pmm-admin purge causes client to panic

pmm-admin purge sends DELETE request to Prometheus, which isn't implemented in Prometheus 2.x.

> DELETE /prometheus/api/v1/series?match[]={job="linux",instance="XXX"} HTTP/1.1
> Host: 127.0.0.1
> User-Agent: Go-http-client/1.1
> Accept-Encoding: gzip
{{> }}
{{> }}
2018/08/21 13:57:45 response:
< HTTP/1.1 500 Internal Server Error
< Access-Control-Allow-Headers: Accept, Authorization, Content-Type, Origin
< Access-Control-Allow-Methods: GET, OPTIONS
< Access-Control-Allow-Origin: *
< Access-Control-Expose-Headers: Date
< Connection: keep-alive
< Content-Type: application/json
< Date: Tue, 21 Aug 2018 11:57:45 GMT
< Server: nginx
{{< }}
< {"status":"error","errorType":"internal","error":"not implemented"}
panic: interface conversion: interface {} is nil, not map[string]interface {}goroutine 1 [running]:
github.com/percona/pmm-client/pmm.(*Admin).PurgeMetrics(0xbe4e60, 0x7ffc6713756e, 0xd, 0xc4201735b8, 0xc42013db68, 0x7ff262)
/tmp/go/src/github.com/percona/pmm-client/pmm/main.go:427 +0x5a4
main.glob..func39(0xbde460, 0xc42017a340, 0x1, 0x2)
/tmp/go/src/github.com/percona/pmm-client/pmm-admin.go:1179 +0xd1
github.com/percona/pmm-client/vendor/github.com/spf13/cobra.(*Command).execute(0xbde460, 0xc42017a300, 0x2, 0x2, 0xbde460, 0xc42017a300)
/tmp/go/src/github.com/percona/pmm-client/vendor/github.com/spf13/cobra/command.go:654 +0x2a2
github.com/percona/pmm-client/vendor/github.com/spf13/cobra.(*Command).ExecuteC(0xbdeb20, 0xc42013ddd0, 0x49b302, 0x0)
/tmp/go/src/github.com/percona/pmm-client/vendor/github.com/spf13/cobra/command.go:729 +0x2d4
github.com/percona/pmm-client/vendor/github.com/spf13/cobra.(*Command).Execute(0xbdeb20, 0xc0263c, 0x92bf0f)
/tmp/go/src/github.com/percona/pmm-client/vendor/github.com/spf13/cobra/command.go:688 +0x2b
main.main()
/tmp/go/src/github.com/percona/pmm-client/pmm-admin.go:1398 +0x18dc

The  API in  Prom2 - https://prometheus.io/docs/prometheus/latest/querying/api/#delete-series 

Additionally, it's better to call  https://prometheus.io/docs/prometheus/latest/querying/api/#clean-tombstones after  a Prevoiuous call 

TODO: 

1) Implement both   Prom1 and Prom2   DELETE  requests   (be ready to get errors on them )

2) If  `tsdb` request is OK - then  also  Do  Clean Tombstones  request to free space