pystylometry 0.1.0__py3-none-any.whl → 1.0.0__py3-none-any.whl

pystylometry/readability/additional_formulas.py

@@ -0,0 +1,985 @@
+"""Additional readability formulas.
+
+This module provides additional readability metrics beyond the core formulas
+(Flesch, SMOG, Gunning Fog, Coleman-Liau, ARI). These formulas offer alternative
+approaches to measuring text difficulty and are valuable for cross-validation
+and comprehensive readability assessment.
+
+Related GitHub Issue:
+    #16 - Additional Readability Formulas
+    https://github.com/craigtrim/pystylometry/issues/16
+
+Formulas implemented:
+    - Dale-Chall: Based on list of 3000 familiar words
+    - Linsear Write: Developed for technical writing assessment
+    - Fry Readability Graph: Visual graph-based assessment
+    - FORCAST: Military formula using only single-syllable words
+    - Powers-Sumner-Kearl: Recalibrated Flesch for primary grades
+
+References:
+    Dale, E., & Chall, J. S. (1948). A formula for predicting readability.
+    Chall, J. S., & Dale, E. (1995). Readability revisited: The new Dale-Chall
+        readability formula. Brookline Books.
+    Klare, G. R. (1974-1975). Assessing readability. Reading Research Quarterly.
+    Fry, E. (1968). A readability formula that saves time. Journal of Reading.
+    Caylor, J. S., et al. (1973). Methodologies for determining reading requirements
+        of military occupational specialties. Human Resources Research Organization.
+    Powers, R. D., Sumner, W. A., & Kearl, B. E. (1958). A recalculation of four
+        adult readability formulas. Journal of Educational Psychology.
+"""
+
+from .._normalize import normalize_for_readability
+from .._types import (
+    DaleChallResult,
+    FORCASTResult,
+    FryResult,
+    LinsearWriteResult,
+    PowersSumnerKearlResult,
+)
+from .._utils import split_sentences, tokenize
+from .syllables import count_syllables
+
+
+# Dale-Chall List of Familiar Words (subset of ~1200 words)
+# GitHub Issue #16: https://github.com/craigtrim/pystylometry/issues/16
+# Full Dale-Chall list has 3000 words that 80% of 4th graders understand.
+# This is a representative subset covering most common everyday words.
+DALE_CHALL_FAMILIAR_WORDS = {
+    # Articles, pronouns, determiners
+    "a", "an", "the", "this", "that", "these", "those", "some", "any", "all",
+    "each", "every", "both", "few", "many", "much", "more", "most", "other",
+    "another", "such", "what", "which", "who", "whom", "whose", "whoever",
+    "i", "me", "my", "mine", "myself", "we", "us", "our", "ours", "ourselves",
+    "you", "your", "yours", "yourself", "yourselves",
+    "he", "him", "his", "himself", "she", "her", "hers", "herself",
+    "it", "its", "itself", "they", "them", "their", "theirs", "themselves",
+    "one", "ones", "someone", "somebody", "something", "anyone", "anybody", "anything",
+    "everyone", "everybody", "everything", "no", "none", "nobody", "nothing",
+
+    # Conjunctions and prepositions
+    "and", "or", "but", "if", "when", "where", "why", "how", "because", "so",
+    "for", "nor", "yet", "after", "before", "while", "since", "until", "unless",
+    "though", "although", "whether", "than", "as", "like",
+    "of", "to", "in", "on", "at", "by", "with", "from", "about", "into",
+    "through", "over", "under", "above", "below", "between", "among", "against",
+    "during", "without", "within", "along", "across", "behind", "beside", "near",
+    "off", "out", "up", "down", "around", "past", "toward", "upon",
+
+    # Common verbs (base, past, -ing, -ed forms included)
+    "be", "am", "is", "are", "was", "were", "been", "being",
+    "have", "has", "had", "having", "do", "does", "did", "doing", "done",
+    "will", "would", "shall", "should", "may", "might", "must", "can", "could",
+    "go", "goes", "went", "gone", "going", "come", "comes", "came", "coming",
+    "make", "makes", "made", "making", "get", "gets", "got", "getting", "gotten",
+    "know", "knows", "knew", "known", "knowing",
+    "think", "thinks", "thought", "thinking",
+    "see", "sees", "saw", "seen", "seeing", "look", "looks", "looked", "looking",
+    "take", "takes", "took", "taken", "taking", "give", "gives", "gave", "given", "giving",
+    "find", "finds", "found", "finding", "tell", "tells", "told", "telling",
+    "ask", "asks", "asked", "asking", "work", "works", "worked", "working",
+    "seem", "seems", "seemed", "seeming", "feel", "feels", "felt", "feeling",
+    "try", "tries", "tried", "trying", "leave", "leaves", "left", "leaving",
+    "call", "calls", "called", "calling", "use", "uses", "used", "using",
+    "want", "wants", "wanted", "wanting", "need", "needs", "needed", "needing",
+    "say", "says", "said", "saying", "talk", "talks", "talked", "talking",
+    "turn", "turns", "turned", "turning", "run", "runs", "ran", "running",
+    "move", "moves", "moved", "moving", "live", "lives", "lived", "living",
+    "believe", "believes", "believed", "believing",
+    "hold", "holds", "held", "holding", "bring", "brings", "brought", "bringing",
+    "happen", "happens", "happened", "happening",
+    "write", "writes", "wrote", "written", "writing",
+    "sit", "sits", "sat", "sitting", "stand", "stands", "stood", "standing",
