If you accept Paul Creston's concept for classifying meters, we can find ways of reasoning about how meters, that are traditionally considered to be odd meters, can be seen as improper divisions of natural numbers into parts that don't make sense. You can take odd meters that are fundamentally "peg-legged" and make them into regular rhythms.
You can't divide the meter 5/4 into regular subdivisions of its two pulses. For example, dividing 9/8 in half gives you two pulses of 4.5 beats which is undefined in the natural numbers. You can see all of these so called odd meters as fragments of meters that have been divided into, most commonly, two or three parts. We can use a new approach that considers 5/4 as half of a bar in 10/4 meter. And by Paul Creston's rules for how metrical groups work, you can transform two bars of 5/4 into one bar of 10/4 binary meter that has two pulses subdivided into regular 5 note groups (regular meaning all the pulses have the same number of subdivisions). This resolves the problem of 5/4 being an unfortunate odd meter and actually feels pretty smooth once you get a feel for how to divide 2 binary beats into 10 evenly spaced quarter notes.
And by simply extending this logic we can use this little python script to find other ways of grouping these incomplete odd meters into beautiful flowing meters with many different ways of subdividing them. When it comes to filling in what Paul Creston calls "patterns", you can use additive concepts that sum to the number that is that same as the subdivision count of the pulse.
In Creston's concept there is a distinction between the actual metrical pulse which is defined by the common subdivision of the meter that is indicated and a metrical beat which is the rhythmic beat that is actually felt. For example, you can divide 3/4 into its metrical pulse, or a metrical beat of 2 dotted quarters.
Combining Creston's concepts for accents, patterns, and polyrhythms with the above regular meters can create incredible worlds of rhythm.