Support to validate Patient IDs against Estonian Patient ID schema in conditions for various rules

[gunterze]

Support to validate Patient IDs against Estonian Patient ID schema in conditions for various rules. Configurable by condition property:

PatientID=§PATIENT_ID_EST - true, if Patient ID is compliant with Estonian Patient ID schema
PatientID!=§PATIENT_ID_EST - true, if Patient ID is not compliant with Estonian Patient ID schema

Estonian Patient ID schema:

            final Pattern PATTERN = Pattern.compile("[1-6]\\d\\d(0[1-9]|1[0-2])(0[1-9]|[12]\\d|3[01])\\d\\d\\d\\d");

            @Override
            public boolean test(String s) {
                return PATTERN.matcher(s).matches() && verifyChecksum(s.toCharArray());
            }

            private boolean verifyChecksum(char[] chars) {
                int checksum = checksum(chars, 1, 2, 3, 4, 5, 6, 7, 8, 9, 1);
                if (checksum == 10) {
                    checksum = checksum(chars, 3, 4, 5, 6, 7, 8, 9, 1, 2, 3);
                    if (checksum == 10)
                        checksum = 0;
                }
                return chars[10] - '0' == checksum;
            }

            private int checksum(char[] chars, int... weights) {
                int sum = 0;
                for (int i = 0; i < 10; i++) {
                    sum += (chars[i] - '0') * weights[i];
                }
                return sum % 11;
            }

[petrkalina]

note: the check can potentially succeed on a random number if:

• it has 11 numerical characters
• matches "[1-6]\\d\\d(0[1-9]|1[0-2])(0[1-9]|[12]\\d|3[01])\\d\\d\\d\\d" pattern
• the last number is valid EE checksum of the previous numbers (on a random number matching the above the probability of this randomly is 10%)

.. in overall, on the space of all PIDs, the P(false-positive) probability i:

P(false positive rnd) = 
P(syntax matches accidentally) * P(validation succeeds accidentally) =
P(first number is 1..6) * P(birth month is a valid month i.e. 1..12) * P(birth day is valid day i.e. 1..31)` * P(checksum matches) =
P(is 11 numerical digits) * 6/10 * 12/100 * 31/100 * 0.1 =
P(11 digits) * 0,002232

on existing PID DB, the P(false positive) can be estimated as

[count(invalid PIDs with valid format)  /  count(all PIDs)] * P(checksum accidentally matches) =
[count(invalid PIDs with valid format)  /  count(all PIDs)] * 0.1