+    "hear", "hears", "heard", "hearing", "let", "lets", "letting",
+    "help", "helps", "helped", "helping", "show", "shows", "showed", "shown", "showing",
+    "play", "plays", "played", "playing", "read", "reads", "reading",
+    "change", "changes", "changed", "changing", "keep", "keeps", "kept", "keeping",
+    "start", "starts", "started", "starting", "stop", "stops", "stopped", "stopping",
+    "learn", "learns", "learned", "learning", "grow", "grows", "grew", "grown", "growing",
+    "open", "opens", "opened", "opening", "close", "closes", "closed", "closing",
+    "walk", "walks", "walked", "walking", "win", "wins", "won", "winning",
+    "begin", "begins", "began", "begun", "beginning", "end", "ends", "ended", "ending",
+    "lose", "loses", "lost", "losing", "send", "sends", "sent", "sending",
+    "buy", "buys", "bought", "buying", "pay", "pays", "paid", "paying",
+    "eat", "eats", "ate", "eaten", "eating", "drink", "drinks", "drank", "drinking",
+    "sleep", "sleeps", "slept", "sleeping", "wake", "wakes", "woke", "waking",
+    "sing", "sings", "sang", "sung", "singing", "dance", "dances", "danced", "dancing",
+    "wait", "waits", "waited", "waiting", "stay", "stays", "stayed", "staying",
+    "fly", "flies", "flew", "flown", "flying", "fall", "falls", "fell", "fallen", "falling",
+    "cut", "cuts", "cutting", "break", "breaks", "broke", "broken", "breaking",
+    "watch", "watches", "watched", "watching", "listen", "listens", "listened", "listening",
+    "remember", "remembers", "remembered", "remembering",
+    "forget", "forgets", "forgot", "forgotten", "forgetting",
+    "meet", "meets", "met", "meeting", "follow", "follows", "followed", "following",
+    "carry", "carries", "carried", "carrying", "catch", "catches", "caught", "catching",
+    "draw", "draws", "drew", "drawn", "drawing", "drive", "drives", "drove", "driven", "driving",
+    "ride", "rides", "rode", "ridden", "riding", "wear", "wears", "wore", "worn", "wearing",
+    "pull", "pulls", "pulled", "pulling", "push", "pushes", "pushed", "pushing",
+    "throw", "throws", "threw", "thrown", "throwing",
+    "reach", "reaches", "reached", "reaching", "pass", "passes", "passed", "passing",
+    "shoot", "shoots", "shot", "shooting", "rise", "rises", "rose", "risen", "rising",
+    "blow", "blows", "blew", "blown", "blowing", "grow", "grows", "grew", "grown", "growing",
+    "hit", "hits", "hitting", "fight", "fights", "fought", "fighting",
+    "die", "dies", "died", "dying", "kill", "kills", "killed", "killing",
+    "speak", "speaks", "spoke", "spoken", "speaking",
+
+    # Common nouns
+    "time", "times", "year", "years", "day", "days", "week", "weeks",
+    "month", "months", "hour", "hours", "minute", "minutes", "second", "seconds",
+    "morning", "afternoon", "evening", "night", "today", "yesterday", "tomorrow",
+    "people", "person", "man", "men", "woman", "women", "child", "children",
+    "boy", "boys", "girl", "girls", "baby", "babies", "friend", "friends",
+    "family", "families", "mother", "father", "parent", "parents",
+    "brother", "brothers", "sister", "sisters", "son", "daughter",
+    "place", "places", "home", "house", "houses", "room", "rooms",
+    "school", "schools", "class", "classes", "student", "students", "teacher", "teachers",
+    "way", "ways", "thing", "things", "part", "parts", "group", "groups",
+    "number", "numbers", "side", "sides", "kind", "kinds", "head", "heads",
+    "hand", "hands", "eye", "eyes", "face", "faces", "body", "bodies",
+    "foot", "feet", "arm", "arms", "leg", "legs", "ear", "ears", "mouth",
+    "water", "food", "air", "land", "earth", "ground", "world",
+    "country", "countries", "state", "states", "city", "cities", "town", "towns",
+    "name", "names", "word", "words", "line", "lines", "page", "pages",
+    "book", "books", "story", "stories", "letter", "letters", "paper", "papers",
+    "point", "points", "end", "ends", "top", "bottom", "front", "back",
+    "life", "lives", "problem", "problems", "question", "questions", "answer", "answers",
+    "work", "works", "job", "jobs", "money", "door", "doors", "window", "windows",
+    "car", "cars", "road", "roads", "street", "streets", "tree", "trees",
+    "animal", "animals", "bird", "birds", "fish", "dog", "dogs", "cat", "cats",
+    "horse", "horses", "sea", "mountain", "mountains", "river", "rivers",
+    "sun", "moon", "star", "stars", "sky", "cloud", "clouds", "rain", "snow",
+    "wind", "fire", "light", "dark", "sound", "sounds", "color", "colors",
+    "white", "black", "red", "blue", "green", "yellow", "brown", "orange",
+    "game", "games", "ball", "music", "song", "songs", "picture", "pictures",
+    "table", "tables", "chair", "chairs", "bed", "beds", "floor", "wall", "walls",
+    "minute", "power", "war", "force", "age", "care", "order", "case",
+
+    # Common adjectives
+    "good", "better", "best", "bad", "worse", "worst",
+    "big", "bigger", "biggest", "small", "smaller", "smallest",
+    "large", "larger", "largest", "little", "less", "least",
+    "long", "longer", "longest", "short", "shorter", "shortest",
+    "high", "higher", "highest", "low", "lower", "lowest",
+    "old", "older", "oldest", "young", "younger", "youngest", "new", "newer", "newest",
