sqlglot.helper

  1from __future__ import annotations
  2
  3import inspect
  4import logging
  5import re
  6import sys
  7import typing as t
  8from collections.abc import Collection
  9from contextlib import contextmanager
 10from copy import copy
 11from enum import Enum
 12
 13if t.TYPE_CHECKING:
 14    from sqlglot import exp
 15    from sqlglot.expressions import Expression
 16
 17    T = t.TypeVar("T")
 18    E = t.TypeVar("E", bound=Expression)
 19
 20CAMEL_CASE_PATTERN = re.compile("(?<!^)(?=[A-Z])")
 21PYTHON_VERSION = sys.version_info[:2]
 22logger = logging.getLogger("sqlglot")
 23
 24
 25class AutoName(Enum):
 26    """This is used for creating enum classes where `auto()` is the string form of the corresponding value's name."""
 27
 28    def _generate_next_value_(name, _start, _count, _last_values):  # type: ignore
 29        return name
 30
 31
 32def seq_get(seq: t.Sequence[T], index: int) -> t.Optional[T]:
 33    """Returns the value in `seq` at position `index`, or `None` if `index` is out of bounds."""
 34    try:
 35        return seq[index]
 36    except IndexError:
 37        return None
 38
 39
 40@t.overload
 41def ensure_list(value: t.Collection[T]) -> t.List[T]:
 42    ...
 43
 44
 45@t.overload
 46def ensure_list(value: T) -> t.List[T]:
 47    ...
 48
 49
 50def ensure_list(value):
 51    """
 52    Ensures that a value is a list, otherwise casts or wraps it into one.
 53
 54    Args:
 55        value: the value of interest.
 56
 57    Returns:
 58        The value cast as a list if it's a list or a tuple, or else the value wrapped in a list.
 59    """
 60    if value is None:
 61        return []
 62    elif isinstance(value, (list, tuple)):
 63        return list(value)
 64
 65    return [value]
 66
 67
 68@t.overload
 69def ensure_collection(value: t.Collection[T]) -> t.Collection[T]:
 70    ...
 71
 72
 73@t.overload
 74def ensure_collection(value: T) -> t.Collection[T]:
 75    ...
 76
 77
 78def ensure_collection(value):
 79    """
 80    Ensures that a value is a collection (excluding `str` and `bytes`), otherwise wraps it into a list.
 81
 82    Args:
 83        value: the value of interest.
 84
 85    Returns:
 86        The value if it's a collection, or else the value wrapped in a list.
 87    """
 88    if value is None:
 89        return []
 90    return (
 91        value if isinstance(value, Collection) and not isinstance(value, (str, bytes)) else [value]
 92    )
 93
 94
 95def csv(*args, sep: str = ", ") -> str:
 96    """
 97    Formats any number of string arguments as CSV.
 98
 99    Args:
100        args: the string arguments to format.
101        sep: the argument separator.
102
103    Returns:
104        The arguments formatted as a CSV string.
105    """
106    return sep.join(arg for arg in args if arg)
107
108
109def subclasses(
110    module_name: str,
111    classes: t.Type | t.Tuple[t.Type, ...],
112    exclude: t.Type | t.Tuple[t.Type, ...] = (),
113) -> t.List[t.Type]:
114    """
115    Returns all subclasses for a collection of classes, possibly excluding some of them.
116
117    Args:
118        module_name: the name of the module to search for subclasses in.
119        classes: class(es) we want to find the subclasses of.
120        exclude: class(es) we want to exclude from the returned list.
121
122    Returns:
123        The target subclasses.
124    """
125    return [
126        obj
127        for _, obj in inspect.getmembers(
128            sys.modules[module_name],
129            lambda obj: inspect.isclass(obj) and issubclass(obj, classes) and obj not in exclude,
130        )
131    ]
132
133
134def apply_index_offset(expressions: t.List[t.Optional[E]], offset: int) -> t.List[t.Optional[E]]:
135    """
136    Applies an offset to a given integer literal expression.
137
138    Args:
139        expressions: the expression the offset will be applied to, wrapped in a list.
