--no-scm option seems not working

[andrewBatutin]

dvc version - 0.71.0

pkg Mac install

can't get dvc stop ignoring data files

i init repo with

dvc init --no-scm

then

dvc add data

creates .gitignore with /data folder in it

is there something wrong i'm doing ?

[efiop]

Hi @andrewBatutin ! --no-scm only affects dvc init. After that dvc dynamically determines if you are using git or not and if you do - it also starts using it. Could you describe your scenario, please? 🙂

[andrewBatutin]

What i want to achieve is to get .dvc file for my data folder and keep data folder in project, so it doesn't fall under gitignore settings

And i thought --no-scm option during init won't ignore files and folders i add to dvc

[efiop]

@andrewBatutin Not sure I follow. What do you mean "keep data folder in project"? You mean dvc add it, while having some git-tracked files inside?

[andrewBatutin]

yes this case

i want to dvc add data and keep files inside data under git tracking

[efiop]

@andrewBatutin It is not possible, as dvc treats data as a whole directory. Maybe you could elaborate on your scenario, so we could better understand your use case and maybe suggest a better approach? 🙂

[andrewBatutin]

Actually i think i can use just git for my scenario) I want to easily track changes at precomputed word embedding files Save to file there SHA checksum every time they changed to keep track of them I think i can just use git + hooks to do that Thnx)

…

On Thu, 28 Nov 2019 at 16:10, Ruslan Kuprieiev ***@***.***> wrote: @andrewBatutin <https://github.com/andrewBatutin> It is not possible, as dvc treats data as a whole directory. Maybe you could elaborate on your scenario, so we could better understand your use case and maybe suggest a better approach? 🙂 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2859?email_source=notifications&email_token=ACCANJSUD4YDTHEWC4STV7LQV7GM7A5CNFSM4JSUFU42YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEFMWYKI#issuecomment-559508521>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACCANJR5DR3FBV2DXGAUQR3QV7GM7ANCNFSM4JSUFU4Q> .

[efiop]

I still don't quite understand your scenario 🙁 But happy that you've found a workaround. Please ping us if you will have any questions 🙂

[shcheklein]

@andrewBatutin I think what you need is to use non-cached data artifacts with DVC. It calculates checksums but does not put file into .gitignore. The command you would need is dvc run in this case:

dvc run -O data -f data.dvc

or you can use full power of it create DAG from multiple stages (all w/o caching or gitignoring files).

In this case you have git to version your actual files but also you have DVC to create .dvc files to capture checksums, etc. I see that it can be helpful if you want to use only pipelines for examples, w/o storing data with DVC>

[andrewBatutin]

Greate It seems what i want ) Спасибо)

…

On Thu, 28 Nov 2019 at 18:39, Ivan Shcheklein ***@***.***> wrote: @andrewBatutin <https://github.com/andrewBatutin> I think what you need is to use non-cached data artifacts with DVC. It calculates checksums but does not put file into .gitignore. The command you would need is dvc run in this case: dvc run -O data -f data.dvc or you can use full power of it create DAG from multiple stages (all w/o caching or gitignoring files). In this case you have git to version your actual files but also you have DVC to create .dvc files to capture checksums, etc. I see that it can be helpful if you want to use only pipelines for examples, w/o storing data with DVC> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2859?email_source=notifications&email_token=ACCANJQNMO2WVBJBIWCOZLLQV7X2LA5CNFSM4JSUFU42YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEFNCLJQ#issuecomment-559556006>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACCANJUSEWYW7FUZWIR6FQTQV7X2LANCNFSM4JSUFU4Q> .