+    "great", "greater", "greatest", "important", "right", "left", "own",
+    "other", "different", "same", "next", "last", "first", "second", "third",
+    "early", "earlier", "earliest", "late", "later", "latest",
+    "easy", "easier", "easiest", "hard", "harder", "hardest",
+    "hot", "hotter", "hottest", "cold", "colder", "coldest",
+    "warm", "warmer", "warmest", "cool", "cooler", "coolest",
+    "fast", "faster", "fastest", "slow", "slower", "slowest",
+    "strong", "stronger", "strongest", "weak", "weaker", "weakest",
+    "happy", "happier", "happiest", "sad", "sadder", "saddest",
+    "nice", "nicer", "nicest", "kind", "kinder", "kindest",
+    "sure", "free", "full", "whole", "ready", "simple", "clear",
+    "real", "true", "certain", "public", "able", "several",
+    "open", "closed", "deep", "wide", "bright", "dark", "heavy", "light",
+    "clean", "dirty", "wet", "dry", "soft", "hard", "quiet", "loud",
+    "quick", "slow", "rich", "poor", "sick", "well", "dead", "alive",
+    "empty", "busy", "pretty", "beautiful", "ugly",
+
+    # Common adverbs
+    "very", "too", "so", "more", "most", "less", "least",
+    "well", "better", "best", "just", "only", "even", "still",
+    "also", "just", "now", "then", "here", "there", "where",
+    "how", "when", "why", "not", "never", "always", "often",
+    "sometimes", "usually", "ever", "again", "back", "away",
+    "together", "once", "twice", "soon", "today", "yesterday", "tomorrow",
+    "already", "almost", "enough", "quite", "rather", "really",
+    "perhaps", "maybe", "probably", "certainly", "surely",
+    "yes", "no", "please", "thank", "sorry",
+
+    # Numbers
+    "zero", "one", "two", "three", "four", "five", "six", "seven", "eight", "nine", "ten",
+    "eleven", "twelve", "thirteen", "fourteen", "fifteen", "sixteen", "seventeen", "eighteen", "nineteen", "twenty",
+    "thirty", "forty", "fifty", "sixty", "seventy", "eighty", "ninety",
+    "hundred", "thousand", "million",
+    "first", "second", "third", "fourth", "fifth", "sixth", "seventh", "eighth", "ninth", "tenth",
+
+    # Additional common words
+    "able", "accept", "across", "act", "add", "afraid", "against", "agree",
+    "allow", "alone", "appear", "apple", "area", "arm", "arrive", "art",
+    "aunt", "ball", "become", "believe", "belong", "boat", "build",
+    "burn", "business", "chair", "chance", "church", "clear", "climb",
+    "clothe", "clothes", "company", "contain", "continue", "control",
+    "cook", "corner", "cost", "count", "course", "cover", "create",
+    "cross", "crowd", "cry", "decide", "depend", "describe", "develop",
+    "die", "direction", "discover", "doctor", "double", "drop", "during",
+    "edge", "effect", "eight", "either", "else", "enjoy", "enough",
+    "enter", "example", "except", "excite", "expect", "explain", "express",
+    "fact", "fair", "farm", "fear", "field", "fill", "final", "fine",
+    "finger", "finish", "flower", "force", "foreign", "forest", "form",
+    "fresh", "front", "garden", "general", "glass", "god", "gold",
+    "hang", "hat", "hope", "hot", "idea", "include", "increase",
+    "instead", "interest", "island", "join", "laugh", "law", "lead",
+    "lie", "lift", "list", "lock", "love", "machine", "mark",
+    "matter", "mean", "measure", "member", "mention", "middle", "mile",
+    "mind", "miss", "moment", "nation", "natural", "nature", "necessary",
+    "neighbor", "notice", "object", "ocean", "offer", "office", "opinion",
+    "paint", "pair", "party", "pattern", "period", "pick", "plan",
+    "plant", "position", "possible", "pound", "prepare", "present", "president",
+    "press", "prince", "print", "probable", "produce", "promise", "proper",
+    "protect", "prove", "purpose", "quarter", "queen", "question", "quick",
+    "quiet", "race", "raise", "range", "rate", "reason", "receive",
+    "record", "region", "remain", "reply", "report", "represent", "require",
+    "rest", "result", "return", "roll", "rule", "sail", "salt",
+    "save", "science", "season", "seat", "seem", "sell", "sense",
+    "sentence", "separate", "serve", "set", "settle", "seven", "shape",
+    "share", "ship", "shore", "sign", "silver", "single", "sir",
+    "six", "size", "skin", "soldier", "solve", "south", "space",
+    "special", "speed", "spell", "spend", "spread", "spring", "square",
+    "step", "stone", "straight", "strange", "stream", "strength", "strike",
+    "subject", "success", "sudden", "suffer", "suggest", "suit", "summer",
+    "supply", "support", "suppose", "surface", "surprise", "sweet", "swim",
+    "system", "tail", "taste", "teach", "team", "telephone", "television",
+    "temperature", "ten", "test", "thick", "thin", "though", "thousand",
+    "three", "tire", "total", "touch", "track", "train", "travel",
+    "trip", "trouble", "type", "uncle", "understand", "unit", "universe",
+    "value", "various", "view", "village", "visit", "voice", "vote",
+    "wagon", "wander", "warm", "wash", "wave", "wealth", "weather",
+    "weight", "welcome", "west", "wheel", "wild", "wind", "winter",
+    "wish", "wonder", "wood", "yard", "yellow",
+}
+
+
+def compute_dale_chall(text: str) -> DaleChallResult:
+    """
+    Compute Dale-Chall Readability Formula.
