How to download a file from private repository using curl

[Adam Grzesiczak]

Hi,

Is there a way to download a file from a private repository using curl or wget?

I'm getting a 401 when hitting below URL. The app password got only read permissions to Repository and I wonder if this is enough or the URL is correct. I can't use the RAW file as the URL to RAW version is changing unless there is URL to always get the RAW version of the file?

curl -s -S --user username:apppassword -L -O https://bitbucket.org/<ORG_NAME>/<REPO>/src/master/<FOLDER>/file.txt

In a browser this will display the file and the website itself so it might not be a good way of doing it but I can't find a correct API URL.

Update: I found a correct URL for the latest RAW file but I always get 401 with curl/wget and an apppassword. What permissions app password needs to get this file?

[Adam Grzesiczak]

Ok found a "correct" way to do it. "Correct" as you shouldn't use passwords in a bash like that.

curl -s -S --user username:apppassword -L -O https://api.bitbucket.org/2.0/repositories/<ORGANISATION>/<REPO>/src/master/<FOLDER>/<FILE>

[aloraine]

Thanks! I have the same question. This worked for me, as well. 

[Adam Grzesiczak]

One way to import passwords to a bash script is from an external file.

script.sh

#!/bin/bash

source passwords.txt

curl -s -S --user ${USER}:${APP_PASSWORD} -L -O https://api.bitbucket.org/2.0/repositories/<ORGANISATION>/<REPO>/src/master/<FOLDER>/<FILE>

passwords.txt

USER="username"

APP_PASSWORD="password"

[Charlie Simon]

Just letting anyone who's curious. Don't use api.bitbucket.org, at least every time I did I was just getting an error. Instead, I've opted to just copy the link to the download from the "Download's" Tab in the repository. Below is my example. 

curl -H "Cache-Control: no-cache" --user <Username>:<Password> -L -O https://bitbucket.org/<orgname>/<Repository>/get/<filename>.zip

You'll need to add -H "Cache-Control: no-cache" don't ask me why but setting a custom header seems to solve the problem.

I also removed the silence, just because I like to know I'm not just sitting on a dead page.

And if anyone is looking to streamline this. I'm using this as a deployment system via an install.sh script built into a custom ISO with a redirect link to a custom domain that points to the BitBucket download link. Meaning all I need to do is update my URL redirect. I've put below my .sh code in case anyone is curious. It's a neat little trick to save some time in the future.  

Obviously, this will also work within the normal terminal. 

url=$(curl -Ls -o /dev/null -w %{url_effective} <redirect url>)

echo $url

curl -H "Cache-Control: no-cache" --user <Username>:<Password>  -L -O $url

[RogerThomas]

As this thread comes up when trying to research this issue at the end of 2021, here is and example line that works for me, when trying to pull down a file in a public repository

curl -s -L https://api.bitbucket.org/2.0/repositories/{workspace)/{repo_name}/src/main/bin/OneLineRoot

This pulls down the file /bin/OneLineRoot within the main branch, from repository {repo_name} within the {workspace}.

Bitbucket's interface needs to be a lot more exact than say the one presented by github.