Parsing Google calendar to DDay.iCal

Question

I'm working on application which parses Google Calendar via Google API to DDay.iCal

The main attributes, properties are handled easily... ev.Summary = evt.Title.Text;

The problem is when I got an recurring event, the XML contains a field like:

<gd:recurrence>
    DTSTART;VALUE=DATE:20100916
    DTEND;VALUE=DATE:20100917
    RRULE:FREQ=YEARLY
</gd:recurrence>

or

<gd:recurrence>
  DTSTART:20100915T220000Z
  DTEND:20100916T220000Z
  RRULE:FREQ=YEARLY;BYMONTH=9;WKST=SU"
</gd:recurrence>

using the following code:

    String[] lines = 
evt.Recurrence.Value.Split(new char[] { '\n', '\r' }, StringSplitOptions.RemoveEmptyEntries);

                    foreach (String line in lines)
                    {

                        if (line.StartsWith("R"))
                        {
                            RecurrencePattern rp = new RecurrencePattern(line);
                            ev.RecurrenceRules.Add(rp);
                        }
                        else 
                        {
                            ISerializationContext ctx = new SerializationContext();
                            ISerializerFactory factory = new DDay.iCal.Serialization.iCalendar.SerializerFactory();

                            ICalendarProperty property = new CalendarProperty();

                            IStringSerializer serializer = factory.Build(property.GetType(), ctx) as IStringSerializer;

                            property = (ICalendarProperty)serializer.Deserialize(new StringReader(line));

                            ev.Properties.Add(property);
                            Console.Out.WriteLine(property.Name + " - " + property.Value);
                        }

                    }

RRULEs are parsed correctly, but the problem is that other property (datetimes) values are empty...

Answer 1

Here is the starting point of what I'm doing, going off of the RFC-5545 spec's recurrence rule. It isn't complete to the spec and may break given certain input, but it should get you going. I think this should all be doable using RegEx, and something as heavy as a recursive decent parser would be overkill.

RRULE:(?:FREQ=(DAILY|WEEKLY|SECONDLY|MINUTELY|HOURLY|DAILY|WEEKLY|MONTHLY|YEARLY);)?(?:COUNT=([0-9]+);)?(?:INTERVAL=([0-9]+);)?(?:BYDAY=([A-Z,]+);)?(?:UNTIL=([0-9]+);)?

I am building this up using http://regexstorm.net/tester.

The test input I'm using is:

DTSTART;TZID=America/Chicago:20140711T133000\nDTEND;TZID=America/Chicago:20140711T163000\nRRULE:FREQ=WEEKLY;INTERVAL=8;BYDAY=FR;UNTIL=20141101

DTSTART;TZID=America/Chicago:20140711T133000\nDTEND;TZID=America/Chicago:20140711T163000\nRRULE:FREQ=WEEKLY;COUNT=5;INTERVAL=8;BYDAY=FR;UNTIL=20141101

DTSTART;TZID=America/Chicago:20140711T133000\nDTEND;TZID=America/Chicago:20140711T163000\nRRULE:FREQ=WEEKLY;BYDAY=FR;UNTIL=20141101

Sample matching results would look like:

Index    Position    Matched String                                                 $1      $2  $3  $4  $5
0        90          RRULE:FREQ=WEEKLY;INTERVAL=8;BYDAY=FR;UNTIL=20141101           WEEKLY      8   FR  20141101
1        236         RRULE:FREQ=WEEKLY;COUNT=5;INTERVAL=8;BYDAY=FR;UNTIL=20141101   WEEKLY  5   8   FR  20141101
2        390         RRULE:FREQ=WEEKLY;BYDAY=FR;UNTIL=20141101                      WEEKLY          FR  20141101

Usage is like:

string freqPattern = @"RRULE:(?:FREQ=(DAILY|WEEKLY|SECONDLY|MINUTELY|HOURLY|DAILY|WEEKLY|MONTHLY|YEARLY);?)?(?:COUNT=([0-9]+);?)?(?:INTERVAL=([0-9]+);?)?(?:BYDAY=([A-Z,]+);?)?(?:UNTIL=([0-9]+);?)?";
MatchCollection mc = Regex.Matches(rule, freqPattern, System.Text.RegularExpressions.RegexOptions.IgnoreCase);
foreach (Match m in mc)
{
    string frequency = m.Groups[1].ToString();
    string count = m.Groups[2].ToString();
    string interval = m.Groups[3].ToString();
    string byday = m.Groups[4].ToString();
    string until = m.Groups[5].ToString();
    System.Console.WriteLine("recurrence => frequency: \"{0}\", count: \"{1}\", interval: \"{2}\", byday: \"{3}\", until: \"{4}\"", frequency, count, interval, byday, until);
}

Answer 2

This is a great example of when to use regular expressions. Try this out for general parsing:

\s*(\w+):((\w+=\w+;)+(\w+=\w+)?|\w+)

Or, you might decide to have something more schema-specific.

\s*(?:DTSTART:)(?'Start'\w+)
\s*(?:DTEND:)(?'End'\w+)
\s*(?:RRULE:)(?'Rule'(\w+=\w+;)+(\w+=\w+)?)