sqlpp11/scripts/ddl2cpp

#!/usr/bin/env python3

##
# Copyright (c) 2013-2022, Roland Bock
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without modification,
# are permitted provided that the following conditions are met:
#
#  * Redistributions of source code must retain the above copyright notice,
#    this list of conditions and the following disclaimer.
#  * Redistributions in binary form must reproduce the above copyright notice,
#    this list of conditions and the following disclaimer in the documentation
#    and/or other materials provided with the distribution.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
# IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT,
# INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
# BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
# LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
# OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
# OF THE POSSIBILITY OF SUCH DAMAGE.
##

import pyparsing as pp
import sys
import re
import os

# error codes, we should refactor this later
ERROR_BAD_ARGS = 1
ERROR_DATA_TYPE = 10
ERROR_STRANGE_PARSING = 20

# Rather crude SQL expression parser.
# This is not geared at correctly interpreting SQL, but at identifying (and ignoring) expressions for instance in DEFAULT expressions
ddlLeft, ddlRight = map(pp.Suppress, "()")
ddlNumber = pp.Word(pp.nums + "+-.", pp.nums + "+-.Ee")
ddlString = (
    pp.QuotedString("'") | pp.QuotedString('"', escQuote='""') | pp.QuotedString("`")
)
ddlTerm = pp.Word(pp.alphas + "_", pp.alphanums + "_.$")
ddlName = pp.Or([ddlTerm, ddlString, pp.Combine(ddlString + "." + ddlString)])
ddlOperator = pp.Or(
    map(pp.CaselessLiteral, ["+", "-", "*", "/", "<", "<=", ">", ">=", "=", "%", "DIV"])
)

ddlBracedExpression = pp.Forward()
ddlFunctionCall = pp.Forward()
ddlCastEnd = "::" + ddlTerm
ddlCast = ddlString + ddlCastEnd
ddlBracedArguments = pp.Forward()
ddlExpression = pp.OneOrMore(
    ddlBracedExpression
    | ddlFunctionCall
    | ddlCastEnd
    | ddlCast
    | ddlOperator
    | ddlString
    | ddlTerm
    | ddlNumber
    | ddlBracedArguments
)

ddlBracedArguments << ddlLeft + pp.delimitedList(ddlExpression) + ddlRight
ddlBracedExpression << ddlLeft + ddlExpression + ddlRight

ddlArguments = pp.Suppress(pp.Group(pp.delimitedList(ddlExpression)))
ddlFunctionCall << ddlName + ddlLeft + pp.Optional(ddlArguments) + ddlRight

# Column and constraint parsers
ddlBooleanTypes = [
    "bool",
    "boolean",
]

ddlBoolean = pp.Or(
    map(pp.CaselessLiteral, sorted(ddlBooleanTypes, reverse=True))
).setParseAction(pp.replaceWith("boolean"))

ddlIntegerTypes = [
    "bigint",
    "int",
    "int2",  # PostgreSQL
    "int4",  # PostgreSQL
    "int8",  # PostgreSQL
    "integer",
    "mediumint",
    "smallint",
    "tinyint",
]
ddlInteger = pp.Or(
    map(pp.CaselessLiteral, sorted(ddlIntegerTypes, reverse=True))
).setParseAction(pp.replaceWith("integer"))

ddlSerialTypes = [
    "bigserial",  # PostgreSQL
    "serial",  # PostgreSQL
    "smallserial",  # PostgreSQL
]
ddlSerial = (
    pp.Or(map(pp.CaselessLiteral, sorted(ddlSerialTypes, reverse=True)))
    .setParseAction(pp.replaceWith("integer"))
    .setResultsName("hasAutoValue")
)

