diff options
Diffstat (limited to 'kgramstats.h')
-rw-r--r-- | kgramstats.h | 12 |
1 files changed, 9 insertions, 3 deletions
diff --git a/kgramstats.h b/kgramstats.h index 069bb90..248b193 100644 --- a/kgramstats.h +++ b/kgramstats.h | |||
@@ -14,15 +14,21 @@ class kgramstats | |||
14 | { | 14 | { |
15 | public: | 15 | public: |
16 | kgramstats(string corpus, int maxK); | 16 | kgramstats(string corpus, int maxK); |
17 | map<string, int>* lookupExts(kgram tk); | ||
18 | int getMaxK(); | ||
19 | vector<string> randomSentence(int n); | 17 | vector<string> randomSentence(int n); |
20 | 18 | ||
21 | private: | 19 | private: |
20 | typedef struct | ||
21 | { | ||
22 | int all; | ||
23 | int titlecase; | ||
24 | int uppercase; | ||
25 | int period; | ||
26 | } token_data; | ||
22 | int maxK; | 27 | int maxK; |
23 | map<kgram, map<string, int>* >* stats; | 28 | map<kgram, map<string, token_data*>* >* stats; |
24 | }; | 29 | }; |
25 | 30 | ||
26 | void printKgram(kgram k); | 31 | void printKgram(kgram k); |
32 | std::string canonize(std::string f); | ||
27 | 33 | ||
28 | #endif \ No newline at end of file | 34 | #endif \ No newline at end of file |