pyinfra-dev
operation for managing an annotated section inside a file
Is your feature request related to a problem? Please describe
there are cases where a user needs to control most of the contents of a file outside of pyinfra, but still have pyinfra manage a particular block in the file, for example:
<some content maintained by the user or other tool>
### BEGIN CONTROL BLOCK ###
<data from pyinfra goes here>
<should support multiple lines>
### END CONTROL BLOCK ###
<other content maintained by user or other tool>
everything between the ### control blocks should be replaced.
Describe the solution you'd like
a new operation, perhaps called file.insert, which could be invoked as follows:
file.insert(
path="/path/to/file",
match_begin="### BEGIN CONTROL BLOCK ###",
match_end = "### END CONTROL BLOCK ###",
content = "something interesting\nand something else",
)
Future Scope
something to consider is inlined cases, where the block start and block end do not have newlines surrounding them. maybe that should be handled by a separate operation?
for example:
my_variable = "<content>"
file.insert(
path="/path/to/file",
match_begin="my_variable = \"",
match_end="\"",
content="some value",
inline=True,
)
other potential parameters for this operation:
- match_greedy to match until the last instance of
match_end - match_index to match blocks other than the first one
- match_all to match all blocks instead of just the first
actually, it looks like this is present in git but not released (or not in documentation) yet:
https://github.com/Fizzadar/pyinfra/blob/2.x/pyinfra/facts/files.py#L359 https://github.com/Fizzadar/pyinfra/blob/2.x/pyinfra/operations/files.py#L1607
however, the documentation for files.block is written as files.marked_block, which should be fixed.