+
+    The Dale-Chall formula estimates reading difficulty based on the percentage
+    of words that are NOT on a list of 3000 familiar words (words understood
+    by 80% of 4th graders). It also considers average sentence length.
+
+    Related GitHub Issue:
+        #16 - Additional Readability Formulas
+        https://github.com/craigtrim/pystylometry/issues/16
+
+    Formula:
+        Raw Score = 0.1579 * (difficult_words_pct) + 0.0496 * (avg_sentence_length)
+
+        If difficult_words_pct > 5%:
+            Adjusted Score = Raw Score + 3.6365
+
+    Grade Level Correspondence:
+        4.9 or lower: Grade 4 and below
+        5.0-5.9: Grades 5-6
+        6.0-6.9: Grades 7-8
+        7.0-7.9: Grades 9-10
+        8.0-8.9: Grades 11-12
+        9.0-9.9: Grades 13-15 (College)
+        10.0+: Grade 16+ (College Graduate)
+
+    Advantages:
+        - Based on empirical word familiarity data
+        - Works well for educational materials
+        - Well-validated across grade levels
+        - Considers both vocabulary and syntax
+
+    Disadvantages:
+        - Requires maintaining 3000-word familiar list
+        - List is dated (1948, updated 1995)
+        - May not reflect modern vocabulary
+        - Doesn't account for concept difficulty
+
+    Args:
+        text: Input text to analyze. Should contain at least one complete
+              sentence. Empty text returns NaN values.
+
+    Returns:
+        DaleChallResult containing:
+            - dale_chall_score: The Dale-Chall readability score
+            - grade_level: Grade range (e.g., "7-8", "College")
+            - difficult_word_count: Words not on familiar list
+            - difficult_word_ratio: Difficult words / total words
+            - avg_sentence_length: Average words per sentence
+            - total_words: Total word count
+            - metadata: List of difficult words, adjusted score flag, etc.
+
+    Example:
+        >>> result = compute_dale_chall("Sample educational text...")
+        >>> print(f"Dale-Chall score: {result.dale_chall_score:.2f}")
+        Dale-Chall score: 7.3
+        >>> print(f"Grade level: {result.grade_level}")
+        Grade level: 7-8
+        >>> print(f"Difficult words: {result.difficult_word_ratio * 100:.1f}%")
+        Difficult words: 12.4%
+
+    Note:
+        - Case-insensitive word matching
+        - Punctuation stripped before word lookup
+        - Proper nouns may be flagged as difficult even if well-known
+        - Technical/specialized texts score higher than general texts
+    """
+    # Tokenize and segment
+    sentences = split_sentences(text)
+    tokens = tokenize(text)
+    word_tokens = normalize_for_readability(tokens)
+
+    if len(sentences) == 0 or len(word_tokens) == 0:
+        return DaleChallResult(
+            dale_chall_score=float("nan"),
+            grade_level="Unknown",
+            difficult_word_count=0,
+            difficult_word_ratio=float("nan"),
+            avg_sentence_length=float("nan"),
+            total_words=0,
+            metadata={
+                "sentence_count": 0,
+                "raw_score": float("nan"),
+                "adjusted": False,
+                "difficult_words_sample": [],
+            },
+        )
+
+    # Count difficult words (not in familiar list)
+    difficult_words = []
+    for word in word_tokens:
+        word_lower = word.lower()
+        if word_lower not in DALE_CHALL_FAMILIAR_WORDS:
+            difficult_words.append(word)
+
+    difficult_word_count = len(difficult_words)
+    difficult_word_ratio = difficult_word_count / len(word_tokens)
+    difficult_word_pct = difficult_word_ratio * 100
+
+    # Calculate average sentence length
+    avg_sentence_length = len(word_tokens) / len(sentences)
+
+    # Calculate raw score
+    raw_score = 0.1579 * difficult_word_pct + 0.0496 * avg_sentence_length
+
+    # Apply adjustment if difficult word % > 5.0
+    adjusted = difficult_word_pct > 5.0
+    if adjusted:
+        dale_chall_score = raw_score + 3.6365
+    else:
+        dale_chall_score = raw_score
+
+    # Map score to grade level
+    if dale_chall_score < 5.0:
+        grade_level = "4 and below"
+    elif dale_chall_score < 6.0:
+        grade_level = "5-6"
+    elif dale_chall_score < 7.0:
+        grade_level = "7-8"
+    elif dale_chall_score < 8.0:
+        grade_level = "9-10"
+    elif dale_chall_score < 9.0:
+        grade_level = "11-12"
+    elif dale_chall_score < 10.0:
+        grade_level = "College"
+    else:
+        grade_level = "College Graduate"
+
+    # Build metadata
+    # Sample up to 20 difficult words for metadata (avoid huge lists)
+    difficult_words_sample = list(set(difficult_words))[:20]
+
+    metadata = {
+        "sentence_count": len(sentences),
+        "raw_score": raw_score,
+        "adjusted": adjusted,
+        "difficult_word_pct": difficult_word_pct,
+        "difficult_words_sample": difficult_words_sample,
+        "familiar_word_list_size": len(DALE_CHALL_FAMILIAR_WORDS),
+    }
+
+    return DaleChallResult(
+        dale_chall_score=dale_chall_score,
+        grade_level=grade_level,
+        difficult_word_count=difficult_word_count,
+        difficult_word_ratio=difficult_word_ratio,
+        avg_sentence_length=avg_sentence_length,
+        total_words=len(word_tokens),
+        metadata=metadata,
+    )
+
+
+def compute_linsear_write(text: str) -> LinsearWriteResult:
+    """
+    Compute Linsear Write Readability Formula.
