Warning: math and theoretical numbers ahead.
1 hp = 745.699872 watts - source google.com "hp in watts"
watts == amps * volts
So if you have a 1HP motor running at 120V that's roughly 6.2 amps (745.699872 / 120).
That's the simple math, the truth is of course more complicated.
Motors aren't anywhere near 100% efficient so you won't get 1HP for 6.2 amps at 120v in the real world. A motor with good efficiency might be 75-80% efficient so would actually draw ~1000watts at full load (and more than twice that at startup for surge - which is one reason you want some battery backing). Its really hard to know what the efficiency of a given motor is without measuring it and they do vary very wildly from 80% down to 50% or sometimes even lower (you
might) get the data from the manufacturer, but its kind of like a snipe - a rare bird indeed). Generally the heavier the motor is the better it is in this regard for a given timeframe (some newer motors are lighter but use more efficient designs).
Countering that once the motor is running its draw depends on actual load. So for a 1HP saw motor it might draw up to 3000 watts for ~1s at startup, then idle at maybe 2-3 watts until you put it under load (i.e. for the duration ripping a 2x4) or maybe even higher if you take it over the rated load (by pushing to fast). I would expect a lathe to have a moderately low "average" draw just based on hand waving numbers because the motor is mostly idling once it gets the piece spun up unless you really start hogging material off (even then momentum is on your side). Adding in power converters (like a VFD for variable speed on a lathe) also imposes overhead, although I don't have handy numbers in front of me for that, there is non trivial loss there as well.
The numbers this guy:
http://www.generatorsales.com/wattage-calculator.asp has are mostly believable as full load numbers, but there is really no substitute for measuring actual draw on your tools under use if you can. The "rated watts" number there is at nominal peak usage so for example the fridge obviously wouldn't take 500w continuous but only while its running (and there is generally a moderately healthy budget on top of actual draw in most of those numbers - source: abusing datacenter power usage by squeezing more computers in than we really should have.. don't turn them all on at once
and then looking at other power supply/motor ratings).
You do have to size the system for all the things that can be running at once though.. So if you have a small system you might consider piggy backing some of the RV power feed onto it, but have a transfer switch so that when the table saw is running the fridge doesn't kick in as well and trip the breaker.
You can get a clamp on ammeter that just clamps over the power cord that is accurate enough for rough estimates for around $50 which imho would be money well spent before dropping a few grand on a solar system that may or may not work (or may be oversized so you spend less than its worth).
On the solar side you need to oversize the panels to absorb loss in the system. Generally you would chain it together vaguely like this:
panels->charge controller[1]->batteries[2]->inverter[3]->power out
1. the charge controller regulates the power to the batteries to make sure that they aren't over voltaged (frying them) and may convert the power from the solar cell type to a battery happy type (i.e AC->DC or 30v to 6v or whatever). Better ones are smart enough to detect bad batteries or other problems. This process obviously introduces some loss; figure maybe 20% would be reasonable (a good system might be less).
2. The batteries provide buffering for when there is no sun (obvious). They also provide backup power for when a motor starts up and you get a surge. Its also useful to put the power through them as ~most (all?) solar arrays have the output voltage vary depending on the amount of sun which if you directly fed that to tools would be bad (brown outs from the clouds).
3. The inverter converts battery voltage/type (xxV DC) to useful power (120v AC). Again some loss, 20% or less depending on the quality of the components.