Support for pythonic sample matching in views

[voxel51]

Support for pythonic sample matching in views

[brimoor]

Here's an implementation of the parser for python-like string matches:

import json

# https://docs.mongodb.com/manual/reference/operator/query-comparison
_OPERATORS = {
    ">": "$gt",
    ">=": "$gte",
    "==": "$eq",
    "!=": "$ne",
    "<": "$lt",
    "<=": "$lte",
    "in": "$in",
    "not in": "$nin",
}

def parse_query(s):
    chunks = s.split()

    if len(chunks) < 3:
        raise ValueError(
            "Invalid query string '%s'; all queries must have the syntax "
            "'<field> <operator> <value>'" % s
        )

    if (len(chunks) > 3) and (chunks[1:3] == ["not", "in"]):
        chunks = [chunks[0], "not in"] + chunks[3:]

    if len(chunks) > 3:
        chunks = chunks[:2] + [" ".join(chunks[2:])]

    field, op, val = chunks

    if op not in _OPERATORS:
        raise ValueError(
            "Unsupported comparison operation '%s'; the supported operations "
            "are {%s}" % (op, ", ".join(_OPERATORS.keys()))
        )

    try:
        return {field: {_OPERATORS[op]: json.loads(val)}}
    except json.JSONDecodeError:
        raise ValueError("Query value '%s' must be JSON parsable" % val)

Example uses:

print(parse_query("metadata.size_bytes > 1200"))
# {'metadata.size_bytes': {'$gt': 1200}}

print(parse_query('tag not in ["test", "train"]'))
# {'tag': {'$nin': ['test', 'train']}}

print(parse_query("metadata.size_bytes ~= 1200"))
# ValueError: Unsupported comparison operation '~='; the supported operations are {>, >=, ==, !=, <, <=, in, not in}

print(parse_query('tag not in {"test", "train"}'))
# ValueError: Query value '{"test", "train"}' must be JSON parsable

[tylerganter]

Resolved by #290