🔪🥩 largonji

« Larlépem-vous louchébem? » — Hybrid French ↔ Louchébem converter

🇫🇷 Version française / French version →

A modern Python converter to transform French into louchébem, the historical slang of Parisian butchers.

Louchébem is the main variant of largonji (English Wikipedia), a family of linguistic deformation techniques used in French slang (including also javanais and other variants).

This package implements louchébem with a hybrid approach: dictionary of authentic historical terms + algorithmic transformation for unknown words.

🚀 Installation

# With uv (recommended)
uv add largonji

# With pip
pip install largonji

💻 Quick Start

from largonji import LouchebemConverter

# Create a converter and transform text
converter = LouchebemConverter()
print(converter.convert_text("Bonjour le boucher"))
# → Lonjourbem le loucherbem

🥩 What is Louchébem?

Louchébem is the main variant of largonji, a family of linguistic deformation techniques developed in French slang. Louchébem was created and popularized by Parisian butchers in the 19th century.

The word "louchébem" itself comes from transforming "boucher" (butcher) according to the process's rules.

📜 A Bit of History

Largonji appeared in the slang of Parisian working classes from the late 18th or early 19th century. The earliest recorded words include:

lomben (← bon, "good") in an 1821 slang glossary
La Lorcefé (← La Force, a Parisian prison) in Vidocq's Memoirs (1828-1829)
loucherbem (← boucher, "butcher") attested around 1876

Some words from louchébem have entered common French language:

loufoque (← fou, "crazy") : bizarre, wacky
larfeuille (← feuille, "leaf/paper", slang for money) : cash, dough
à loilpé (← à poil, "naked") : in the buff

🎯 The Basic Principle

The louchébem process follows a simple rule:

Replace the first consonant (or consonant cluster) with "l"
Move that consonant to the end of the word
Add a suffix (often related to the moved consonant)

Examples:

boucher → loucherbem
jargon → largonji
prix → liprem ("pr" cluster moved together)

🔧 The Naive Approach

A basic louchébem implementation might look like this:

def louchebem_naive(word):
    """Simplified version (only works for basic cases)"""
    if not word:
        return word
    
    # Extract first consonant
    first_consonant = word[0]
    rest = word[1:]
    
    # Build transformed word
    return 'l' + rest + first_consonant + 'em'

# Examples
print(louchebem_naive("boucher"))  # → loucherbem ✓
print(louchebem_naive("prix"))     # → lixprem ✗ (should be liprem)
print(louchebem_naive("entendre")) # → lntendree ✗ (should be enlendreté)

⚠️ Limitations of the Naive Approach

This method doesn't handle:

Consonant clusters (pr, tr, fr, etc.)
Vowel-initial words
Silent consonants at word end (discret → discrè)
French phonetics (qu, gu, s/z, etc.)
Varied suffixes based on consonant type
Historical words with established spelling

Our hybrid implementation solves all these problems! 🎉

✨ Our Implementation Choices

Louchébem was never a standardized language: not all butchers spoke exactly the same slang. Suffixes in particular varied from person to person, and the language evolved deliberately to remain less obvious to outsiders (non-deterministic aspect). Some cases are not really clear in historical sources, notably the treatment of vowel-initial words.

This implementation is therefore based on deliberate choices: it relies on available sources and what sounds best. This section details these technical decisions.

1. 📚 Hybrid Approach: Dictionary + Algorithm

Problem: Some louchébem words have established historical spelling that may differ from algorithmic transformation.

Solution:

Dictionary of ~100 authentic historical words (sources: Lorédan Larchey 1858, Gaston Esnault 1965, louchebem.fr)
Algorithmic transformation for unlisted words
Dictionary takes priority when word exists

# Example: "argot" is in the dictionary
converter.convert_text("argot")  # → largomuche (historical form)

# "robot" doesn't exist in the dictionary
converter.convert_text("robot")  # → lobotrem (algorithmic transformation)

2. 🛡️ Selective Word Preservation

Problem: Transforming all words makes text unreadable. Which words to preserve?