ddlFloatingPointTypes = [
    "decimal",  # MYSQL
    "double",
    "float8",  # PostgreSQL
    "float",
    "float4",  # PostgreSQL
    "numeric",  # PostgreSQL
    "real",
]
ddlFloatingPoint = pp.Or(
    map(pp.CaselessLiteral, sorted(ddlFloatingPointTypes, reverse=True))
).setParseAction(pp.replaceWith("floating_point"))

ddlTextTypes = [
    "char",
    "varchar",
    "character varying",  # PostgreSQL
    "text",
    "clob",
    "enum",  # MYSQL
    "set",
    "longtext",  # MYSQL
    "jsonb",  # PostgreSQL
    "json",  # PostgreSQL
    "tinytext",  # MYSQL
    "mediumtext",  # MYSQL
    "rational", # PostgreSQL pg_rationale extension
]

ddlText = pp.Or(
    map(pp.CaselessLiteral, sorted(ddlTextTypes, reverse=True))
).setParseAction(pp.replaceWith("text"))

ddlBlobTypes = [
    "bytea",
    "tinyblob",
    "blob",
    "mediumblob",
    "longblob",
    "binary",  # MYSQL
    "varbinary",  # MYSQL
]

ddlBlob = pp.Or(
    map(pp.CaselessLiteral, sorted(ddlBlobTypes, reverse=True))
).setParseAction(pp.replaceWith("blob"))


ddlDateTypes = [
    "date",
]

ddlDate = (
    pp.Or(map(pp.CaselessLiteral, sorted(ddlDateTypes, reverse=True)))
    .setParseAction(pp.replaceWith("day_point"))
    .setResultsName("warnTimezone")
)

ddlDateTimeTypes = [
    "datetime",
    "timestamp",
    "timestamp without time zone",  # PostgreSQL
    "timestamp with time zone",  # PostgreSQL
    "timestamptz",  # PostgreSQL
]

ddlDateTime = pp.Or(
    map(pp.CaselessLiteral, sorted(ddlDateTimeTypes, reverse=True))
).setParseAction(pp.replaceWith("time_point"))

ddlTimeTypes = [
    "time",
    "time without time zone",  # PostgreSQL
    "time with time zone",  # PostgreSQL
]

ddlTime = pp.Or(
    map(pp.CaselessLiteral, sorted(ddlTimeTypes, reverse=True))
).setParseAction(pp.replaceWith("time_of_day"))


ddlUnknown = pp.Word(pp.alphanums).setParseAction(pp.replaceWith("UNKNOWN"))

ddlType = (
    ddlBoolean
    | ddlInteger
    | ddlSerial
    | ddlFloatingPoint
    | ddlText
    | ddlBlob
    | ddlDateTime
    | ddlDate
    | ddlTime
    | ddlUnknown
)

ddlUnsigned = pp.CaselessLiteral("UNSIGNED").setResultsName("isUnsigned")
ddlDigits = "," + pp.Word(pp.nums)
ddlWidth = ddlLeft + pp.Word(pp.nums) + pp.Optional(ddlDigits) + ddlRight
ddlTimezone = (
    (pp.CaselessLiteral("with") | pp.CaselessLiteral("without"))
    + pp.CaselessLiteral("time")
    + pp.CaselessLiteral("zone")
)

ddlNotNull = pp.Group(
    pp.CaselessLiteral("NOT") + pp.CaselessLiteral("NULL")
).setResultsName("notNull")
ddlDefaultValue = pp.CaselessLiteral("DEFAULT").setResultsName("hasDefaultValue")

ddlAutoKeywords = [
    "AUTO_INCREMENT",
    "AUTOINCREMENT",
    "SMALLSERIAL",
    "SERIAL",
    "BIGSERIAL",
    "GENERATED",
]
ddlAutoValue = pp.Or(map(pp.CaselessLiteral, sorted(ddlAutoKeywords, reverse=True)))

