From f83cd2a05ac84ef4bf28f75c763b0688df3f4201 Mon Sep 17 00:00:00 2001 From: David Bremner Date: Tue, 24 Aug 2021 08:17:19 -0700 Subject: [PATCH] lib/parse-sexp: support and, not, and or. All operations and (Xapian) fields will eventually have an entry in the prefixes table. The flags field is just a placeholder for now, but will eventually distinguish between various kinds of prefixes. --- doc/man7/notmuch-sexp-queries.rst | 16 ++++--- lib/parse-sexp.cc | 76 +++++++++++++++++++++++++++++-- test/T081-sexpr-search.sh | 31 +++++++++++-- 3 files changed, 109 insertions(+), 14 deletions(-) diff --git a/doc/man7/notmuch-sexp-queries.rst b/doc/man7/notmuch-sexp-queries.rst index d177934d..0304759e 100644 --- a/doc/man7/notmuch-sexp-queries.rst +++ b/doc/man7/notmuch-sexp-queries.rst @@ -51,7 +51,9 @@ subqueries. (for most fields) or *or*. See :any:`fields` for more information. ``(`` *operator* |q1| |q2| ... |qn| ``)`` - Combine queries |q1| to |qn|. See :any:`operators` for more information. + Combine queries |q1| to |qn|. Currently supported operators are + ``and``, ``or``, and ``not``. ``(not`` |q1| ... |qn| ``)`` is equivalent + to ``(and (not`` |q1| ``) ... (not`` |qn| ``))``. ``(`` *modifier* |q1| |q2| ... |qn| ``)`` Combine queries |q1| to |qn|, and reinterpret the result (e.g. as a regular expression). @@ -62,11 +64,6 @@ subqueries. FIELDS `````` -.. _operators: - -OPERATORS -````````` - .. _modifiers: MODIFIERS @@ -82,6 +79,13 @@ EXAMPLES Match all messages containing "added", but also those containing "add", "additional", "Additional", "adds", etc... via stemming. +``(and Bob Marley)`` + Match messages containing words "Bob" and "Marley", or their stems + The words need not be adjacent. + +``(not Bob Marley)`` + Match messages containing neither "Bob" nor "Marley", nor their stems, + .. |q1| replace:: :math:`q_1` .. |q2| replace:: :math:`q_2` .. |qn| replace:: :math:`q_n` diff --git a/lib/parse-sexp.cc b/lib/parse-sexp.cc index f031d790..0d2c0ba8 100644 --- a/lib/parse-sexp.cc +++ b/lib/parse-sexp.cc @@ -7,12 +7,69 @@ /* _sexp is used for file scope symbols to avoid clashing with * definitions from sexp.h */ +typedef enum { + SEXP_FLAG_NONE = 0, +} _sexp_flag_t; + +typedef struct { + const char *name; + Xapian::Query::op xapian_op; + Xapian::Query initial; + _sexp_flag_t flags; +} _sexp_prefix_t; + +static _sexp_prefix_t prefixes[] = +{ + { "and", Xapian::Query::OP_AND, Xapian::Query::MatchAll, + SEXP_FLAG_NONE }, + { "not", Xapian::Query::OP_AND_NOT, Xapian::Query::MatchAll, + SEXP_FLAG_NONE }, + { "or", Xapian::Query::OP_OR, Xapian::Query::MatchNothing, + SEXP_FLAG_NONE }, + { } +}; + +static notmuch_status_t _sexp_to_xapian_query (notmuch_database_t *notmuch, + const _sexp_prefix_t *parent, + const sexp_t *sx, + Xapian::Query &output); + +static notmuch_status_t +_sexp_combine_query (notmuch_database_t *notmuch, + const _sexp_prefix_t *parent, + Xapian::Query::op operation, + Xapian::Query left, + const sexp_t *sx, + Xapian::Query &output) +{ + Xapian::Query subquery; + + notmuch_status_t status; + + /* if we run out elements, return accumulator */ + + if (! sx) { + output = left; + return NOTMUCH_STATUS_SUCCESS; + } + + status = _sexp_to_xapian_query (notmuch, parent, sx, subquery); + if (status) + return status; + + return _sexp_combine_query (notmuch, + parent, + operation, + Xapian::Query (operation, left, subquery), + sx->next, output); +} + /* Here we expect the s-expression to be a proper list, with first * element defining and operation, or as a special case the empty * list */ static notmuch_status_t -_sexp_to_xapian_query (notmuch_database_t *notmuch, const sexp_t *sx, +_sexp_to_xapian_query (notmuch_database_t *notmuch, const _sexp_prefix_t *parent, const sexp_t *sx, Xapian::Query &output) { @@ -32,11 +89,20 @@ _sexp_to_xapian_query (notmuch_database_t *notmuch, const sexp_t *sx, return NOTMUCH_STATUS_SUCCESS; } - if (sx->list->ty == SEXP_VALUE) - _notmuch_database_log (notmuch, "unknown prefix '%s'\n", sx->list->val); - else + if (sx->list->ty == SEXP_LIST) { _notmuch_database_log (notmuch, "unexpected list in field/operation position\n", sx->list->val); + return NOTMUCH_STATUS_BAD_QUERY_SYNTAX; + } + + for (_sexp_prefix_t *prefix = prefixes; prefix && prefix->name; prefix++) { + if (strcmp (prefix->name, sx->list->val) == 0) { + return _sexp_combine_query (notmuch, parent, prefix->xapian_op, prefix->initial, + sx->list->next, output); + } + } + + _notmuch_database_log (notmuch, "unknown prefix '%s'\n", sx->list->val); return NOTMUCH_STATUS_BAD_QUERY_SYNTAX; } @@ -54,6 +120,6 @@ _notmuch_sexp_string_to_xapian_query (notmuch_database_t *notmuch, const char *q return NOTMUCH_STATUS_BAD_QUERY_SYNTAX; } - return _sexp_to_xapian_query (notmuch, sx, output); + return _sexp_to_xapian_query (notmuch, NULL, sx, output); } #endif diff --git a/test/T081-sexpr-search.sh b/test/T081-sexpr-search.sh index 4e3d755c..5e1bb18d 100755 --- a/test/T081-sexpr-search.sh +++ b/test/T081-sexpr-search.sh @@ -9,9 +9,34 @@ fi add_email_corpus -test_begin_subtest "all messages: ()" -notmuch search '*' > EXPECTED -notmuch search --query=sexp "()" > OUTPUT +for query in '()' '(not)' '(and)' '(or ())' '(or (not))' '(or (and))' \ + '(or (and) (or) (not (and)))'; do + test_begin_subtest "all messages: $query" + notmuch search '*' > EXPECTED + notmuch search --query=sexp "$query" > OUTPUT + test_expect_equal_file EXPECTED OUTPUT +done + +for query in '(or)' '(not ())' '(not (not))' '(not (and))' \ + '(not (or (and) (or) (not (and))))'; do + test_begin_subtest "no messages: $query" + notmuch search --query=sexp "$query" > OUTPUT + test_expect_equal_file /dev/null OUTPUT +done + +test_begin_subtest "and of exact terms" +notmuch search --query=sexp '(and "wonderful" "wizard")' | notmuch_search_sanitize > OUTPUT +cat < EXPECTED +thread:XXX 2009-11-18 [1/3] Carl Worth| Jan Janak; [notmuch] What a great idea! (inbox unread) +EOF +test_expect_equal_file EXPECTED OUTPUT + +test_begin_subtest "or of exact terms" +notmuch search --query=sexp '(or "php" "wizard")' | notmuch_search_sanitize > OUTPUT +cat < EXPECTED +thread:XXX 2010-12-29 [1/1] François Boulogne; [aur-general] Guidelines: cp, mkdir vs install (inbox unread) +thread:XXX 2009-11-18 [1/3] Carl Worth| Jan Janak; [notmuch] What a great idea! (inbox unread) +EOF test_expect_equal_file EXPECTED OUTPUT test_begin_subtest "single term in body"