How do I write a file?

[74th]

I want to read a command output and write a new file.
What is good way to do it?

from invoke import task,Context,Result

@task
def replace(c):
    c:Context
    r:Result = c.run("cat file")
    t = r.stdout.replace("hoge","fuga")
    c.run(f'echo "{t}" > new_file')

from invoke import task,Context,Result

@task
def replace(c):
    c:Context
    r:Result = c.run("cat file")
    t = r.stdout.replace("hoge","fuga")
    with open("new_file", "w" ) as f:
        f.write(t)

If I use fabric2, I can use fabric.transfer.Transfer.put(). But must write a temporary file at local.
I think it's not good way.

[74th]

using heredocument

@task
def replace2(c):
    c: Context
    r: Result = c.run("cat file")
    t = r.stdout.replace("hoge", "fuga")
    c.run("cat << EOS > new_file\n" + t + "EOS")

[bitprophet]

Is there a reason you're trying to use shell commands for this instead of regular Python file I/O (e.g. open())? I can't really recommend a course of action without knowing the use case :)

[74th]

I'm translating shell scripts to tasks.py or fabfile.py.
My script has cat file | sed 's/__IP__/1.2.3.4/g' > new_fle.
I thinks I wants '> new_file' operator.
Of course, I'm able to use open(), but they do not work with with c.cd(dic) and fabric2, and I have to think current directory.

[74th]

I think I want to use pyinvoke functions instead of shell |, >, < operators. Is not it easy?

[74th]

I need #553, stdin closing. It can make it easier to pipe the standard input.

@task
def replace3(c):
    c: Context
    r: Result = c.run("cat file")
    t = r.stdout.replace("hoge", "fuga")
    tio = io.StringIO(t)
    c.run("cat > new_file", in_stream=tio)

[dperetti]

Trying something like this with no avail... It just stalls.

    with open(file, 'r') as f:
        c.run(f'kubectl exec -i {pod} -- pg_restore -d \'{database_url}\' --no-owner -n public',
                in_stream=f)