RAW, it probably does 0 damage, because nothing says otherwise
It says “half,” so you probably round both sides down and get 0. Note that the “minimum 1” rule is for a hit specifically, and this spell doesn’t involve one.
That said, every group I’ve played with has allowed the caster to choose which way the odd damage went in a half-and-half situation. Usually, it’s 1 damage, who cares? And certainly the minimum 1 thing should also apply to non-hit damage, as here. With thunderhead, it is the entire spell, though, so I wouldn’t allow it to apply to each half, which would effectively double thunderhead’s damage. Alternating bolts of electricity and sonic damage probably makes the most sense.
The stunning effect comes when the spell “concludes”
I actually don’t think it’s either the first or each bolt, but rather the last bolt, when the spell is complete.
That said, I don’t think the term “concludes” is actually defined, and it’s conceivable that they meant the casting of the spell is complete, so it would happen as the first step. Either way, though, I don’t see any way to justify it on each bolt.
The weapons will deal their normal, magically-enhanced damage. Minus 5, for the golem's DR, unless the weapons are also adamantine. Immunity to magic doesn't automatically destroy all magic the golem comes into contact with, it just can't be directly affected by magical effects.
PFSRD wrote:
Immunity to Magic (Ex)
A flesh golem is immune to any spell or spell-like ability that allows spell resistance. In addition, certain spells and effects function differently against the creature...
Emphasis mine.
If something allows spell resistance, immunity to magic functions like infinite spell resistance against it. Things that aren't stopped by spell resistance (like sword swings, acid arrows, and being crushed under a ton of solid rock during an earthquake spell) aren't stopped by immunity to magic. It also specifically only works against spells and spell-like abilities, so it doesn't stop things like dragon breath or magic swords.
So we've established that "immune to magic" doesn't turn off a magic weapon's +1. What does that mean for the golem's DR? In this case, nothing.
PFSRD wrote:
Weapons with an enhancement bonus of +3 or greater can ignore some types of damage reduction, regardless of their actual material or alignment. The following table shows what type of enhancement bonus is needed to overcome some common types of damage reduction.
Cold iron / silver +3
Adamantine* +4
Alignment-based +5
So the weapons would need to be +4 to bypass the golem's DR 5/adamantine. If they were +4, they would overcome the DR despite the golem's immune to magic trait; but it's a moot point since they're only +1 weapons to start with, therefore they only overcome DR /magic (and the various damage types for their weapon types).
Best Answer
What happens in general varies a lot based on the relative timing of the two effects, the specific effects involved, and the nature of their sources' descriptions. The general answer is that either one of the effects happens, or the other effect happens, or some third thing happens which is neither the first effect nor the second, but in no case should a logically contradictory state result.
In your explanation both of the example effects have the two damage sources occurring simultaneously and the effects are not so much contradictory as conflicting, so we will from here on deal exclusively with this case.
The exact nature of attack resolution has been a subject of some debate for some time (see this pathfinder question which links to this 3.5 question), and this question to some extent hinges on how atomic the attack action is. While some groups declare that the source of a simultaneous group of effects chooses in what order the effects are resolved, most groups, my own included, declare that all effects in a simultaneous group should be resolved simultaneously. It should be noted that this is a different question, as many groups that play with non-atomic attack actions hold that the dealing of damage is a single atomic step in an attack.
The minority opinion is, in this case, much simpler to adjudicate: The warrior swinging the sword or the mage casting the spell is asked whether he wants the fire or electricity damage to be dealt first (or the Gm just has him declare what the damage dice are from as he rolls if the GM wants to be sneaky, using the order of declaration as the order of resolution). If the electricity damage happens first, the golem is healed and slowed. If the fire damage happens first, the golem is just healed.
The majority opinion is somewhat more complicated since the electricity and fire damage happen at the same time. I personally use the Waving Hands resolution system for simultaneous effects, because I think it makes the most sense and does a good job of being internally consistent. In this case, the electricity effect triggers at the same time as the fire damage is causing the creature to have just begun being slow, so the electricity effect counters the fire one and the creature is just healed by the electricity effect.