ddlColumn = pp.Group(
    ddlName("name")
    + ddlType("type")
    + pp.Suppress(pp.Optional(ddlWidth))
    + pp.Suppress(pp.Optional(ddlTimezone))
    + pp.ZeroOrMore(
        ddlUnsigned("isUnsigned")
        | ddlNotNull("notNull")
        | pp.CaselessLiteral("null")
        | ddlAutoValue("hasAutoValue")
        | ddlDefaultValue("hasDefaultValue")
        | pp.Suppress(ddlExpression)
    )
)

ddlConstraintKeywords = [
    "CONSTRAINT",
    "PRIMARY",
    "FOREIGN",
    "KEY",
    "FULLTEXT",
    "INDEX",
    "UNIQUE",
    "CHECK",
    "PERIOD",
]
ddlConstraint = pp.Group(
    pp.Or(map(pp.CaselessLiteral, sorted(ddlConstraintKeywords, reverse=True)))
    + ddlExpression
).setResultsName("isConstraint")

# CREATE TABLE parser
ddlIfNotExists = pp.Group(
    pp.CaselessLiteral("IF") + pp.CaselessLiteral("NOT") + pp.CaselessLiteral("EXISTS")
).setResultsName("ifNotExists")
ddlOrReplace = pp.Group(
    pp.CaselessLiteral("OR") + pp.CaselessLiteral("REPLACE")
).setResultsName("orReplace")
ddlCreateTable = pp.Group(
    pp.CaselessLiteral("CREATE")
    + pp.Suppress(pp.Optional(ddlOrReplace))
    + pp.CaselessLiteral("TABLE")
    + pp.Suppress(pp.Optional(ddlIfNotExists))
    + ddlName.setResultsName("tableName")
    + ddlLeft
    + pp.Group(pp.delimitedList(pp.Suppress(ddlConstraint) | ddlColumn)).setResultsName(
        "columns"
    )
    + ddlRight
).setResultsName("create")
# ddlString.setDebug(True) #uncomment to debug pyparsing

ddl = pp.OneOrMore(pp.Suppress(pp.SkipTo(ddlCreateTable, False)) + ddlCreateTable)

ddlComment = pp.oneOf(["--", "#"]) + pp.restOfLine
ddl.ignore(ddlComment)


def testBoolean():
    for t in ddlBooleanTypes:
        result = ddlType.parseString(t, parseAll=True)
        assert result[0] == "boolean"


def testInteger():
    for t in ddlIntegerTypes:
        result = ddlType.parseString(t, parseAll=True)
        assert result[0] == "integer"


def testSerial():
    for t in ddlSerialTypes:
        result = ddlType.parseString(t, parseAll=True)
        assert result[0] == "integer"
        assert result.hasAutoValue


def testFloatingPoint():
    for t in ddlFloatingPointTypes:
        result = ddlType.parseString(t, parseAll=True)
        assert result[0] == "floating_point"


def testText():
    for t in ddlTextTypes:
        result = ddlType.parseString(t, parseAll=True)
        assert result[0] == "text"


def testBlob():
    for t in ddlBlobTypes:
        result = ddlType.parseString(t, parseAll=True)
        assert result[0] == "blob"


def testDate():
    for t in ddlDateTypes:
        result = ddlType.parseString(t, parseAll=True)
        assert result[0] == "day_point"


def testDateTime():
    for t in ddlDateTimeTypes:
        result = ddlType.parseString(t, parseAll=True)
        assert result[0] == "time_point"


def testTime():
    for t in ddlTimeTypes:
        result = ddlType.parseString(t, parseAll=True)
        assert result[0] == "time_of_day"


def testUnknown():
    for t in ["cheesecake", "blueberry"]:
        result = ddlType.parseString(t, parseAll=True)
        assert result[0] == "UNKNOWN"


def testAutoValue():
    def test(s, expected):
        results = ddlAutoValue.parseString(s, parseAll=True)
        print(results)


