From 8e2ca365c2de9e222b248ed1701ebf91a95ad6ed Mon Sep 17 00:00:00 2001 From: Richie Cahill Date: Fri, 3 Jul 2026 18:26:12 -0400 Subject: [PATCH] feat(ebook): update protected phrases with additional tokens and phrases --- .../protected_phrases/bad_ends.toml | 1 + .../protected_phrases/bad_starts.toml | 22 +++++ .../protected_phrases/ignored_phrases.toml | 82 ++++++++++++++++++- 3 files changed, 103 insertions(+), 2 deletions(-) diff --git a/python/ebook_search/protected_phrases/bad_ends.toml b/python/ebook_search/protected_phrases/bad_ends.toml index a2c9c23..1a8f600 100644 --- a/python/ebook_search/protected_phrases/bad_ends.toml +++ b/python/ebook_search/protected_phrases/bad_ends.toml @@ -27,4 +27,5 @@ tokens = [ "will", "with", "would", + "did", ] diff --git a/python/ebook_search/protected_phrases/bad_starts.toml b/python/ebook_search/protected_phrases/bad_starts.toml index a3f0b55..bc92079 100644 --- a/python/ebook_search/protected_phrases/bad_starts.toml +++ b/python/ebook_search/protected_phrases/bad_starts.toml @@ -1,5 +1,27 @@ tokens = [ "a", "an", + "did", + "didn't", + "he", + "here", + "how", + "i", + "it", + "she", + "that", "the", + "there", + "they", + "this", + "we", + "what", + "when", + "where", + "which", + "who", + "whom", + "whose", + "why", + "you", ] diff --git a/python/ebook_search/protected_phrases/ignored_phrases.toml b/python/ebook_search/protected_phrases/ignored_phrases.toml index 35c344f..801d209 100644 --- a/python/ebook_search/protected_phrases/ignored_phrases.toml +++ b/python/ebook_search/protected_phrases/ignored_phrases.toml @@ -1,18 +1,71 @@ phrases = [ - "ended up", - "had a", + "a little", "across the", + "and she", "anyone in", "are you", "around him", "around the", "as much", "as soon", + "at all", + "at least", "before the", "behind him", "between the", + "but she", + "could not", + "did he", + "did i", + "did it", + "did not believe", + "did not care", + "did not even", + "did not know what", + "did not know", + "did not like", + "did not look", + "did not mean", + "did not move", + "did not need", + "did not see", + "did not seem", + "did not think", + "did not understand", + "did not want", + "did not", + "did she", + "did so", + "did that", + "did the", + "did they", + "did what", "did you", + "didn't answer", + "didn't care", + "didn't even", + "didn't expect", + "didn't feel", + "didn't get", + "didn't i", + "didn't know", + "didn't like", + "didn't look", + "didn't make", + "didn't mean", + "didn't need", + "didn't really", + "didn't say", + "didn't see", + "didn't seem", + "didn't think", + "didn't want", + "didn't you", "end up", + "ended up", + "had a", + "had been", + "have been", "he asked", "he concluded", "he continued", @@ -49,26 +102,34 @@ phrases = [ "how many", "i am", "i don't", + "i said", "i was", "i wouldn't", "i'm not", + "if he", + "if they", "is in", "is not", "is that", "is the", "it had", "it had", + "it is", "it was", "it wasn't", "it wasn't", + "no one", "of course", "of force", + "of it", "of magic", "of marines", "of power", "of those", "old man", "older man", + "one of", + "out of", "set up", "she admitted", "she asked", @@ -84,7 +145,22 @@ phrases = [ "tell you", "that had", "that is", + "that she", "that was", + "the dark", + "the door", + "the first", + "the last", + "the man", + "the one", + "the only", + "the other", + "the rest", + "the room", + "the same", + "the two", + "the way", + "the world", "there are", "there was", "there were", @@ -98,6 +174,7 @@ phrases = [ "through his", "through the", "to find", + "to get", "to keep", "to stay", "to stop", @@ -129,6 +206,7 @@ phrases = [ "you are", "you can't", "you don't", + "you know", "you need", "you were", ]