+
+    Developed for the U.S. Air Force to assess technical writing, the Linsear
+    Write formula classifies words as "easy" (1-2 syllables) or "hard" (3+
+    syllables) and uses sentence length to estimate grade level.
+
+    Related GitHub Issue:
+        #16 - Additional Readability Formulas
+        https://github.com/craigtrim/pystylometry/issues/16
+
+    Formula:
+        1. Count "easy" words (1-2 syllables): multiply count by 1
+        2. Count "hard" words (3+ syllables): multiply count by 3
+        3. Divide sum by number of sentences
+        4. If result > 20, divide by 2 to get grade level
+        5. If result <= 20, subtract 2, then divide by 2
+
+    The formula is optimized for technical writing and works best with
+    passages of about 100 words.
+
+    Advantages:
+        - Simple binary classification (easy/hard)
+        - Effective for technical documents
+        - Fast computation
+        - Developed specifically for instructional materials
+
+    Disadvantages:
+        - Less well-known than other formulas
+        - Binary word classification is crude
+        - May overestimate difficulty of technical terms
+        - Limited validation compared to Flesch or Dale-Chall
+
+    Args:
+        text: Input text to analyze. Works best with 100-word samples.
+              Empty text returns NaN values.
+
+    Returns:
+        LinsearWriteResult containing:
+            - linsear_score: The Linsear Write score
+            - grade_level: Corresponding U.S. grade level (integer)
+            - easy_word_count: Words with 1-2 syllables
+            - hard_word_count: Words with 3+ syllables
+            - avg_sentence_length: Average words per sentence
+            - metadata: Calculation details, sentence count, etc.
+
+    Example:
+        >>> result = compute_linsear_write("Technical manual text...")
+        >>> print(f"Linsear Write score: {result.linsear_score:.2f}")
+        Linsear Write score: 11.3
+        >>> print(f"Grade level: {result.grade_level}")
+        Grade level: 11
+        >>> print(f"Easy words: {result.easy_word_count}")
+        Easy words: 78
+        >>> print(f"Hard words: {result.hard_word_count}")
+        Hard words: 22
+
+    Note:
+        - Syllable counting required (use existing syllable module)
+        - Punctuation and numbers typically excluded
+        - Most accurate with 100-word samples
+        - Grade level is rounded to nearest integer
+    """
+    # Tokenize and segment
+    sentences = split_sentences(text)
+    tokens = tokenize(text)
+    word_tokens = normalize_for_readability(tokens)
+
+    if len(sentences) == 0 or len(word_tokens) == 0:
+        return LinsearWriteResult(
+            linsear_score=float("nan"),
+            grade_level=0,
+            easy_word_count=0,
+            hard_word_count=0,
+            avg_sentence_length=float("nan"),
+            metadata={"sentence_count": 0, "total_words": 0, "raw_score": float("nan")},
+        )
+
+    # Classify words as easy (1-2 syllables) or hard (3+ syllables)
+    easy_word_count = 0
+    hard_word_count = 0
+
+    for word in word_tokens:
+        syllable_count = count_syllables(word)
+        if syllable_count <= 2:
+            easy_word_count += 1
+        else:
+            hard_word_count += 1
+
+    # Calculate weighted sum
+    weighted_sum = (easy_word_count * 1) + (hard_word_count * 3)
+
+    # Calculate score
+    raw_score = weighted_sum / len(sentences)
+
+    # Convert to grade level
+    if raw_score > 20:
+        grade_level = round(raw_score / 2)
+    else:
+        grade_level = round((raw_score - 2) / 2)
+
+    # Ensure grade level is non-negative
+    grade_level = max(0, grade_level)
+
+    # Calculate average sentence length
+    avg_sentence_length = len(word_tokens) / len(sentences)
+
+    # Build metadata
+    metadata = {
+        "total_words": len(word_tokens),
+        "sentence_count": len(sentences),
+        "raw_score": raw_score,
+        "weighted_sum": weighted_sum,
+    }
+
+    return LinsearWriteResult(
+        linsear_score=raw_score,
+        grade_level=grade_level,
+        easy_word_count=easy_word_count,
+        hard_word_count=hard_word_count,
+        avg_sentence_length=avg_sentence_length,
+        metadata=metadata,
+    )
+
+
+def compute_fry(text: str) -> FryResult:
+    """
+    Compute Fry Readability Graph metrics.
+
+    The Fry Readability Graph plots average sentence length against average
+    syllables per 100 words to determine reading difficulty. This implementation
+    provides the numerical coordinates and estimated grade level.
+
+    Related GitHub Issue:
+        #16 - Additional Readability Formulas
+        https://github.com/craigtrim/pystylometry/issues/16
+
+    Method:
+        1. Select three 100-word samples from text
+        2. Count average sentence length across samples
+        3. Count average syllables per 100 words across samples
+        4. Plot coordinates on Fry graph (or use numerical approximation)
+        5. Determine grade level from graph zone
+
+    The original Fry graph has zones corresponding to grade levels 1-17+.