def testColumn():
    text = "\"id\" int(8) unsigned NOT NULL DEFAULT nextval('dk_id_seq'::regclass)"
    result = ddlColumn.parseString(text, parseAll=True)
    column = result[0]
    assert column.name == "id"
    assert column.type == "integer"
    assert column.isUnsigned
    assert column.notNull
    assert not column.hasAutoValue


def testConstraint():
    for text in [
        "CONSTRAINT unique_person UNIQUE (first_name, last_name)",
        "UNIQUE (id)",
        "UNIQUE (first_name,last_name)"
    ]:
        result = ddlConstraint.parseString(text, parseAll=True)
        assert result.isConstraint

def testMathExpression():
        text = "2 DIV 2"
        result = ddlExpression.parseString(text, parseAll=True)
        assert len(result) == 3
        assert result[0] == "2"
        assert result[1] == "DIV"
        assert result[2] == "2"


def testRational():
    for text in [
        "pos RATIONAL NOT NULL DEFAULT nextval('rational_seq')::integer",
     ]:
        result = ddlColumn.parseString(text, parseAll=True)
        column = result[0]
        assert column.name == "pos"
        assert column.type == "text"
        assert column.notNull


def testTable():
    text = """
  CREATE TABLE "public"."dk" (
  "id" int8 NOT NULL DEFAULT nextval('dk_id_seq'::regclass),
  "strange" NUMERIC(314, 15),
  "last_update" timestamp(6) DEFAULT now(),
   PRIMARY KEY (id)
)
"""
    result = ddlCreateTable.parseString(text, parseAll=True)


def testParser():
    testBoolean()
    testInteger()
    testSerial()
    testFloatingPoint()
    testText()
    testBlob()
    testDate()
    testTime()
    testUnknown()
    testDateTime()
    testColumn()
    testConstraint()
    testMathExpression()
    testRational()
    testTable()


# CODE GENERATOR
# HELPERS


def get_include_guard_name(namespace, inputfile):
    val = re.sub("[^A-Za-z0-9]+", "_", namespace + "_" + os.path.basename(inputfile))
    return val.upper()


def identity_naming_func(s):
    return s


def repl_camel_case_func(m):
    if m.group(1) == "_":
        return m.group(2).upper()
    else:
        return m.group(1) + m.group(2).upper()


def class_name_naming_func(s):
    s = s.replace(".", "_")
    return re.sub("(^|\s|[_0-9])(\S)", repl_camel_case_func, s)


def member_name_naming_func(s):
    s = s.replace(".", "_")
    return re.sub("(\s|_|[0-9])(\S)", repl_camel_case_func, s)


def repl_func_for_args(m):
    if m.group(1) == "-":
        return m.group(2).upper()
    else:
        return m.group(1) + m.group(2).upper()


def setArgumentBool(s, bool_value):
    first_lower = (
        lambda s: s[:1].lower() + s[1:] if s else ""
    )  # http://stackoverflow.com/a/3847369/5006740
    var_name = first_lower(re.sub("(\s|-|[0-9])(\S)", repl_func_for_args, s))
    globals()[var_name] = bool_value


def escape_if_reserved(name):
    reserved_names = [
        "BEGIN",
        "END",
        "GROUP",
        "ORDER",
    ]
    if name.upper() in reserved_names:
        return "!{}".format(name)
    return name


def beginHeader(pathToHeader, namespace, nsList):
    header = open(pathToHeader, "w")
    print("// generated by " + " ".join(sys.argv), file=header)
    print("#ifndef " + get_include_guard_name(namespace, pathToHeader), file=header)
    print("#define " + get_include_guard_name(namespace, pathToHeader), file=header)
    print("", file=header)
    print("#include <sqlpp11/table.h>", file=header)
    print("#include <sqlpp11/data_types.h>", file=header)
    print("#include <sqlpp11/char_sequence.h>", file=header)
    print("", file=header)
    for ns in nsList:
        print("namespace " + ns, file=header)
        print("{", file=header)
    return header


