diff options
author | Jehan <jehan@girinstud.io> | 2022-12-15 14:31:10 +0100 |
---|---|---|
committer | Jehan <jehan@girinstud.io> | 2022-12-15 14:31:10 +0100 |
commit | 362086bf568974272f7cf654cf215d556fea7fc1 (patch) | |
tree | d583e826999b64d441d44d0447ee336c630bc41e | |
parent | 598fe90c910ec89b104f0d15c30868c0ed5793ad (diff) |
script: fix BuildLangModel.py.
-rwxr-xr-x | script/BuildLangModel.py | 10 |
1 files changed, 6 insertions, 4 deletions
diff --git a/script/BuildLangModel.py b/script/BuildLangModel.py index 7049ee7..684ece6 100755 --- a/script/BuildLangModel.py +++ b/script/BuildLangModel.py @@ -414,7 +414,7 @@ elif lang.alphabet is not None: logfd.write("\n[{:2}] Char {}: {} %".format(order, chr(char), ratio * 100)) accumulated_ratios += ratio freq_count += 1 - if accumulated_ratios < 0.4: + if very_freq_ratio < 0.4: very_freq_count += 1 very_freq_ratio += ratio else: @@ -440,14 +440,16 @@ elif lang.frequent_ranges is not None: logfd.write("\n[{:2}] Char {}: {} %".format(order, chr(char), ratio * 100)) frequent_ranges_size -= 1 break - if accumulated_ratios < 0.4: - very_freq_count += 1 - very_freq_ratio += ratio else: # A frequent character in the non-frequent range. logfd.write("\n[{:2}] Char {}: {} %".format(order, chr(char), ratio * 100)) freq_count += 1 accumulated_ratios += ratio + + if very_freq_ratio < 0.4: + very_freq_count += 1 + very_freq_ratio += ratio + if frequent_ranges_size <= 0: break |