hunch alternatives and similar packages
Based on the "Console" category.
Alternatively, view hunch alternatives based on common mentions on social networks and blogs.
9.3 3.7 hunch VS docoptA command-line interface description language and parser that will make you smile
7.1 0.0 hunch VS hflagsCommand line flag parser for Haskell, conceptually very similar to Google's gflags
Do you think we are missing an alternative of hunch or a related project?
Hunch is a command-line, system-agnostic application that allows quick creation of file system entries with a custom DSL resembling CSS selectors.
- CSS selectors to navigate the file system
- Custom numbering of duplicate entries
- External names placeholders to draw file names from external sources
- Template utilities: create files and directories from base templates instead of empty entries
- Simulation mode: preview the file system tree before creating entries
Basic usage, demonstrating automatic numbering of duplicate entries with the
Drawing names from external sources:
Hunch syntax is a (really) small subset of CSS selectors and emmet-inspired operators:
>stands for "go down" or "child". It means that the directory on the left side is the parent of the expression on the right side.
+stands for "stay on the same level" or "sibling". It means that the expression on the left side is on the same file system location than the expression on the right side.
*stands for "repeat" or "times". It means that the expression at one side of the operator is repeated as many times as the number on the other side.
=stands for "copy from" or "template". It means that the file or directory on the left side should be copied from the file or directory on the right side, instead of being created empty. An error is raised if a file template is used for a directory, and conversely.
_?NAMEstands for "take from" or "external source". It means that the name of the entry is not currently determined, but will be drawn from the external source "NAME" passed as argument.
Expressions can be grouped between parentheses. Logical constraints apply :
hunch '(foobar * 3) > file' will raise an error, as
a file system entry can't have more than one parent.
Hunch treats every entry as a file by default. If it should be a directory:
- It should have at least one child (
- ...its name should end with a '/' character.
This allows you to write most expressions without consideration for the "file or directory?" question. If an entry is a directory without any children, just add a '/' at the end of its name.
Absolute path dilemma
For now, you can't use absolute paths in Hunch syntax; an error will be raised before any processing. This behavior is intended.
Absolute path in the context provided by Hunch are awkward: should they be allowed everywhere and screw logic with constructs like
"/home/foo" + "/tmp" where Hunch operators are ignored ?
Should they only be allowed in the root position ? If so, should constructs like
"/home/foo" + baz be interpreted as creating "/home/foo" and "/home/foo/baz", or "/home/foo" and "./baz" ?
And what about duplicate absolute paths ?
As I didn't want to impose an arbitrary flawed view on Hunch syntax, the problem is, until a better solution arises, bypassed with the absolute paths interdiction
and a command-line option,
--root-dir. By specifying a root directory (default "."), all operations will be processed inside this directory. It is equivalent to
cd'ing into the directory before processing.
Numbering of duplicate entries uses a printf-like syntax. A given token (default
$) will be replaced in group in entry names.
The number of consecutive token characters indicates optional padding. If the length of the consecutive token characters is smaller than the length of the current number, the whole number will be displayed anyway.
file-$$$ * 150 will be replaced by "file-000", "file-001" to "file-150", while
file-$ * 150 will be replaced by "file-0", "file-1" to "file-150".
Replacement occurs at every occurrence, so
$$$-file-$$$ will be replaced by "000-file-000".
Input and options
Hunch takes two arguments: the input string to parse and a variable number of "key value" pairs representing external sources in the form "source_id source_value", where 'source_value' is a string of names separated by a given delimiter (default ',').
Hunch can be configured from command-line options:
--root-dir: The base directory in which all Hunch actions should be processed. Default: ".".
--templates-path: A directory containing template files and directories. Setting this option will allow you to reference template files without specifying the same base path over and over again. You can manually overrie this setting by using an absolute path in a template construct.
--delimiter: A delimiter character or string, used to split external names. Default: ",".
--number-token: A token character used as a placeholder for numbering replacement. Default "$".
--start-at: A number to start the numbering. Default: "0".
--override: Switch. If a file in input expression already exists, override it, otherwise do nothing.
--no-check: Switch. Do not run integrity tests on entries and template files (invalid names, template not found, ...). Switching on this option while not in simulation mode may result in runtime errors if entries or template files are ill-formed.
--verbose: Switch. Print debugging information.
--simulate: Switch. Print the preview of the resulting file system tree after parsing the given input, and exit.
As well as the traditional
Hunch is written in Haskell 2010. You will need a Haskell distribution (such as Haskell Platform) and a
build/package management tool like
cabal install hunch (
cabal update may be needed beforehand) or
stack install hunch.
Building from sources (Stack):
Building from sources (cabal):
cabal install --upgrade-dependencies --dependencies-only
- Test suite
MIT License. See file LICENSE.
*Note that all licence references and agreements mentioned in the hunch README section above are relevant to that project's source code only.