File edition helpers working on top of mem-fs
import { create as createMemFs } from 'mem-fs';
import { create as createEditor } from 'mem-fs-editor';
const store = createMemFs();
const fs = createEditor(store);
fs.write('somefile.js', 'var a = 1;');
await fs.commit()
Read a file and return its contents as a string.
You can alternatively get the raw contents buffer if you pass options.raw = true
.
By default, calling read()
on a file path that does not exist throws error. You can, however, pass options.defaults = 'your default content'
to get a default content you pass in, if you prefer to not deal with try/catch.
Read a file and parse its contents as JSON.
readJSON()
internally calls read()
but will not throw an error if the file path you pass in does not exist. If you pass in an optional defaults
, the defaults
content will be returned in case of the target file is missing, instead of undefined
. (Error would still be thrown if JSON.parse
failed to parse your target file.)
Replace the content of a file (existing or new) with a string or a buffer.
Replace the content of a file (existing or new) with an object that is to be converted by calling JSON.stringify()
.
contents
should usually be a JSON object, but it can technically be anything that is acceptable by JSON.stringify.
Optionally pass replacer
and space
as the last two arguments, as defined by JSON.stringify. spacer
is used to format the output string (prettify).
Default value for space
is 2
, when not specified.
Append the new contents to the current file contents.
options.trimEnd
(defaulttrue
). Trim trailing whitespace of the current file contents.options.separator
(defaultos.EOL
). Separator to insert between current and new contents.options.create
(defaultfalse
). Create the file if doesn't exists.
Append the new contents
to the exsting filepath
content and parse the new contents as an ejs template where context
is the template context (the variable names available inside the template).
options.trimEnd
(defaulttrue
). Trim trailing whitespace of the current file contents.options.separator
(defaultos.EOL
). Separator to insert between current and new contents.
Extend the content of an existing JSON file with the partial objects provided as argument.
Optionally take the same JSON formatting arguments than #writeJSON()
.
Delete a file or a directory.
filePath
can also be a glob
. If filePath
is glob, you can optionally pass in an options.globOptions
object to change its pattern matching behavior. The full list of options are being described here. The sync
flag is forced to be true
in globOptions
.
Copy file(s) from the from
path to the to
path.
When passing array, you should pass options.fromBasePath
to be used to calculate the to
relative path. The common directory will be detected and used as fromBasePath
otherwise.
Optionally, pass an options.process
function (process(contents)
) returning a string or a buffer who'll become the new file content. The process function will take a single contents argument who is the copied file contents as a Buffer
.
options.ignoreNoMatch
can be used to silence the error thrown if no files match the from
pattern.
options.append
can be used to append from
contents to to
instead of copying, when the file is already loaded in mem-fs (safe for regeneration).
from
can be a glob pattern that'll be match against the file system. If that's the case, then to
must be an output directory. For a globified from
, you can optionally pass in an options.globOptions
object to change its pattern matching behavior. The full list of options are being described here. The nodir
flag is forced to be true
in globOptions
to ensure a vinyl object representing each matching directory is marked as deleted
in the mem-fs
store.
Optionally, when from
is a glob pattern, pass an options.processDestinationPath
function (processDestinationPath(destinationFile)
) returning a string who'll become the new file name.
options.noGlob
can be used to by bypass glob matching entirely. In that case, from
will directly match file paths against the file system.
Async version of copy
.
copy
loads from
to memory and copy its contents to to
.
copyAsync
copies directly from the disk to to
, falling back to copy
behavior if the file doesn't exists on disk.
Same parameters of copy
(see copy() documentation for more details).
Copy the from
file and, if it is not a binary file, parse its content as an ejs template where context
is the template context (the variable names available inside the template).
You can optionally pass a templateOptions
object. mem-fs-editor
automatically setup the filename option so you can easily use partials.
You can also optionally pass a copyOptions
object (see copy() documentation for more details).
Templates syntax looks like this:
<%= value %>
<%- include('partial.ejs', { name: 'Simon' }) %>
Dir syntax looks like this:
/some/path/dir<%= value %>/...
Refer to the ejs documentation for more details.
Async version of copyTpl
that uses copyAsync
instead of copy
.
Can be used for best performance. Reduces overheads.
Same parameters of copyTpl
(see copyTpl() documentation for more details).
Move/rename a file from the from
path to the to
path.
#move
internally uses #copy
and #delete
, so from
can be a glob pattern, and you can provide options.globOptions
with it.
Returns true
if a file exists. Returns false
if the file is not found or deleted.
Pass stored files through a pipeline and persist every changes made to files in the mem-fs store to disk.
If provided, options
is the pipeline options.
By default only modified files are passed through the pipeline.
Pass a custom filter options.filter
to customize files passed through the pipeline.
If provided, ...transforms
is a vararg of TransformStream to be applied on a stream of vinyl files (like gulp plugins).
commitTransform
is appended to transforms
and persists modified files to disk, non modified files are passed through.
returns promise that is resolved once the pipeline is finished.
Dump files to compare expected result.
Provide a cwd
for relative path. Allows to omit temporary path.
Provide a filter
function or a pattern to focus on specific files.
dump
returns only modified (committed or not) files when no filter or a pattern is provided.