def endHeader(header, nsList):
    for ns in reversed(nsList):
        print("} // namespace " + ns, file=header)
    print("#endif", file=header)
    header.close()


def help_message():
    arg_string = ""
    pad = 0
    for argument in list(optionalArgs.keys()):
        if len(argument) > pad:
            pad = len(argument)
    for argument in list(optionalArgs.keys()):
        if len(argument) < pad:
            padding = " " * (pad - len(argument))
        else:
            padding = ""
        arg_string = (
            arg_string + argument + ":  " + padding + optionalArgs[argument] + "\n"
        )
    print(
        "Usage:\n"
        "ddl2cpp [optional args] <path to ddl> <path to target> <namespace>\n\n"
        "OPTIONAL ARGUMENTS:\n" + arg_string + "\n"
        "<path to ddl>    path to your SQL database/table definitions (SHOW CREATE TABLE SomeTable) \n"
        "<path to target> path to a generated C++ header file without extension (no *.h). \n"
        "<namespace>      namespace you want. Usually a project/database name\n"
    )
    sys.exit(0)


optionalArgs = {
    # if  -some-key is present, it will set variable someKey to True
    "-no-timestamp-warning": "show warning about date / time data types",  # noTimeStampWarning = True
    "-auto-id": "Assume column 'id' to have an automatic value as if AUTO_INCREMENT was specified (e.g. implicit for SQLite ROWID)",  # autoId = True
    "-identity-naming": "Use table and column names from the ddl (defaults to UpperCamelCase for tables and lowerCamelCase for columns)",  # identityNaming = True
    "-split-tables": "Make a header for each table name, using target as a directory",  # splitTables = True
    "--help": "show this help",
    "--test": "run parser self-test",
}

noTimestampWarning = False
autoId = False
identityNaming = False
splitTables = False


def createHeader():
    global noTimestampWarning
    # ARGUMENT PARSING
    if len(sys.argv) < (4):
        help_message()
        sys.exit(ERROR_BAD_ARGS)

    firstPositional = 1
    if len(sys.argv) >= 4:
        for arg in sys.argv:
            if arg in list(optionalArgs.keys()):
                setArgumentBool(arg, True)
                firstPositional += 1
            else:
                pass

    if identityNaming:
        toClassName = identity_naming_func
        toMemberName = identity_naming_func
    else:
        toClassName = class_name_naming_func
        toMemberName = member_name_naming_func

    pathToDdl = sys.argv[firstPositional]

    pathToHeader = sys.argv[firstPositional + 1] + ("/" if splitTables else ".h")
    namespace = sys.argv[firstPositional + 2]

    try:
        tableCreations = ddl.parseFile(pathToDdl)
    except pp.ParseException as e:
        print("ERROR: Could not parse any CREATE TABLE statement in " + pathToDdl)
        # print(pp.parseError)
        sys.exit(ERROR_STRANGE_PARSING)

    nsList = namespace.split("::")

