I can help only to some degree here.

I've seen issues like that at times including errors for code that only existed in the past and isn't there anymore, forcing a theory: Maybe algos are pieced together on build from changes and that process becomes confused sometimes, but I don't know. Creating a new algo seemed to help.

One thing I do know. This sort of thing will only honor one of them:

price_filter = ( price_low < prices.latest < price_high )
is_small_cap = (3000000 <= dat.Fundamentals.market_cap.latest <= 2000000000)

So you'll want to separate them out for both to take effect. That's been my experience. Then it is fine.
m is the mask:

m   = (dat.Fundamentals.market_cap.latest <= 2000000000)  # is_small_cap  
m & = (3000000 <= dat.Fundamentals.market_cap.latest)  

@Blue Seahawk
Tried creating a new algo and copy/pasting the code into it and nothing changed. Oh well. For the time being I've just deleted that portion and I'll just put the calculations somewhere else in the code. Thank you.

This may help,

from quantopian.algorithm import attach_pipeline, pipeline_output  
from quantopian.pipeline.data import Fundamentals  
from quantopian.pipeline import Pipeline, filters  
# -------------------------------------------------------------------  
STK_SET, UB, LB = filters.QTradableStocksUS(), 2000000000, 300000000  
# -------------------------------------------------------------------  
def initialize(context):  
    market_cap = Fundamentals.market_cap.latest  
    small_cap = (STK_SET & (market_cap < UB ) & (market_cap > LB))  
    attach_pipeline(Pipeline(columns = {'market_cap': market_cap}, screen = small_cap) , 'market_cap_pipe')

def before_trading_start(context, data):  
    stocks = pipeline_output('market_cap_pipe').index  
    MC = pipeline_output('market_cap_pipe').market_cap

    small_cap_MC = 0  
    for stock in stocks:  
        small_cap_MC += MC[stock] 

    print (small_cap_MC, len(stocks))  
    record(small_cap_MC = small_cap_MC)