diff options
| author | Kelly Rauchenberger <fefferburbia@gmail.com> | 2016-02-13 23:43:13 -0500 |
|---|---|---|
| committer | Kelly Rauchenberger <fefferburbia@gmail.com> | 2016-02-13 23:43:13 -0500 |
| commit | d62c340f1841c6fc46968643fab63841083aec6f (patch) | |
| tree | b4f9cf6a3464bd4ac3882e7a99eb398b65580193 /kgramstats.cpp | |
| parent | cc8167a392eb435f657e1411d6bf6d7cbc6f3e3f (diff) | |
| download | rawr-ebooks-d62c340f1841c6fc46968643fab63841083aec6f.tar.gz rawr-ebooks-d62c340f1841c6fc46968643fab63841083aec6f.tar.bz2 rawr-ebooks-d62c340f1841c6fc46968643fab63841083aec6f.zip | |
Fixed issue where queries with both the wildcard token and a terminating token would reset the prefix
Diffstat (limited to 'kgramstats.cpp')
| -rw-r--r-- | kgramstats.cpp | 19 |
1 files changed, 5 insertions, 14 deletions
| diff --git a/kgramstats.cpp b/kgramstats.cpp index ac694f3..f788cb1 100644 --- a/kgramstats.cpp +++ b/kgramstats.cpp | |||
| @@ -344,21 +344,12 @@ kgramstats::kgramstats(std::string corpus, int maxK) | |||
| 344 | td.titlecase++; | 344 | td.titlecase++; |
| 345 | } | 345 | } |
| 346 | 346 | ||
| 347 | kgram term_prefix; | 347 | if (std::begin(prefix)->tok.suffix == suffixtype::terminating) |
| 348 | bool changed = false; | ||
| 349 | std::transform(prefix.begin(), prefix.end(), std::back_inserter(term_prefix), [&] (query& q) { | ||
| 350 | if (q.tok.suffix == suffixtype::terminating) | ||
| 351 | { | ||
| 352 | changed = true; | ||
| 353 | |||
| 354 | return wildcardQuery; | ||
| 355 | } else { | ||
| 356 | return q; | ||
| 357 | } | ||
| 358 | }); | ||
| 359 | |||
| 360 | if (changed) | ||
| 361 | { | 348 | { |
| 349 | kgram term_prefix(prefix); | ||
| 350 | term_prefix.pop_front(); | ||
| 351 | term_prefix.push_front(wildcardQuery); | ||
| 352 | |||
| 362 | if (tstats[term_prefix].count(f) == 0) | 353 | if (tstats[term_prefix].count(f) == 0) |
| 363 | { | 354 | { |
| 364 | tstats[term_prefix].emplace(f, f); | 355 | tstats[term_prefix].emplace(f, f); |