Diff in diff gives strange output

[BDC fire]

Hi all
(see images below)

For a project, I have to check whether anti dumping measures against china had an effect on eu imports after 2016 (first measures, see image). Japan is the control group.
I cut my data from 2011 upto 2018, and it seems that the lines go rather parallel, as such i thought of doing diff in diff. However it does seem like things go wrong. estat trendplots gives strange increasing results(?) and estat ptrends gives no output at all.
The ATET coefficient on my interaction term is also positive, which would be the opposite of what i would expect.

Coefficient Std. err. t P>|t| [95% conf. interval]
treatXpost | 3.32445 4.467837 0.74 0.458 -5.496659 12.14556


Code and images are included, thanks in advance for any help.

import excel "C:\Users\........\Panel data format for stata.xlsx", sheet("Sheet1") firstrow

gen newdate = date(Yeardate, "DMY")
format newdate %td
order newdate, after(Yeardate)

encode Country, gen(country_num)
order country_num, after(Country)

gen treated = 0
replace treated = 1 if Country == "China"

gen postfirsttreatment = 0
replace postfirsttreatment=1 if newdate > date("01/07/2016","DMY")

generate treatXpost = treated * postfirsttreatment

drop if newdate >= date("01/01/2019","DMY")  /**avoid data where the second peak is starting*/


/** didregress (index12monthslidingwindow) (treatXpost), group(country_num) time(newdate) **/

xtset country_num
xtdidregress (index12monthslidingwindow treated postfirsttreatment) (treatXpost), group(country_num) time(newdate) aeq

estat trendplots
estat ptrends