Oyemi in Action

pip install oyemi

from Oyemi import Encoder

# Initialize encoder
enc = Encoder()

# Encode a simple word
codes = enc.encode("happy")
print("Codes for 'happy':", codes)

# Polysemous word (multiple meanings)
codes = enc.encode("bank")
print("Codes for 'bank':", codes[:3])  # First 3 senses

# Check lexicon size
print(f"Lexicon contains {enc.word_count:,} words")

# Get parsed semantic codes
parsed = enc.encode_parsed("fear")

# Examine the primary sense
primary = parsed[0]
print(f"Word: fear")
print(f"  Code: {primary.raw}")
print(f"  Superclass: {primary.superclass}")
print(f"  Part of Speech: {primary.pos_name}")
print(f"  Abstractness: {primary.abstractness_name}")
print(f"  Valence: {primary.valence_name}")

# Compare positive vs negative words
for word in ["love", "hate", "table"]:
    p = enc.encode_parsed(word)[0]
    print(f"{word:10} -> {p.valence_name}")

# Analyze sentiment of a sentence
sentence = "The manager was incompetent and the layoffs were devastating"

# Tokenize and analyze
words = sentence.lower().split()
valence_counts = {'positive': 0, 'negative': 0, 'neutral': 0}

for word in words:
    try:
        parsed = enc.encode_parsed(word, raise_on_unknown=False)
        if parsed:
            valence = parsed[0].valence_name
            valence_counts[valence] += 1
            if valence != 'neutral':
                print(f"  {word}: {valence}")
    except:
        pass

print(f"\nValence Summary: {valence_counts}")

# Calculate sentiment score
total = sum(valence_counts.values())
score = (valence_counts['positive'] - valence_counts['negative']) / total
print(f"Sentiment Score: {score:.2f}")

from Oyemi import find_synonyms

# Find synonyms for emotional words
for word in ["happy", "angry", "fired"]:
    syns = find_synonyms(word, limit=5)
    print(f"{word}: {syns}")

# Get weighted synonyms (higher weight = closer match)
weighted = find_synonyms("fear", return_weighted=True, limit=5)
print("\nWeighted synonyms for 'fear':")
for syn, weight in weighted:
    print(f"  {syn}: {weight:.2f}")

from Oyemi import semantic_similarity

# Compare word pairs
pairs = [
    ("happy", "joyful"),     # Synonyms
    ("happy", "sad"),        # Antonyms
    ("dog", "cat"),          # Same category
    ("dog", "computer"),    # Different categories
    ("layoff", "fired"),    # Related workplace terms
]

print("Semantic Similarity Scores:")
for w1, w2 in pairs:
    sim = semantic_similarity(w1, w2)
    print(f"  {w1:12} <-> {w2:12}: {sim:.2f}")

from Oyemi import cluster_by_superclass

# Words from employee feedback
words = [
    "manager", "boss", "supervisor",  # Leadership
    "salary", "bonus", "compensation",  # Money
    "layoff", "fired", "terminated",  # Employment
    "stress", "anxiety", "fear",  # Emotions
]

# Cluster by semantic category
clusters = cluster_by_superclass(words)

print("Semantic Clusters:")
for superclass, cluster_words in clusters.items():
    print(f"\n  [{superclass}]")
    for w in cluster_words:
        print(f"    - {w}")