Logstash config - S3 Filenames? - ruby

Question

I'm attempting to use Logstash to collect sales data for multiple clients from multiple vendors.
So far i've an S3 (Inbox) bucket that I can drop my files (currently CSVs) into and according the client code prefix on the file, the data gets pushed into an Elastic index for each client. This all works nicely.
The problem I have is that I have data from multiple vendors and need a way of identifying which file is from which vendor. Adding an extra column to the CSV is not an option, so my plan was to add this to the filename, so i'd end up with a filenaming convention something like clientcode_vendorcode_reportdate.csv.
I may be missing something, but it seems that on S3 I can't get access to the filename inside my config, which seems crazy given that the prefix is clearly being read at some point. I was intending to use a Grok or Ruby filter to simply split the filename on the underscore, giving me three key variables that I can use in my config, but all attempts so far have failed. I can't even seem to get the full source_path or filename as a variable.
My config so far looks something like this - i've removed failed attempts at using Grok/Ruby filters.
input {
s3 {
access_key_id =>"MYACCESSKEYID"
bucket => "sales-inbox"
backup_to_bucket => "sales-archive"
delete => "true"
secret_access_key => "MYSECRETACCESSKEY"
region => "eu-west-1"
codec => plain
prefix => "ABC"
type => "ABC"
}
s3 {
access_key_id =>"MYACCESSKEYID"
bucket => "sales-inbox"
backup_to_bucket => "sales-archive"
delete => "true"
secret_access_key => "MYSECRETACCESSKEY"
region => "eu-west-1"
codec => plain
prefix => "DEF"
type => "DEF"
}
}
filter {
if [type] == "ABC" {
csv {
columns => ["Date","Units","ProductID","Country"]
separator => ","
}
mutate {
add_field => { "client_code" => "abc" }
}
}
else if [type] == "DEF" {
csv {
columns => ["Date","Units","ProductID","Country"]
separator => ","
}
mutate {
add_field => { "client_code" => "def" }
}
}
mutate
{
remove_field => [ "message" ]
}
}
output {
elasticsearch {
codec => json
hosts => "myelasticcluster.com:9200"
index => "sales_%{client_code}"
document_type => "sale"
}
stdout { codec => rubydebug }
}
An guidance from those well versed in Logstash configs would be much appreciated!