    # PROCESS DDL
    tableCreations = ddl.parseFile(pathToDdl)

    header = 0
    if not splitTables:
        header = beginHeader(pathToHeader, namespace, nsList)
    DataTypeError = False
    for create in tableCreations:
        sqlTableName = create.tableName
        if splitTables:
            header = beginHeader(pathToHeader + sqlTableName + ".h", namespace, nsList)
        tableClass = toClassName(sqlTableName)
        tableMember = toMemberName(sqlTableName)
        tableNamespace = tableClass + "_"
        tableTemplateParameters = tableClass
        print("  namespace " + tableNamespace, file=header)
        print("  {", file=header)
        for column in create.columns:
            if column.isConstraint:
                continue
            sqlColumnName = column.name
            columnClass = toClassName(sqlColumnName)
            tableTemplateParameters += (
                ",\n               " + tableNamespace + "::" + columnClass
            )
            columnMember = toMemberName(sqlColumnName)
            columnType = column.type
            if columnType == "UNKNOWN":
                print(
                    "Error: datatype of %s.%s is not supported."
                    % (sqlTableName, sqlColumnName)
                )
                DataTypeError = True
            if columnType == "integer" and column.isUnsigned:
                columnType = columnType + "_unsigned"
            if columnType == "time_point" and not noTimestampWarning:
                print(
                    "Warning: date and time values are assumed to be without timezone."
                )
                print(
                    "Warning: If you are using types WITH timezones, your code has to deal with that."
                )
                print("You can disable this warning using -no-timestamp-warning")
                noTimestampWarning = True
            traitslist = ["sqlpp::" + columnType]
            columnCanBeNull = not column.notNull
            print("    struct " + columnClass, file=header)
            print("    {", file=header)
            print("      struct _alias_t", file=header)
            print("      {", file=header)
            print(
                '        static constexpr const char _literal[] =  "'
                + escape_if_reserved(sqlColumnName)
                + '";',
                file=header,
            )
            print(
                "        using _name_t = sqlpp::make_char_sequence<sizeof(_literal), _literal>;",
                file=header,
            )
            print("        template<typename T>", file=header)
            print("        struct _member_t", file=header)
            print("          {", file=header)
            print("            T " + columnMember + ";", file=header)
            print(
                "            T& operator()() { return " + columnMember + "; }",
                file=header,
            )
            print(
                "            const T& operator()() const { return "
                + columnMember
                + "; }",
                file=header,
            )
            print("          };", file=header)
            print("      };", file=header)
            requireInsert = True
            hasAutoValue = column.hasAutoValue or (autoId and sqlColumnName == "id")
            if hasAutoValue:
                traitslist.append("sqlpp::tag::must_not_insert")
                traitslist.append("sqlpp::tag::must_not_update")
                requireInsert = False
            if not column.notNull:
                traitslist.append("sqlpp::tag::can_be_null")
                requireInsert = False
            if column.hasDefaultValue:
                requireInsert = False
            if requireInsert:
                traitslist.append("sqlpp::tag::require_insert")
            print(
                "      using _traits = sqlpp::make_traits<"
                + ", ".join(traitslist)
                + ">;",
                file=header,
            )
            print("    };", file=header)
        print("  } // namespace " + tableNamespace, file=header)
        print("", file=header)

        print(
            "  struct "
            + tableClass
            + ": sqlpp::table_t<"
            + tableTemplateParameters
            + ">",
            file=header,
        )
        print("  {", file=header)
        print("    struct _alias_t", file=header)
        print("    {", file=header)
        print(
            '      static constexpr const char _literal[] =  "' + sqlTableName + '";',
            file=header,
        )
        print(
            "      using _name_t = sqlpp::make_char_sequence<sizeof(_literal), _literal>;",
            file=header,
        )
        print("      template<typename T>", file=header)
        print("      struct _member_t", file=header)
        print("      {", file=header)
        print("        T " + tableMember + ";", file=header)
        print("        T& operator()() { return " + tableMember + "; }", file=header)
        print(
            "        const T& operator()() const { return " + tableMember + "; }",
            file=header,
        )
        print("      };", file=header)
        print("    };", file=header)
        print("  };", file=header)
        if splitTables:
            endHeader(header, nsList)

    if not splitTables:
        endHeader(header, nsList)
    if DataTypeError:
        print("Error: unsupported datatypes.")
        print("Possible solutions:")
        print("A) Implement this datatype (examples: sqlpp11/data_types)")
        print("B) Extend/upgrade ddl2cpp (edit types map)")
        print("C) Raise an issue on github")
        sys.exit(10)  # return non-zero error code, we might need it for automation


if __name__ == "__main__":
    if "--help" in sys.argv:
        help_message()
        sys.exit()
    elif "--test" in sys.argv:
        testParser()
        sys.exit()
    else:
        createHeader()