path: root/bin/shyaml

#!/usr/bin/env python

## Note: to launch test, you can use:
##   python -m doctest -d shyaml.py
## or
##   nosetests

from __future__ import print_function

import sys
import yaml
import os.path
import re

EXNAME = os.path.basename(sys.argv[0])


def tokenize(s):
    r"""Returns an iterable in all subpart of a '.' separated string

    So:

        >>> list(tokenize('foo.bar.wiz'))
        ['foo', 'bar', 'wiz']

    this function has to deal with any type of data in the string. So it
    actually interprets the string. Characters with meaning are '.' and '\'.
    Both of these can be included in a token by quoting them with '\'.

    So dot of slashes can be contained in token:

        >>> print('\n'.join(tokenize(r'foo.dot<\.>.slash<\\>')))
        foo
        dot<.>
        slash<\>

    Notice that empty keys are also supported:

        >>> list(tokenize(r'foo..bar'))
        ['foo', '', 'bar']

    Given an empty string:

        >>> list(tokenize(r''))
        ['']

    And a None value:

        >>> list(tokenize(None))
        []

    """
    if s is None:
        raise StopIteration
    tokens = (re.sub(r'\\(\\|\.)', r'\1', m.group(0))
              for m in re.finditer(r'((\\.|[^.\\])*)', s))
    ## an empty string superfluous token is added after all non-empty string token:
    for token in tokens:
        if len(token) != 0:
            next(tokens)
        yield token


def mget(dct, key, default=None):
    r"""Allow to get values deep in a dict with doted keys

    Accessing leaf values is quite straightforward:

        >>> dct = {'a': {'x': 1, 'b': {'c': 2}}}
        >>> mget(dct, 'a.x')
        1
        >>> mget(dct, 'a.b.c')
        2

    But you can also get subdict if your key is not targeting a
    leaf value:

        >>> mget(dct, 'a.b')
        {'c': 2}

    As a special feature, list access is also supported by providing a
    (possibily signed) integer, it'll be interpreted as usual python
    sequence access using bracket notation:

        >>> mget({'a': {'x': [1, 5], 'b': {'c': 2}}}, 'a.x.-1')
        5
        >>> mget({'a': {'x': 1, 'b': [{'c': 2}]}}, 'a.b.0.c')
        2

    Keys that contains '.' can be accessed by escaping them:

        >>> dct = {'a': {'x': 1}, 'a.x': 3, 'a.y': 4}
        >>> mget(dct, 'a.x')
        1
        >>> mget(dct, r'a\.x')
        3
        >>> mget(dct, r'a.y')
        >>> mget(dct, r'a\.y')
        4

    As a consequence, if your key contains a '\', you should also escape it:

        >>> dct = {r'a\x': 3, r'a\.x': 4, 'a.x': 5, 'a\\': {'x': 6}}
        >>> mget(dct, r'a\\x')
        3
        >>> mget(dct, r'a\\\.x')
        4
        >>> mget(dct, r'a\\.x')
        6
        >>> mget({'a\\': {'b': 1}}, r'a\\.b')
        1
        >>> mget({r'a.b\.c': 1}, r'a\.b\\\.c')
        1

    And even empty strings key are supported:

        >>> dct = {r'a': {'': {'y': 3}, 'y': 4}, 'b': {'': {'': 1}}, '': 2}
        >>> mget(dct, r'a..y')
        3
        >>> mget(dct, r'a.y')
        4
        >>> mget(dct, r'')
        2
        >>> mget(dct, r'b..')
        1

    mget support also default value if the key is not found:

        >>> mget({'a': 1}, 'b.y', default='N/A')
        'N/A'

    but will complain if you are trying to get into a leaf:

        >>> mget({'a': 1}, 'a.y', default='N/A')  # doctest: +ELLIPSIS
        Traceback (most recent call last):
        ...
        TypeError: 'int' object ...

    if the key is None, the whole dct should be sent back:

        >>> mget({'a': 1}, None)
        {'a': 1}

    """
    return aget(dct, tokenize(key), default)


def aget(dct, key, default=None):
    r"""Allow to get values deep in a dict with iterable keys

    Accessing leaf values is quite straightforward:

        >>> dct = {'a': {'x': 1, 'b': {'c': 2}}}
        >>> aget(dct, ('a', 'x'))
        1
        >>> aget(dct, ('a', 'b', 'c'))
        2

    If key is empty, it returns unchanged the ``dct`` value.

        >>> aget({'x': 1}, ())
        {'x': 1}

    """
    key = iter(key)
    try:
        head = next(key)
    except StopIteration:
        return dct
    try:
        value = dct[int(head)] if isinstance(dct, list) else dct[head]
    except KeyError:
        return default
    return aget(value, key, default)


def stderr(msg):
    sys.stderr.write(msg + "\n")


def die(msg, errlvl=1, prefix="Error: "):
    stderr("%s%s" % (prefix, msg))
    sys.exit(errlvl)

SIMPLE_TYPES = (str, int, float)
COMPLEX_TYPES = (list, dict)


def dump(value):
    return value if isinstance(value, SIMPLE_TYPES) \
      else yaml.dump(value, default_flow_style=False)

def type_name(value):
    """Returns pseudo-YAML type name of given value."""
    return "struct" if isinstance(value, dict) else \
          "sequence" if isinstance(value, (tuple, list)) else \
          type(value).__name__

def stdout(value):
    sys.stdout.write(value)

def main(args):
    usage = """usage:
    %(exname)s {get-value{,-0},get-type,keys{,-0},values{,-0}} KEY DEFAULT
    """ % {"exname": EXNAME}
    if len(args) == 0:
        die(usage, errlvl=0, prefix="")
    action = args[0]
    key_value = None if len(args) == 1 else args[1]
    default = args[2] if len(args) > 2 else ""
    contents = yaml.load(sys.stdin)
    try:
        value = mget(contents, key_value, default)
    except IndexError:
        die("list index error in path %r." % key_value)
    except (KeyError, TypeError):
        die("invalid path %r." % key_value)

    tvalue = type_name(value)
    termination = "\0" if action.endswith("-0") else "\n"

    if action == "get-value":
        print(dump(value), end='')
    elif action in ("get-values", "get-values-0"):
        if isinstance(value, dict):
            for k, v in value.iteritems():
                stdout("%s%s%s%s" % (dump(k), termination,
                                     dump(v), termination))
        elif isinstance(value, list):
            for l in value:
                stdout("%s%s" % (dump(l), termination))
        else:
            die("%s does not support %r type. "
                "Please provide or select a sequence or struct."
                % (action, tvalue))
    elif action == "get-type":
        print(tvalue)
    elif action in ("keys", "keys-0", "values", "values-0"):
        if isinstance(value, dict):
            method = value.keys if action.startswith("keys") else value.values
            for k in method():
                stdout("%s%s" % (dump(k), termination))
        else:
            die("%s does not support %r type. "
                "Please provide or select a struct." % (action, tvalue))
    else:
        die("Invalid argument.\n%s" % usage)


if __name__ == "__main__":
    sys.exit(main(sys.argv[1:]))