附件文章在使用空间双重差分时进行了平行趋势检验，比较了直接效应系数的差异。请教各位，如何在stata上实现这种检验及作图？原文“Fig. 3 plots the direct effect and the 95% confidence interval of Dk it with the spatial weight matrix ofW1. The horizontal axis is the number of years before and after the initiation of HSR operations. ”

我仿照常规did平行趋势检验写了如下命令：
gen policy = year-HSR
replace policy=-5 if policy<=-5
replace policy=10 if policy>=10

gen policy_d=policy+5
xsmle y ib5.policy_d , fe model(sdm) wmat(invdis) type(both) nolog effects


但结果无法计算出直接效应和间接效应（提示Warning: direct and indirect effects cannot be computed if factor variables are specified         option -effects- ignored. Notice that total effects can be obtained using -margins-）
后续使用coefplot命令作图也是空白，请问如何计算出直接效应，并把直接效应系数及置信区间作出图来？