+    This implementation uses numerical approximation to estimate grade level.
+
+    Advantages:
+        - Visual/graphical approach (intuitive)
+        - Uses two independent dimensions (length & syllables)
+        - Well-validated for educational materials
+        - Covers wide range of grade levels (1-17+)
+
+    Disadvantages:
+        - Requires exactly 100-word samples (padding/truncation needed)
+        - Graph reading can be subjective
+        - Less precise than formula-based methods
+        - Multiple samples needed for reliability
+
+    Args:
+        text: Input text to analyze. Should contain at least 100 words.
+              Shorter texts are padded or return limited results.
+
+    Returns:
+        FryResult containing:
+            - avg_sentence_length: Average words per sentence
+            - avg_syllables_per_100: Average syllables per 100 words
+            - grade_level: Estimated grade level (e.g., "5", "7", "College")
+            - graph_zone: Which zone of Fry graph (for validity checking)
+            - metadata: Sample details, total sentences, syllables, etc.
+
+    Example:
+        >>> result = compute_fry("Educational text for grade assessment...")
+        >>> print(f"Avg sentence length: {result.avg_sentence_length:.1f}")
+        Avg sentence length: 14.3
+        >>> print(f"Syllables/100 words: {result.avg_syllables_per_100:.1f}")
+        Syllables/100 words: 142.7
+        >>> print(f"Grade level: {result.grade_level}")
+        Grade level: 6
+
+    Note:
+        - Original method uses three 100-word samples
+        - Implementation may use single sample or whole text
+        - Syllable counting required
+        - Grade level estimation uses zone boundaries
+        - Some texts fall outside graph zones (marked as invalid)
+    """
+    # Tokenize and segment
+    sentences = split_sentences(text)
+    tokens = tokenize(text)
+    word_tokens = normalize_for_readability(tokens)
+
+    if len(sentences) == 0 or len(word_tokens) == 0:
+        return FryResult(
+            avg_sentence_length=float("nan"),
+            avg_syllables_per_100=float("nan"),
+            grade_level="Unknown",
+            graph_zone="invalid",
+            metadata={
+                "total_sentences": 0,
+                "total_syllables": 0,
+                "total_words": 0,
+                "sample_size": 0,
+            },
+        )
+
+    # Use first 100 words for sample (or entire text if < 100 words)
+    sample_size = min(100, len(word_tokens))
+    sample_tokens = word_tokens[:sample_size]
+
+    # Count syllables in sample
+    total_syllables = sum(count_syllables(word) for word in sample_tokens)
+
+    # Count sentences within the sample
+    # We need to determine how many sentences are in the first sample_size words
+    word_count_so_far = 0
+    sentences_in_sample = 0
+    for sent in sentences:
+        sent_tokens = tokenize(sent)
+        sent_word_tokens = normalize_for_readability(sent_tokens)
+        if word_count_so_far + len(sent_word_tokens) <= sample_size:
+            sentences_in_sample += 1
+            word_count_so_far += len(sent_word_tokens)
+        else:
+            # Partial sentence in sample
+            if word_count_so_far < sample_size:
+                sentences_in_sample += 1
+            break
+
+    # Ensure at least 1 sentence for division
+    sentences_in_sample = max(1, sentences_in_sample)
+
+    # Calculate avg_sentence_length (for the sample)
+    avg_sentence_length = sample_size / sentences_in_sample
+
+    # Calculate avg_syllables_per_100 (scale if sample < 100)
+    avg_syllables_per_100 = (total_syllables / sample_size) * 100
+
+    # Map to grade level using Fry graph approximation
+    # Fry graph zones (simplified numerical approximation):
+    # These are rough boundaries based on Fry graph zones
+    # X-axis: avg sentences per 100 words (inverse of avg_sentence_length)
+    # Y-axis: avg syllables per 100 words
+
+    # Determine grade level based on avg_sentence_length and avg_syllables_per_100
+    # Higher syllables per 100 = higher grade
+    # Longer sentences = higher grade
+    # Simplified zone mapping:
+    if avg_syllables_per_100 < 125:
+        if avg_sentence_length < 7:
+            grade_level = "1"
+            graph_zone = "valid"
+        elif avg_sentence_length < 11:
+            grade_level = "2"
+            graph_zone = "valid"
+        else:
+            grade_level = "3"
+            graph_zone = "valid"
+    elif avg_syllables_per_100 < 135:
+        if avg_sentence_length < 8:
+            grade_level = "2"
+            graph_zone = "valid"
+        elif avg_sentence_length < 12:
+            grade_level = "3"
+            graph_zone = "valid"
+        else:
+            grade_level = "4"
+            graph_zone = "valid"
+    elif avg_syllables_per_100 < 145:
+        if avg_sentence_length < 9:
+            grade_level = "3"
+            graph_zone = "valid"
+        elif avg_sentence_length < 13:
+            grade_level = "5"
+            graph_zone = "valid"
+        else:
+            grade_level = "6"
+            graph_zone = "valid"
+    elif avg_syllables_per_100 < 155:
+        if avg_sentence_length < 10:
+            grade_level = "4"
+            graph_zone = "valid"
+        elif avg_sentence_length < 14:
+            grade_level = "7"
+            graph_zone = "valid"
+        else:
+            grade_level = "8"
+            graph_zone = "valid"
+    elif avg_syllables_per_100 < 165:
+        if avg_sentence_length < 12:
+            grade_level = "6"
+            graph_zone = "valid"
+        elif avg_sentence_length < 16:
+            grade_level = "9"
+            graph_zone = "valid"
+        else:
+            grade_level = "10"
+            graph_zone = "valid"
+    elif avg_syllables_per_100 < 175:
+        if avg_sentence_length < 14:
+            grade_level = "8"
+            graph_zone = "valid"
+        elif avg_sentence_length < 18:
+            grade_level = "11"
+            graph_zone = "valid"
+        else:
+            grade_level = "12"
+            graph_zone = "valid"
+    else:  # avg_syllables_per_100 >= 175
+        if avg_sentence_length < 16:
+            grade_level = "10"
+            graph_zone = "valid"
+        elif avg_sentence_length < 20:
+            grade_level = "College"
+            graph_zone = "valid"
+        else:
+            grade_level = "College+"
+            graph_zone = "valid"
+
+    # Check if outside typical graph bounds
+    if avg_syllables_per_100 > 185 or avg_sentence_length > 25:
+        graph_zone = "above_graph"
+    elif avg_syllables_per_100 < 110:
+        graph_zone = "below_graph"
+
+    # Build metadata
+    metadata = {
+        "total_sentences": len(sentences),
+        "total_syllables": sum(count_syllables(w) for w in word_tokens),
+        "total_words": len(word_tokens),
+        "sample_size": sample_size,
+        "sentences_in_sample": sentences_in_sample,
+        "syllables_in_sample": total_syllables,
+    }
+
+    return FryResult(
+        avg_sentence_length=avg_sentence_length,
+        avg_syllables_per_100=avg_syllables_per_100,
+        grade_level=grade_level,
+        graph_zone=graph_zone,
+        metadata=metadata,
+    )
+
+
+def compute_forcast(text: str) -> FORCASTResult:
+    """
+    Compute FORCAST Readability Formula.
+
+    FORCAST (FORmula for CASTing readability) was developed by the U.S. military
+    to assess readability without counting syllables. It uses only the count of
+    single-syllable words as its metric, making it fast and simple.
+
+    Related GitHub Issue:
+        #16 - Additional Readability Formulas
+        https://github.com/craigtrim/pystylometry/issues/16
+
+    Formula:
+        Grade Level = 20 - (N / 10)
+
+        Where N is the number of single-syllable words in a 150-word sample.
+
+    The formula is optimized for technical and military documents and works
+    best with standardized 150-word samples.
+
+    Advantages:
+        - Extremely simple (only counts single-syllable words)
+        - No sentence segmentation required
+        - Fast computation
+        - Developed specifically for military/technical texts
+
+    Disadvantages:
+        - Less well-known and validated than other formulas
+        - Requires exactly 150-word samples
+        - Single dimension (doesn't consider sentence length)
+        - May not generalize well beyond military context
+
+    Args:
+        text: Input text to analyze. Works best with 150-word samples.
+              Shorter texts are padded or scored proportionally.
+              Longer texts use first 150 words or multiple samples.
+
+    Returns:
+        FORCASTResult containing:
+            - forcast_score: The FORCAST readability score
+            - grade_level: Corresponding U.S. grade level (integer)
+            - single_syllable_ratio: Single-syllable words / total words
+            - single_syllable_count: Count of single-syllable words
+            - total_words: Total word count analyzed
+            - metadata: Sample details, calculation specifics, etc.
+
+    Example:
+        >>> result = compute_forcast("Military technical document...")
+        >>> print(f"FORCAST score: {result.forcast_score:.2f}")
+        FORCAST score: 9.7
+        >>> print(f"Grade level: {result.grade_level}")
+        Grade level: 10
+        >>> print(f"Single-syllable ratio: {result.single_syllable_ratio:.3f}")
+        Single-syllable ratio: 0.687
+
+    Note:
+        - Syllable counting required (but only to identify 1-syllable words)
+        - Recommended sample size is 150 words
+        - Multiple samples can be averaged for longer texts
+        - Simpler than most readability formulas
+        - Grade levels typically range from 5-12
+    """
+    # Tokenize
+    tokens = tokenize(text)
+    word_tokens = normalize_for_readability(tokens)
+
+    if len(word_tokens) == 0:
+        return FORCASTResult(
+            forcast_score=float("nan"),
+            grade_level=0,
+            single_syllable_ratio=float("nan"),
+            single_syllable_count=0,
+            total_words=0,
+            metadata={"sample_size": 0, "scaled_n": float("nan")},
+        )
+
+    # Use first 150 words for sample (or entire text if < 150 words)
+    sample_size = min(150, len(word_tokens))
+    sample_tokens = word_tokens[:sample_size]
+
+    # Count single-syllable words in sample
+    single_syllable_count = 0
+    for word in sample_tokens:
+        if count_syllables(word) == 1:
+            single_syllable_count += 1
+
+    # Scale N to 150-word basis if sample < 150
+    if sample_size < 150:
+        scaled_n = single_syllable_count * (150 / sample_size)
+    else:
+        scaled_n = single_syllable_count
+
+    # Calculate grade level: 20 - (N / 10)
+    forcast_score = 20 - (scaled_n / 10)
+    grade_level = round(forcast_score)
+
+    # Ensure grade level is in reasonable range (0-20)
+    grade_level = max(0, min(20, grade_level))
+
+    # Calculate single syllable ratio (for the sample)
+    single_syllable_ratio = single_syllable_count / sample_size
+
+    # Build metadata
+    metadata = {
+        "sample_size": sample_size,
+        "scaled_n": scaled_n,
+        "total_words_in_text": len(word_tokens),
+    }
+
+    return FORCASTResult(
+        forcast_score=forcast_score,
+        grade_level=grade_level,
+        single_syllable_ratio=single_syllable_ratio,
+        single_syllable_count=single_syllable_count,
+        total_words=sample_size,
+        metadata=metadata,
+    )
+
+
+def compute_powers_sumner_kearl(text: str) -> PowersSumnerKearlResult:
+    """
+    Compute Powers-Sumner-Kearl Readability Formula.