Solution: System of individually toggleable preservation rules:

Category	Examples	Reason
Function words	le, la, de, un, et, à	Grammatical structure
Ultra-common verbs	être, avoir, faire, aller	Readability
Interjections	oh, ah, hein, ben	Oral expression
Numbers & dates	123, XIV, 31/12/2023	Precise information
Proper nouns	Paris, Marie	Identification
Acronyms	SNCF, UNESCO	Abbreviations
Already louchébem	loucherbem, louf	Avoid double transformation

# Disable certain preservations
config = LouchebemConfig(
    preserve_stopwords=True,        # Keep "le", "la", etc.
    preserve_proper_nouns=False,    # Transform "Paris" too!
    preserve_numbers=True,          # Keep "123"
)
converter = LouchebemConverter(config=config)

3. 🔤 Vowel-Initial Words

Problem: How to transform "entendre", "attention", "orange" that start with a vowel?

Solution: Find the attack consonant (first consonant cluster after the initial vowel sound). This is an implementation choice — historical sources are not clear on this case. This approach sounds better and allows keeping words not all starting with "l" in a text, which is more pleasant to read.

"entendre"  → "en" (nasal vowel) + "t" (attack) + "endre"
            → en + l + endre + t + és
            → "enlendretès"

"attention" → "a" + "t" (simplified from "tt") + "ention"
            → a + l + ention + t + és
            → "alentiontès"

"orange"    → "o" + "r" + "ange"
            → o + l + ange + r + em
            → "olangerem"

Tip: The code handles nasal vowels ("an", "en", "in", "on", "un") as vowel sounds.

4. 🎯 Weighted Suffixes by Consonant Type

Problem: Not all suffixes are equally probable. Historically, certain suffixes match better with certain consonants.

Solution: Weighted random suffix selection organized by consonant, based on historical usage.

# Example suffixes for different consonants
D → dé (35%), dem (25%), doc (5%), dique (5%), ...
P → pem (30%), puche (25%), poc (15%), pique (5%), ...
F → fès (35%), foc (20%), fem (15%), fique (10%), ...

Each suffix already contains its consonant to ensure phonetic harmony.

5. 🎵 Multi-Consonant Clusters

Problem: How to handle "prix", "train", "fromage" that start with 2+ consonants?

Solution: Extract the complete cluster of consonants before the first vowel and move it as a block.

"prix"    → "p" + "r" + "ix" → l + ix + pr + em → "liprem"
"train"   → "t" + "r" + "ain" → l + ain + tr + em → "laintrem"
"fromage" → "f" + "r" + "omage" → l + omaj + fr + é → "lomajfré"

Technical detail: Clusters are always preserved in full, even if the suffix already contains one of the letters (e.g., "pl" stays "pl", not just "p").

6. 🔇 Silent Consonants and Phonetic Adjustments

Problem: Written French ≠ spoken French. How to handle silent consonants?

Solution: Detection and removal of silent consonants with vowel adjustments.

"discret"  → discrè + t (silent removed, e→è to preserve sound)
           → l + iscrè + d + oc
           → "liscrèdoc"

"employée" → employé (ée→é, extra 'e' is silent)
           → e + l + oyé + pl + é
           → "emloyéplé"

"parler"   → parlé (er→é, identical sound)
           → l + arlé + p + em
           → "larlépem"

Applied rules:

-et → -è (discret → discrè)
-ent → -en (moment → momen, 't' is silent)
-er → -é (infinitive verbs)
-ée → -é (extra 'e' is redundant)

6. 🎲 Doubled Consonant Simplification

Problem: What to do with "attention" (two 't's)? What if we create duplicates (pl + lé = pllé)?

Solution:

Before moving: Simplify doubled consonants (tt→t, nn→n, mm→m)
After construction: Simplify any accidentally created duplicates

"attention" → "a" + "tt" → "a" + "t" (simplified) + "ention"
            → alentiontès

"employée"  → "em" + "pl" + "oyé" 
            → em + l + oyé + pl + é
            → emloyéplé

8. 📝 Apostrophe Handling (Elisions)

Problem: How to handle "l'argot", "d'autre", "j'aime"?

Solution: Specific rules based on prefix:

# Special case: l' + word → merge
"l'argot"   → "largot" → "largomuche" (then check dictionary)
"l'origine" → "lorigine" → "loriginelé"

# Other apostrophes: preserve prefix
"d'autre" → "d'" + "autre" transformed → "d'auletrem"
"j'aime"  → "j'" + "aime" transformed → "j'ailemem"

Logic: Merging with 'l' makes sense since the word will start with 'l' anyway!

