taler-docs

Documentation for GNU Taler components, APIs and protocols
Log | Files | Refs | README | LICENSE

commit 179c07b75471025ed89f14a51650f2632723a537
parent 11f992c7a5bcc74c9c638e27826ce94250dc1671
Author: Christian Grothoff <christian@grothoff.org>
Date:   Tue,  3 Jun 2025 16:17:38 +0200

new sanctionscheck options

Diffstat:
Mmanpages/taler-exchange.conf.5.rst | 29++++++++++++++++++++++++++++-
1 file changed, 28 insertions(+), 1 deletion(-)

diff --git a/manpages/taler-exchange.conf.5.rst b/manpages/taler-exchange.conf.5.rst @@ -672,6 +672,11 @@ UNIXPATH EXCHANGE SANCTIONSCHECK OPTIONS ------------------------------- +This section configures the taler-exchange-sanctionlist tool to enforce +sanctions. The tool scores KYC records in two dimensions, the quality +of the match (rating) and the number of attributes that were found in +both sets (confidence). + The following options must be in the section "[exchange-sanctionscheck]". MIN_ROW_FILENAME @@ -681,11 +686,33 @@ MIN_ROW_FILENAME transactionality is needed or even desirable. There should be no need to change the default. +FREEZE_RATING_LIMIT + Match quality rating (on a scale of [0,1]) that must be exceeded + for the exchange to automatically freeze an account. The account + must satisfy both the rating limit and the confidence limit to be + frozen. + +FREEZE_CONFIDENCE_LIMIT + Match confidence rating (on a scale of [0,1]) that must be exceeded + for the exchange to automatically freeze an account. A rating of 0 + means no attributes were available, while a rating of 1 indicates that + we could compare the acconts across all dimensions. + +INVESTIGATION_LIMIT + Limit (on a scale of [0,1]) that must be exceeded for an account to + be flagged for manual investigation. Only applies of the account + is not automatically frozen. The limit given is compared against + the quality rating divided by the confidence score. + If set at 0.95, a modest 0.8 match with low 0.5 confidence will + trigger (1.6 > 0.95), but a good 0.9 match with a super-high 1.0 + confidence would not (0.9 < 0.95). On the other hand, a bad 0.2 match + with super-low 0.1 confidence would again trigger (2.0 > 0.95). + SEE ALSO ======== -taler-exchange-dbinit(1), taler-exchange-httpd(1), taler-exchange-offline(1) +taler-exchange-dbinit(1), taler-exchange-httpd(1), taler-exchange-offline(1), taler-exchange-sanctionscheck(1) BUGS ====