021, 18:25

:white_check_mark: svg is already on the list

In GitLab by @moedn on Sep 22, 2021, 18:26

jepp indeed; it was already mentioned in the original discussion (linked above in the issue description)

In GitLab by @moedn on Sep 22, 2021, 18:42

@damamitesh reviewed the file, fixed a few things and continued working on it. looks already pretty neat, thanks!

In GitLab by @moedn on Sep 22, 2021, 18:43

I found that a few formats exist both, in binary and plain text form (e.g. dxf or stl). People will most likely share the binary format since it's smaller. However, @hoijui we could still notify people in the log that those files also exist in plain text form which would make them git compatible

I did something like that in an earlier bash based version of a similar tool, for libre-office and ms office files. would be to have a table of such formats too, with header bin-ext,text-ext,format-name, with data like doc,docx,MS Office Word document

In GitLab by @moedn on Nov 11, 2021, 17:59

@hoijui here some ideas how to use these MIME types in the OSH tool:

when export files are found, but no source files, returns a message (no error): "we found export files (STEP, STL ) in this repository, but no source files → replicating an OSH project is nice, but please consider adding the source files in their native data format to the repository so others can effectively contribute in the development of your project"
when export AND source files are found, returns a message (no error): "we found both, export (STEP, STL,…) AND source files (FCStd,…) in this repository → please consider cleaning up your repository, removing those export files and add them to a release instead (or have them automatically created by some sophisticated CI script"
when neither one nor the other is found, returns a message (no error): "we couldn't identify any OSH-related source (or even export) files in this repository → are you sure, this is an OSH project?"

very good, will do that! :-)

In GitLab by @moedn on Nov 15, 2021, 17:53

I think, now I remember @hoijui :) when the OSH tool can build manifest files, it could automatically guess components by the extensions of their source files. I wrote a little pseudo code for this in Alec's Krawler repo (which is private now, so I cannot access the issue). If that's interesting for the OSH tool, I could just ask him to grant me permission :)

This is now implemented. Not exactly, logically like this, but .. basically the same info is portrayed.

This is now tracked in #4.

hoijui / osh-tool

get MIME types for source files #3