algospeak / algospeak_dictionary.py
Veronyka
Add AlgoSpeak AI app with dictionary and translation features
8febfa7
"""
AlgoSpeak Dictionary - Coded terms used to circumvent algorithmic moderation
Sources: Public research, content creator documentation, social media analysis
To add new terms: simply add them to the dictionary below
"""
ALGOSPEAK_DICT = {
# Violence and death
"unalive": "die, kill, suicide",
"unalived": "dead, killed",
"sewerslide": "suicide",
"self-delete": "suicide",
"game over": "death",
# Weapons and conflict
"pew pew": "gun, gunshots",
"noodle": "weapon",
"spicy eggplant": "firearm",
"cornucopia": "war, armed conflict",
"kaboom": "explosion, bomb",
# Adult content and sexuality
"seggs": "sex",
"secs": "sex",
"spicy time": "sex",
"corn": "pornography",
"corn site": "pornographic website",
"le$bian": "lesbian",
"le dollar bean": "lesbian",
"g@y": "gay",
"accountant": "sex worker",
"corn star": "porn actor/actress",
"spicy content": "adult/sexual content",
# Sexual violence
"SA": "sexual assault",
"grape": "rape",
"graping": "rape",
"r-word": "rape",
# Health and body
"mascara": "COVID-19, coronavirus",
"backshot": "vaccine",
"panini": "pandemic",
"spicy cough": "COVID-19",
"leg booty": "LGBTQ+",
"leg booty community": "LGBTQ+ community",
# Crime
"PDF file": "pedophile",
"PDF": "pedophile",
"minor attracted person": "pedophile (dangerous euphemism)",
"unalive juice": "lethal drugs",
# Body and anatomy
"chest feeding": "breastfeeding",
"no-no square": "genitalia",
"front butt": "female genitalia",
"boy parts": "male genitalia",
"girl parts": "female genitalia",
# Social issues
"ouid": "marijuana, weed",
"lettuce": "marijuana",
"devil's lettuce": "marijuana",
"skripper": "stripper",
"sewer slide": "suicide",
# Other common terms
"camping": "human trafficking",
"camping trip": "kidnapping, trafficking",
"disappeared": "dead, killed",
"word": "replaces censored words (context-dependent)",
}
# Categories for better organization
CATEGORIES = {
"violence": ["unalive", "unalived", "sewerslide", "self-delete", "game over", "disappeared"],
"weapons": ["pew pew", "noodle", "spicy eggplant", "kaboom"],
"war": ["cornucopia"],
"sexuality": ["seggs", "secs", "spicy time", "le$bian", "g@y", "leg booty"],
"pornography": ["corn", "corn site", "corn star"],
"sex_work": ["accountant", "skripper"],
"sexual_violence": ["SA", "grape", "graping", "r-word"],
"health": ["mascara", "backshot", "panini", "spicy cough"],
"drugs": ["ouid", "lettuce", "devil's lettuce"],
"crime": ["PDF file", "camping", "camping trip"],
"anatomy": ["chest feeding", "no-no square", "front butt", "boy parts", "girl parts"],
}
def get_algospeak_context() -> str:
"""Returns formatted context for use in LLM prompts"""
lines = ["AlgoSpeak Dictionary (coded terms to avoid algorithmic censorship):\n"]
for term, meaning in sorted(ALGOSPEAK_DICT.items()):
lines.append(f"- \"{term}\" = {meaning}")
return "\n".join(lines)
def get_category_terms(category: str) -> list:
"""Returns terms from a specific category"""
return CATEGORIES.get(category, [])
if __name__ == "__main__":
print(f"AlgoSpeak Dictionary loaded with {len(ALGOSPEAK_DICT)} terms")
print(f"Available categories: {', '.join(CATEGORIES.keys())}")
print("\n" + get_algospeak_context()[:500] + "...")