⚙️ Advanced Configuration

The converter offers many configuration options:

from largonji import LouchebemConfig, LouchebemConverter

# Maximum configuration (default behavior)
config = LouchebemConfig.maximal()

# Minimal configuration (transforms almost everything)
config = LouchebemConfig.minimal()

# Configuration for reproducible tests
config = LouchebemConfig.for_testing(seed=42)

# Custom configuration
config = LouchebemConfig(
    # Word preservation
    preserve_stopwords=True,              # Keep "le", "la", "de", "un", etc. (function words)
    preserve_ultra_common_verbs=True,     # Keep "être", "avoir", "faire", "aller" (readability)
    preserve_interjections=True,          # Keep "oh", "ah", "hein", "ben" (oral expressions)
    preserve_numbers=True,                # Keep numbers and dates (123, XIV, 31/12/2023)
    preserve_proper_nouns=True,           # Keep detected proper nouns (Paris, Marie)
    preserve_acronyms=True,               # Keep acronyms (SNCF, UNESCO)
    preserve_already_louchebem=True,      # Don't re-transform words already in louchébem
    
    # Transformation features
    enable_apostrophe_merging=True,       # Merge "l'argot" into "largot" before transformation
    enable_l_initial_transform=True,      # Transform words starting with "l" (otherwise preserved)
    enable_silent_consonants=True,        # Remove silent consonants (discret → discrè)
    enable_doubled_consonant_simplification=True,  # Simplify "tt" → "t", "ll" → "l", etc.
    enable_infinitive_verbs=True,         # Transform "-er" to "-é" for infinitive verbs
    
    # General behavior
    preserve_case=True,                   # Preserve uppercase/lowercase from original text
    preserve_punctuation=True,            # Keep punctuation intact
    random_seed=None,                     # Random seed for suffixes (None = random, int = reproducible)
)

converter = LouchebemConverter(config=config)

📚 Sources & Acknowledgments

This project is based on historical and linguistic sources:

Wikipedia - Largonji (French): Detailed article on the history and processes of largonji
louchebem.fr: Reference site for louchébem. We drew inspiration from it, but our implementation differs in several aspects.
Sylvain Macouin: "À propos du Ladukteurtrès Largonjem" — Analysis of automatic transformation challenges
Jacques Haddad: Documentation on historical louchébem

📄 License

WTFPL - see the LICENSE file for details.

🤝 Contributing

Contributions are welcome! Feel free to:

🐛 Report bugs
💡 Suggest new features
📝 Improve documentation
➕ Add words to the historical dictionary

Made with ❤️ to preserve the largonji of the louchébems

« Dans le cochon, tout est bon ! »

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

🔪🥩 largonji

🚀 Installation

💻 Quick Start

📖 Table of Contents

🥩 What is Louchébem?

📜 A Bit of History

🎯 The Basic Principle

🔧 The Naive Approach

⚠️ Limitations of the Naive Approach

✨ Our Implementation Choices

1. 📚 Hybrid Approach: Dictionary + Algorithm

2. 🛡️ Selective Word Preservation

3. 🔤 Vowel-Initial Words

4. 🎯 Weighted Suffixes by Consonant Type

5. 🎵 Multi-Consonant Clusters

6. 🔇 Silent Consonants and Phonetic Adjustments

6. 🎲 Doubled Consonant Simplification

8. 📝 Apostrophe Handling (Elisions)

⚙️ Advanced Configuration

📚 Sources & Acknowledgments

📄 License

🤝 Contributing

FilesExpand file tree

README_EN.md

Latest commit

History

README_EN.md

File metadata and controls

🔪🥩 largonji

🚀 Installation

💻 Quick Start

📖 Table of Contents

🥩 What is Louchébem?

📜 A Bit of History

🎯 The Basic Principle

🔧 The Naive Approach

⚠️ Limitations of the Naive Approach

✨ Our Implementation Choices

1. 📚 Hybrid Approach: Dictionary + Algorithm

2. 🛡️ Selective Word Preservation

3. 🔤 Vowel-Initial Words

4. 🎯 Weighted Suffixes by Consonant Type

5. 🎵 Multi-Consonant Clusters

6. 🔇 Silent Consonants and Phonetic Adjustments

6. 🎲 Doubled Consonant Simplification

8. 📝 Apostrophe Handling (Elisions)

⚙️ Advanced Configuration

📚 Sources & Acknowledgments

📄 License

🤝 Contributing