The output circuit is missing but I'll try to interpret the rest.
The bottom row is your basic sine-wave generator. The 4017 in the lower-left corner is a ten-step decoded counter. The ten resistors on its output form a D-to-A converter specialized to produce a ten-step approximation of half a sine wave from the moving 1. (Note that the schematic has an error: R17 should be 27k, not 47k, to be consistent with the rest of the circuit.) The 4013 on the right counts which half of the sine wave you're on, so you get a twenty-step sine wave. Q1 is an emitter-follower amplifier, to "harden up" the output voltage by sourcing more current, so the following stages don't load it down and/or distort it.
The top row is a voltage level monitor, monitoring the AC voltage and adjusting the drive to regulate it to the correct level. The right-hand side is a "power supply", generating a DC voltage in proportion to the AC voltage. The section with the zener provides a reference voltage, and the 3140 op-amp is a subtractor, resulting in an error voltage at the junction between diodes D2 and D2.
I THINK what's happening with D3 is that it's acting as a variable load on the sine wave from the lower row, adjusting the level of the half-sine-wave on the vertical wire on the left. If that's right, D2 might be there to "tell" the op-amp what the shape of D3's response is. I'd have to think about that a bit more to get it right, though.
The middle row is the basic oscilator and pulse former. The SG3525 is a pulse-width modulator, which sets the frequency and creates a train of variable width pulses going right. (I don't have a pinout diagram for it so I can't give you the details story.) The pulses feed down to the 4017, advancing it one step with each pulse, and the lower row controls the pulse width. The variable width pulses go to the 4082A dual and gate, where they are anded by the "which half-cycle" signal from the 4013.
The result is twenty variable-width drive pulses per cycle, ten on Drv-1 and ten on Drv-2. The width of the pulses is adjusted to produce an approximation of a regulated sine wave (plus a bunch of energy at the 20th harmonic and multiples of it, which for 60 Hz would start at 1,200 Hz.)
If you have an output circuit composed of:
- Two switching transistors
- Two free-wheeling diodes
- A transformer, and
- A radio interference filter
you should get a darned good sine wave out of it.
You might also be able to substitute a couple inductors and a couple capacitors for the transformer, switching-regulator style, although 1.2 kilohertz seems a bit low for that.