+
+    The Powers-Sumner-Kearl (PSK) formula is a recalibration of the Flesch
+    Reading Ease formula, optimized for primary grade levels (grades 1-4).
+    It uses the same inputs (sentence length, syllables per word) but with
+    different coefficients.
+
+    Related GitHub Issue:
+        #16 - Additional Readability Formulas
+        https://github.com/craigtrim/pystylometry/issues/16
+
+    Formula:
+        Grade Level = 0.0778 * avg_sentence_length + 0.0455 * avg_syllables_per_word - 2.2029
+
+    The formula was derived from analysis of primary-grade texts and provides
+    more accurate grade-level estimates for beginning readers than the original
+    Flesch formula.
+
+    Advantages:
+        - Optimized for primary grades (1-4)
+        - More accurate than Flesch for young readers
+        - Uses same inputs as Flesch (easy to compare)
+        - Well-validated on educational materials
+
+    Disadvantages:
+        - Less accurate for higher grade levels
+        - Less well-known than Flesch
+        - Limited range (not suitable for college-level texts)
+        - Requires syllable counting
+
+    Args:
+        text: Input text to analyze. Optimized for children's literature
+              and primary-grade educational materials. Empty text returns
+              NaN values.
+
+    Returns:
+        PowersSumnerKearlResult containing:
+            - psk_score: The Powers-Sumner-Kearl score
+            - grade_level: Corresponding grade (decimal, e.g., 2.5 = mid-2nd grade)
+            - avg_sentence_length: Average words per sentence
+            - avg_syllables_per_word: Average syllables per word
+            - total_sentences: Total sentence count
+            - total_words: Total word count
+            - total_syllables: Total syllable count
+            - metadata: Comparison to Flesch, calculation details, etc.
+
+    Example:
+        >>> result = compute_powers_sumner_kearl("Children's book text...")
+        >>> print(f"PSK score: {result.psk_score:.2f}")
+        PSK score: 2.3
+        >>> print(f"Grade level: {result.grade_level:.1f}")
+        Grade level: 2.3
+        >>> print(f"Avg sentence length: {result.avg_sentence_length:.1f}")
+        Avg sentence length: 8.5
+
+    Note:
+        - Most accurate for grades 1-4
+        - Can produce negative scores for very simple texts
+        - Grade level is continuous (can be decimal)
+        - Syllable counting required (same as Flesch)
+        - Compare to Flesch results for validation
+    """
+    # Tokenize and segment
+    sentences = split_sentences(text)
+    tokens = tokenize(text)
+    word_tokens = normalize_for_readability(tokens)
+
+    if len(sentences) == 0 or len(word_tokens) == 0:
+        return PowersSumnerKearlResult(
+            psk_score=float("nan"),
+            grade_level=float("nan"),
+            avg_sentence_length=float("nan"),
+            avg_syllables_per_word=float("nan"),
+            total_sentences=0,
+            total_words=0,
+            total_syllables=0,
+            metadata={
+                "flesch_reading_ease": float("nan"),
+                "flesch_kincaid_grade": float("nan"),
+            },
+        )
+
+    # Count syllables
+    total_syllables = sum(count_syllables(word) for word in word_tokens)
+
+    # Calculate metrics
+    avg_sentence_length = len(word_tokens) / len(sentences)
+    avg_syllables_per_word = total_syllables / len(word_tokens)
+
+    # Apply Powers-Sumner-Kearl formula
+    # Grade = 0.0778 * avg_sentence_length + 0.0455 * avg_syllables_per_word - 2.2029
+    psk_score = (
+        0.0778 * avg_sentence_length + 0.0455 * avg_syllables_per_word - 2.2029
+    )
+    grade_level = round(psk_score, 1)  # Round to 1 decimal place
+
+    # Optional: Calculate Flesch scores for comparison
+    flesch_reading_ease = (
+        206.835 - 1.015 * avg_sentence_length - 84.6 * avg_syllables_per_word
+    )
+    flesch_kincaid_grade = (
+        0.39 * avg_sentence_length + 11.8 * avg_syllables_per_word - 15.59
+    )
+
+    # Build metadata
+    metadata = {
+        "flesch_reading_ease": flesch_reading_ease,
+        "flesch_kincaid_grade": flesch_kincaid_grade,
+        "difference_from_flesch": psk_score - flesch_kincaid_grade,
+        "words_per_sentence": avg_sentence_length,
+        "syllables_per_word": avg_syllables_per_word,
+    }
+
+    return PowersSumnerKearlResult(
+        psk_score=psk_score,
+        grade_level=grade_level,
+        avg_sentence_length=avg_sentence_length,
+        avg_syllables_per_word=avg_syllables_per_word,
+        total_sentences=len(sentences),
+        total_words=len(word_tokens),
+        total_syllables=total_syllables,
+        metadata=metadata,
+    )