On Thu, Oct 24, 2024 at 02:58:00AM +1100, Viktor Dukhovni via mailop wrote:
> The below works if it is not too costly to hold the entire thing in
> memory. It can be optimised for streaming, but the "jq" script becomes
> harder to understand.
>
The streaming version (execute as: jq --stream -nr -f script.jq
input.json) is below. This also filters to the values prior to output,
rejecting data with embedded control characters or addresses with
something other than hex digits '.' or ':' (plausibly, but not
necessarily v4 or v6 addresses).
The sample input therefore needs more realistic addresses (no XXX).
--
Viktor.
def init: { ix: -1, wip: {} };
def update($pv):
$pv[0][0] as $ix
| $pv[0][1:] as $p
| $pv[1] as $v
| if .ix != $ix or $p == []
then .done = .wip | .wip = { addrs: [] }
else .done = {}
end
| .ix = $ix
| if ([$v, $p[-1]] | map(type)) == ["string", "string"]
then if $p[0] == "client"
then .wip |= setpath($p; $v)
elif $p[0] == "ips" and $p[-2] == "active"
then .wip.addrs += [[$p[-1], $v]]
else .
end
else .
end;
def extract:
.done
| select(.client.name? | type == "string" and . != "")
| select(.client.complaint_address? | type == "string" and . != "")
| select(.addrs | type == "array" and . != [])
| select(.client.name | test("^[ -~]*$"))
| select(.client.complaint_address | test("^[ -~]*$"))
| select(.addrs | all(.[0] | test("^[a-fA-F0-9.:]*$")))
| select(.addrs | all(.[1] | test("^[ -~]*$")))
| ( "## Name: \(.client.name)"
, "## Abuse: \(.client.complaint_address)"
, ( .addrs[]
| ( "# \(.[1])"
, "\(.[0]) permit"
)
)
, ""
);
foreach inputs as $pathval (init; update($pathval); extract)
_______________________________________________
mailop mailing list
mailop@mailop.org
https://list.mailop.org/listinfo/mailop