140        offset: the offset that will be applied.
141
142    Returns:
143        The original expression with the offset applied to it, wrapped in a list. If the provided
144        `expressions` argument contains more than one expressions, it's returned unaffected.
145    """
146    if not offset or len(expressions) != 1:
147        return expressions
148
149    expression = expressions[0]
150
151    if expression and expression.is_int:
152        expression = expression.copy()
153        logger.warning("Applying array index offset (%s)", offset)
154        expression.args["this"] = str(int(expression.this) + offset)  # type: ignore
155        return [expression]
156
157    return expressions
158
159
160def camel_to_snake_case(name: str) -> str:
161    """Converts `name` from camelCase to snake_case and returns the result."""
162    return CAMEL_CASE_PATTERN.sub("_", name).upper()
163
164
165def while_changing(
166    expression: t.Optional[Expression], func: t.Callable[[t.Optional[Expression]], E]
167) -> E:
168    """
169    Applies a transformation to a given expression until a fix point is reached.
170
171    Args:
172        expression: the expression to be transformed.
173        func: the transformation to be applied.
174
175    Returns:
176        The transformed expression.
177    """
178    while True:
179        start = hash(expression)
180        expression = func(expression)
181        if start == hash(expression):
182            break
183    return expression
184
185
186def tsort(dag: t.Dict[T, t.List[T]]) -> t.List[T]:
187    """
188    Sorts a given directed acyclic graph in topological order.
189
190    Args:
191        dag: the graph to be sorted.
192
193    Returns:
194        A list that contains all of the graph's nodes in topological order.
195    """
196    result = []
197
198    def visit(node: T, visited: t.Set[T]) -> None:
199        if node in result:
200            return
201        if node in visited:
202            raise ValueError("Cycle error")
203
204        visited.add(node)
205
206        for dep in dag.get(node, []):
207            visit(dep, visited)
208
209        visited.remove(node)
210        result.append(node)
211
212    for node in dag:
213        visit(node, set())
214
215    return result
216
217
218def open_file(file_name: str) -> t.TextIO:
219    """Open a file that may be compressed as gzip and return it in universal newline mode."""
220    with open(file_name, "rb") as f:
221        gzipped = f.read(2) == b"\x1f\x8b"
222
223    if gzipped:
224        import gzip
225
226        return gzip.open(file_name, "rt", newline="")
227
228    return open(file_name, encoding="utf-8", newline="")
229
230
231@contextmanager
232def csv_reader(read_csv: exp.ReadCSV) -> t.Any:
233    """
234    Returns a csv reader given the expression `READ_CSV(name, ['delimiter', '|', ...])`.
235
236    Args:
237        read_csv: a `ReadCSV` function call
238
239    Yields:
240        A python csv reader.
241    """
242    args = read_csv.expressions
243    file = open_file(read_csv.name)
244
245    delimiter = ","
246    args = iter(arg.name for arg in args)
247    for k, v in zip(args, args):
248        if k == "delimiter":
249            delimiter = v
250
251    try:
252        import csv as csv_
253
254        yield csv_.reader(file, delimiter=delimiter)
255    finally:
256        file.close()
257
258
259def find_new_name(taken: t.Collection[str], base: str) -> str:
260    """
261    Searches for a new name.
262
263    Args:
264        taken: a collection of taken names.
265        base: base name to alter.
266
267    Returns:
268        The new, available name.
269    """
270    if base not in taken:
271        return base
272
273    i = 2
274    new = f"{base}_{i}"
275    while new in taken:
276        i += 1
277        new = f"{base}_{i}"
278
279    return new
280
281
282def object_to_dict(obj: t.Any, **kwargs) -> t.Dict:
283    """Returns a dictionary created from an object's attributes."""
