Help with custom factor

Hello,
I'm new to quant trading, but I had an idea that I really wanted to test. I have a list of stocks and I want to build an algorithm that buys the ones with the most statistically low value of a custom factor and short the ones with the highest.
I've attached my algorithm so far and I'm having some trouble with pipeline.
I've built a 10 day moving average of this factor and a 30 day moving average. What I want is to calculate the percent difference and attach it to the pipeline and use my custom filter to screen stocks. I get an error saying that there is nothing to graph. I assume that the pieline output is different from what I think it is.
I also noticed that there are some errors with my execution code.
I would greatly appreciate any help

Thank you so much,
Dan

from quantopian.pipeline import Pipeline, CustomFilter, CustomFactor  
from quantopian.algorithm import attach_pipeline, pipeline_output  
from quantopian.pipeline.factors import Latest  
from quantopian.pipeline.data.builtin import USEquityPricing  
from quantopian.pipeline.data import morningstar  
import numpy as np

class SidInList(CustomFilter):  
    inputs = []  
    window_length = 1  
    params = ('sid_list',)  
    def compute(self, today, assets, out, sid_list):  
        out[:] = np.in1d(assets, sid_list)  
class CustomFactor_10 (CustomFactor):  
    inputs = [morningstar.balance_sheet.net_assets, morningstar.valuation.market_cap, USEquityPricing.close]  
    window_length=10  
    def compute (self, today, assets, out, net, market, close):  
        out[:] = close[-1]/ (net[-1]/market[-1])  
class CustomFactor_30 (CustomFactor):  
    inputs = [morningstar.balance_sheet.net_assets, morningstar.valuation.market_cap, USEquityPricing.close]  
    window_length=30  
    def compute (self, today, assets, out, net, market, close):  
        out[:] = close[-1]/ (net[-1]/market[-1])  
CF_10= CustomFactor_10()  
CF_30= CustomFactor_30()  
percent_diff = (CF_10-CF_30)/CF_30

def my_pipeline(context):  
    set_symbol_lookup_date('2016-1-1')  
    my_sid_filter = SidInList(  
        sid_list = (  
            symbol('AFT').sid,  
            symbol('ARDC').sid,  
            symbol('ACP').sid,  
            symbol('BHL').sid,  
            symbol('FRA').sid,  
            symbol('BGT').sid,  
            symbol('BSL').sid,  
            symbol('BGX').sid,  
            symbol('BGB').sid,  
            symbol('EFT').sid,  
            symbol('EFF').sid,  
            symbol('EFR').sid,  
            symbol('EVF').sid,  
            symbol('FCT').sid,  
            symbol('VVR').sid,  
            symbol('VTA').sid,  
            symbol('JQC').sid,  
            symbol('JRO').sid,  
            symbol('JFR').sid,  
            symbol('NSL').sid,  
            symbol('JSD').sid,  
            symbol('OXLC').sid,  
            symbol('TSLF').sid,  
            symbol('PPR').sid,  
            symbol('PHD').sid,  
            symbol('TLI').sid,  
                )  
            )  
    close = Latest(  
            inputs=[USEquityPricing.close,percent_diff],  
            mask = my_sid_filter,  
            )           

    pipe = Pipeline(  
            columns = {'close' : close,  
                       'percent_diff' : percent_diff,  
                      },  
            screen = my_sid_filter,  
            )  
    return pipe

def initialize(context):  
    attach_pipeline(my_pipeline, 'my_pipeline')  
    schedule_function(trade, date_rules.every_day(), time_rules.market_open())  
def trade (context,data):  
    results = pipeline_output('my_pipeline')  
    securities_in_results = results.index  
    for sec in securities_in_results:  
        if data.can_trade(sec) & (percent_diff < 0):  
            order_target_percent(sec, 1.0 / len(securities_in_results))  
        if data.can_trade(sec) & (percent_diff > 0):  
            order_target_percent(sec,-1.0 / len (securities_in_results))

Hi Dan,

Sorry about the misleading error. What's happening is that you're calling attach_pipeline on a function (my_pipeline); that's the function that gets complained about having no .to_graph (Pipelines themselves have that method). To avoid that error, call the function to use attach_pipeline on the Pipeline object that is returned:

attach_pipeline(my_pipeline(context), 'my_pipeline')

A second issue with your code is that you're passing percent_diff into Latest. Latest converts a BoundColumn into a Factor; BoundColumns are columns of data like USEquityPricing.close. However, percent_diff is already a Factor, since it's constructed by doing arithmetic on custom Factors. So it's not necessary to pass percent_diff into Latest; it can be used by itself as a Pipeline column already. So remove percent_diff from the inputs argument to Latest.

With those two changes your algorithm should be up and running. Let me know if there's any more trouble.