search for standard sites
pub-search.waow.tech
search
zig
blog
atproto
add prefix matching to FTS search
"cat" now matches "catch", "category", etc. by appending * to each search term.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
+59
-8
+59
-8
backend/src/db.zig
+59
-8
backend/src/db.zig
···
190
190
191
191
const temp_alloc = gpa.allocator();
192
192
193
-
// normalize query to match FTS5 tokenization (dots become spaces)
194
-
const normalized_query = try alloc.dupe(u8, query);
195
-
for (normalized_query) |*c| {
196
-
if (c.* == '.') c.* = ' ';
197
-
}
193
+
// normalize query: dots become spaces, add prefix matching with *
194
+
const fts_query = try buildFtsQuery(alloc, query);
198
195
199
196
var jw: json.Stringify = .{ .writer = &output.writer };
200
197
try jw.beginArray();
···
224
221
\\JOIN document_tags dt ON d.uri = dt.document_uri
225
222
\\WHERE documents_fts MATCH ? AND dt.tag = ?
226
223
\\ORDER BY rank LIMIT 40
227
-
, &.{ normalized_query, tag }) catch null
224
+
, &.{ fts_query, tag }) catch null
228
225
else
229
226
execSql(
230
227
\\SELECT f.uri, d.did, d.title,
···
236
233
\\LEFT JOIN publications p ON d.publication_uri = p.uri
237
234
\\WHERE documents_fts MATCH ?
238
235
\\ORDER BY rank LIMIT 40
239
-
, &.{normalized_query}) catch null;
236
+
, &.{fts_query}) catch null;
240
237
241
238
if (doc_result) |result| {
242
239
defer temp_alloc.free(result);
···
284
281
\\JOIN publications p ON f.uri = p.uri
285
282
\\WHERE publications_fts MATCH ?
286
283
\\ORDER BY rank LIMIT 10
287
-
, &.{normalized_query}) catch null;
284
+
, &.{fts_query}) catch null;
288
285
289
286
if (pub_result) |result| {
290
287
defer temp_alloc.free(result);
···
361
358
},
362
359
else => 0,
363
360
};
361
+
}
362
+
363
+
// build FTS5 query with prefix matching: "cat dog" -> "cat* dog*"
364
+
fn buildFtsQuery(alloc: Allocator, query: []const u8) ![]const u8 {
365
+
if (query.len == 0) return "";
366
+
367
+
// normalize dots to spaces
368
+
const normalized = try alloc.dupe(u8, query);
369
+
for (normalized) |*c| {
370
+
if (c.* == '.') c.* = ' ';
371
+
}
372
+
373
+
// count words to calculate output size
374
+
var word_count: usize = 0;
375
+
var in_word = false;
376
+
for (normalized) |c| {
377
+
if (c == ' ') {
378
+
in_word = false;
379
+
} else if (!in_word) {
380
+
word_count += 1;
381
+
in_word = true;
382
+
}
383
+
}
384
+
385
+
if (word_count == 0) return "";
386
+
387
+
// allocate: original length + one '*' per word + spaces
388
+
const result = try alloc.alloc(u8, normalized.len + word_count);
389
+
var pos: usize = 0;
390
+
in_word = false;
391
+
392
+
for (normalized) |c| {
393
+
if (c == ' ') {
394
+
if (in_word) {
395
+
result[pos] = '*';
396
+
pos += 1;
397
+
in_word = false;
398
+
}
399
+
result[pos] = ' ';
400
+
pos += 1;
401
+
} else {
402
+
result[pos] = c;
403
+
pos += 1;
404
+
in_word = true;
405
+
}
406
+
}
407
+
408
+
// add final * if ended in a word
409
+
if (in_word) {
410
+
result[pos] = '*';
411
+
pos += 1;
412
+
}
413
+
414
+
return result[0..pos];
364
415
}
365
416
366
417
fn execSql(sql: []const u8, args: []const []const u8) ![]const u8 {