feat(ebook): add protected phrase extraction library with config-driven tuning

Refactor protected phrase handling from a single module into a
python/ebook_search/protected_phrases package covering extraction,
storage, and runtime matching. Phrase filtering is now data-driven via
bundled TOML files: ignored_phrases, bad_starts, bad_ends, and
most_common_words.

Add phrase-tuning settings to EbookSearchConfig so candidate generation,
scoring, LLM judging, and matching are configurable rather than hardcoded:
token bounds, entity token limit, raw n-gram min count, frequency and
chapter-spread score thresholds, candidate/LLM/target caps, confidence
threshold, nesting defaults, and the phrase hit boost.

python/ebook_search/config.py

@@ -86,6 +86,21 @@ class EbookSearchConfig(BaseSettings):
     validate_citations_enabled: bool = True
     bm25_index_dir: str = ".ebook_search_bm25"
     bm25_refresh_delay_seconds: int = 60
+    protected_phrase_max_candidates_per_book: int = 5000
+    protected_phrase_llm_candidates_per_book: int = 500
+    protected_phrase_confidence_threshold: float = 0.80
+    phrase_hit_boost: float = 0.25
+    phrase_min_tokens: int = 2
+    phrase_max_tokens: int = 5
+    phrase_max_entity_tokens: int = 8
+    phrase_raw_ngram_min_count: int = 2
+    phrase_raw_count_score_threshold: int = 3
+    phrase_raw_count_high_score_threshold: int = 10
+    phrase_chapter_count_score_threshold: int = 2
+    phrase_chapter_count_high_score_threshold: int = 5
+    phrase_target_protected_per_book: int = 100
+    phrase_default_allow_nested: bool = False
+    phrase_default_suppress_children: bool = True
 
     @field_validator("library_paths", mode="before")
     @classmethod

python/ebook_search/protected_phrases/bad_ends.toml

@@ -0,0 +1,30 @@
+tokens = [
+  "a",
+  "an",
+  "and",
+  "any",
+  "as",
+  "at",
+  "be",
+  "because",
+  "but",
+  "by",
+  "can",
+  "could",
+  "do",
+  "for",
+  "from",
+  "have",
+  "if",
+  "of",
+  "or",
+  "some",
+  "than",
+  "the",
+  "these",
+  "this",
+  "to",
+  "will",
+  "with",
+  "would",
+]

python/ebook_search/protected_phrases/bad_starts.toml

@@ -0,0 +1,5 @@
+tokens = [
+    "a",
+    "an",
+    "the",
+]

python/ebook_search/protected_phrases/ignored_phrases.toml

@@ -0,0 +1,134 @@
+phrases = [
+  "ended up",
+  "had a",
+  "across the",
+  "anyone in",
+  "are you",
+  "around him",
+  "around the",
+  "as much",
+  "as soon",
+  "before the",
+  "behind him",
+  "between the",
+  "did you",
+  "end up",
+  "he asked",
+  "he concluded",
+  "he continued",
+  "he couldn't",
+  "he did",
+  "he didn't",
+  "he felt",
+  "he had",
+  "he hadn't",
+  "he knew",
+  "he noted",
+  "he pointed",
+  "he realized",
+  "he replied",
+  "he said",
+  "he saw",
+  "he tapped",
+  "he told",
+  "he was",
+  "he wasn't",
+  "his body",
+  "his chair",
+  "his feet",
+  "his hands",
+  "his head",
+  "his office",
+  "his own",
+  "his pc",
+  "his power",
+  "his shield",
+  "his sight",
+  "his voice",
+  "his wrist",
+  "how many",
+  "i am",
+  "i don't",
+  "i was",
+  "i wouldn't",
+  "i'm not",
+  "is in",
+  "is not",
+  "is that",
+  "is the",
+  "it had",
+  "it had",
+  "it was",
+  "it wasn't",
+  "it wasn't",
+  "of course",
+  "of force",
+  "of magic",
+  "of marines",
+  "of power",
+  "of those",
+  "old man",
+  "older man",
+  "set up",
+  "she admitted",
+  "she asked",
+  "she had",
+  "she replied",
+  "she said",
+  "she snapped",
+  "she told",
+  "she was",
+  "she'd been",
+  "shook his",
+  "sure he",
+  "tell you",
+  "that had",
+  "that is",
+  "that was",
+  "there are",
+  "there was",
+  "there were",
+  "they are",
+  "they had",
+  "they were",
+  "they weren't",
+  "this is",
+  "this place",
+  "though he",
+  "through his",
+  "through the",
+  "to find",
+  "to keep",
+  "to stay",
+  "to stop",
+  "to tell",
+  "to try",
+  "told her",
+  "told him",
+  "under his",
+  "was a",
+  "was enough",
+  "was going",
+  "was in",
+  "was no",
+  "was not",
+  "was now",
+  "was on",
+  "was one",
+  "was only",
+  "was still",
+  "was that",
+  "was the",
+  "was there",
+  "were in",
+  "what had",
+  "what happened",
+  "what was",
+  "where the",
+  "while i",
+  "you are",
+  "you can't",
+  "you don't",
+  "you need",
+  "you were",
+]

python/ebook_search/protected_phrases/most_common_words.toml

@@ -0,0 +1,102 @@
+words = [
+    "the",
+    "be",
+    "to",
+    "of",
+    "and",
+    "a",
+    "in",
+    "that",
+    "have",
+    "I",
+    "it",
+    "for",
+    "not",
+    "on",
+    "with",
+    "he",
+    "as",
+    "you",
+    "do",
+    "at",
+    "this",
+    "but",
+    "his",
+    "by",
+    "from",
+    "they",
+    "we",
+    "say",
+    "her",
+    "she",
+    "or",
+    "an",
+    "will",
+    "my",
+    "one",
+    "all",
+    "would",
+    "there",
+    "their",
+    "what",
+    "so",
+    "up",
+    "out",
+    "if",
+    "about",
+    "who",
+    "get",
+    "which",
+    "go",
+    "me",
+    "when",
+    "make",
+    "can",
+    "like",
+    "time",
+    "no",
+    "just",
+    "him",
+    "know",
+    "take",
+    "people",
+    "into",
+    "year",
+    "your",
+    "good",
+    "some",
+    "could",
+    "them",
+    "see",
+    "other",
+    "than",
+    "then",
+    "now",
+    "look",
+    "only",
+    "come",
+    "its",
+    "over",
+    "think",
+    "also",
+    "back",
+    "after",
+    "use",
+    "two",
+    "how",
+    "our",
+    "work",
+    "first",
+    "well",
+    "way",
+    "even",
+    "new",
+    "want",
+    "because",
+    "any",
+    "these",
+    "give",
+    "day",
+    "most",
+    "us",
+]