284    return {**{k: copy(v) for k, v in vars(obj).copy().items()}, **kwargs}
285
286
287def split_num_words(
288    value: str, sep: str, min_num_words: int, fill_from_start: bool = True
289) -> t.List[t.Optional[str]]:
290    """
291    Perform a split on a value and return N words as a result with `None` used for words that don't exist.
292
293    Args:
294        value: the value to be split.
295        sep: the value to use to split on.
296        min_num_words: the minimum number of words that are going to be in the result.
297        fill_from_start: indicates that if `None` values should be inserted at the start or end of the list.
298
299    Examples:
300        >>> split_num_words("db.table", ".", 3)
301        [None, 'db', 'table']
302        >>> split_num_words("db.table", ".", 3, fill_from_start=False)
303        ['db', 'table', None]
304        >>> split_num_words("db.table", ".", 1)
305        ['db', 'table']
306
307    Returns:
308        The list of words returned by `split`, possibly augmented by a number of `None` values.
309    """
310    words = value.split(sep)
311    if fill_from_start:
312        return [None] * (min_num_words - len(words)) + words
313    return words + [None] * (min_num_words - len(words))
314
315
316def is_iterable(value: t.Any) -> bool:
317    """
318    Checks if the value is an iterable, excluding the types `str` and `bytes`.
319
320    Examples:
321        >>> is_iterable([1,2])
322        True
323        >>> is_iterable("test")
324        False
325
326    Args:
327        value: the value to check if it is an iterable.
328
329    Returns:
330        A `bool` value indicating if it is an iterable.
331    """
332    return hasattr(value, "__iter__") and not isinstance(value, (str, bytes))
333
334
335def flatten(values: t.Iterable[t.Iterable[t.Any] | t.Any]) -> t.Iterator[t.Any]:
336    """
337    Flattens an iterable that can contain both iterable and non-iterable elements. Objects of
338    type `str` and `bytes` are not regarded as iterables.
339
340    Examples:
341        >>> list(flatten([[1, 2], 3, {4}, (5, "bla")]))
342        [1, 2, 3, 4, 5, 'bla']
343        >>> list(flatten([1, 2, 3]))
344        [1, 2, 3]
345
346    Args:
347        values: the value to be flattened.
348
349    Yields:
350        Non-iterable elements in `values`.
351    """
352    for value in values:
353        if is_iterable(value):
354            yield from flatten(value)
355        else:
356            yield value
357
358
359def count_params(function: t.Callable) -> int:
360    """
361    Returns the number of formal parameters expected by a function, without counting "self"
362    and "cls", in case of instance and class methods, respectively.
363    """
364    count = function.__code__.co_argcount
365    return count - 1 if inspect.ismethod(function) else count
366
367
368def dict_depth(d: t.Dict) -> int:
369    """
370    Get the nesting depth of a dictionary.
371
372    For example:
373        >>> dict_depth(None)
374        0
375        >>> dict_depth({})
376        1
377        >>> dict_depth({"a": "b"})
378        1
379        >>> dict_depth({"a": {}})
380        2
381        >>> dict_depth({"a": {"b": {}}})
382        3
383
384    Args:
385        d (dict): dictionary
386
387    Returns:
388        int: depth
389    """
390    try:
391        return 1 + dict_depth(next(iter(d.values())))
392    except AttributeError:
393        # d doesn't have attribute "values"
394        return 0
395    except StopIteration:
396        # d.values() returns an empty sequence
397        return 1
398
399
400def first(it: t.Iterable[T]) -> T:
401    """Returns the first element from an iterable.
402
403    Useful for sets.
404    """
405    return next(i for i in it)
406
407
408def should_identify(text: str, identify: str | bool) -> bool:
409    """Checks if text should be identified given an identify option.
410
411    Args:
412        text: the text to check.
413        identify: "always" | True - always returns true, "safe" - true if no upper case
414
415    Returns:
416        Whether or not a string should be identified.
417    """
418    if identify is True or identify == "always":
419        return True
420    if identify == "safe":
421        return not any(char.isupper() for char in text)
422    return False