mappings.py

convokit_ds = {
    "discourse": "reddit-coarse-discourse-corpus",
    "persuasion": "winning-args-corpus",
    "toxicity": "conversations-gone-awry-corpus",
    "politeness": "wikipedia-politeness-corpus",
    "stance": "supreme-corpus",
    "power": "wiki-corpus",
}

# (Where is the Label Stored, What is the Label Applied to, how much context)
convokit_labels = {
    "power": ("speaker", "meta.is-admin", "all"),
    "stance": ("speaker", "meta.side", "all"),
    "politeness": ("utterance", "meta.Binary", "all"),
    "toxicity": ("conversation", "meta.conversation_has_personal_attack", "first_two"),
    "persuasion": ("utterance", "meta.success", "first_two"),
    "discourse": ("utterance", "meta.majority_type", "first_two"),
}

# (context, label)
csv_column_map = {
    "hate": ("post", "implicit_class"),
    "wikievents": ("context", "targets"),
    "reframe": ("original_text", "reframed_text"),
    "humor": (3, 1),
    "flute-explanation": (("premise", "hypothesis"), ("label", "explanation")),
    "flute-classification": (("premise", "hypothesis"), "type"),
    "media_ideology": ("content", "bias_text"),
    "hippocorpus": ("story", "events"),
    "indian_english_dialect": ("Example", "Feature"),
    "ibc": ("sentence", "leaning"),
    "semeval_stance": ("Tweet", "trump_stance"),
    "tempowic": (("text1", "text2", "word"), "label"),
    "talklife": (("Seeker", "Response"), "remap_level"),
    "sbic": ("post", "targetStereotype"),
    "raop": ("sents", "labels"),
    "emotion": ("text", "labels"),
    "mrf-explanation": ("headline", "writer_intent"),
    "mrf-classification": ("headline", "gold_label"),
    "tropes": ("Quotes", "Tropes"),
    "positive_reframing": ("original_text", "reframed_text")
}

drop_labels = {
    "flute-classification": ["CreativeParaphrase"],
    "flute-explanation": ["CreativeParaphrase"],
    "hate": ["other"],
    "discourse": ["announcement", "other", "negativereaction"],
}

csv_download = {
    "hate": "https://huggingface.co/datasets/SALT-NLP/ImplicitHate/raw/main/implicit_hate.csv",
    "reframe": "https://raw.githubusercontent.com/SALT-NLP/positive-frames/main/data/wholetest.csv",
    "humor": "https://raw.githubusercontent.com/orionw/RedditHumorDetection/master/data/reddit_full/test.tsv",
    "positive_reframing": "https://huggingface.co/datasets/SALT-NLP/positive_reframing/raw/main/positive_reframing.csv",
}

jsonl_download = {
    "flute-explanation": "https://huggingface.co/datasets/ColumbiaNLP/FLUTE/raw/main/train.jsonl",
    "flute-classification": "https://huggingface.co/datasets/ColumbiaNLP/FLUTE/raw/main/train.jsonl",
    # "tempowic": "https://github.com/cardiffnlp/TempoWiC/raw/main/data/test-codalab-10k.data.jl",
}

labelsets = {
    "power": ["True", "False"],
    "stance": ["True", "False"],
    "politeness": ["A", "B", "C"],
    "persuasion": ["True", "False"],
    "discourse": ["A", "B", "C", "D", "E", "F", "G"],
    "hate": ["A", "B", "C", "D", "E", "F"],
    "reframe": None,
    "humor": ["True", "False"],
    "conv_go_awry": ["True", "False"],
    "flute-explanation": None,
    "flute-classification": ["A", "B", "C", "D"],
    "media_ideology": ["A", "B", "C"],
    "hippocorpus": [
        "A",
        "B",
        "C",
        "D",
        "E",
        "F",
        "G",
        "H",
        "I",
        "J",
        "K",
        "L",
        "M",
        "N",
        "O",
        "P",
        "Q",
        "R",
        "S",
        "T",
        "U",
        "V",
        "W",
        "X",
        "Y",
        "Z",
        "AA",
        "AB",
        "AC",
        "AD",
        "AE",
        "AF",
        "AG",
        "AH",
        "AI",
        "AJ",
        "AK",
        "AL",
        "AM",
        "AN",
        "AO",
        "AP",
        "AQ",
        "AR",
        "AS",
        "AT",
        "AU",
        "AV",
        "AW",
        "AX",
        "AY",
        "AZ",
        "BA",
        "BB",
        "BC",
        "BD",
        "BE",
        "BF",
        "BG",
        "BH",
        "BI",
        "BJ",
        "BK",
        "BL",
        "BM",
        "BN",
        "BO",
        "BP",
        "BQ",
        "BR",
        "BS",
        "BT",
    ],
    "indian_english_dialect": [
        "A",
        "B",
        "C",
        "D",
        "E",
        "F",
        "G",
        "H",
        "I",
        "J",
        "K",
        "L",
        "M",
        "N",
        "O",
        "P",
        "Q",
        "R",
        "S",
        "T",
        "U",
        "V",
    ],
    "ibc": ["A", "B", "C"],
    "semeval_stance": ["A", "B", "C"],
    "tempowic": ["A", "B"],
    "wikievents": None,
    "sbic": None,
    "talklife": ["A", "B", "C"],  # fix!
    "raop": ["A", "B", "C", "D", "E", "F", "G"],  # fix!
    "emotion": ["A", "B", "C", "D", "E", "F"],
    "mrf-explanation": None,
    "mrf-classification": ["A", "B"],
    "tropes": [
        "A",
        "B",
        "C",
        "D",
        "E",
        "F",
        "G",
        "H",
        "I",
        "J",
        "K",
        "L",
        "M",
        "N",
        "O",
        "P",
        "Q",
        "R",
        "S",
        "T",
        "U",
        "V",
        "W",
        "X",
        "Y",
        "Z",
        "AA",
        "AB",
        "AC",
        "AD",
        "AE",
        "AF",
        "AG",
        "AH",
        "AI",
        "AJ",
        "AK",
        "AL",
        "AM",
        "AN",
        "AO",
        "AP",
        "AQ",
        "AR",
        "AS",
        "AT",
        "AU",
        "AV",
        "AW",
        "AX",
        "AY",
        "AZ",
        "BA",
        "BB",
        "BC",
        "BD",
        "BE",
        "BF",
        "BG",
        "BH",
        "BI",
        "BJ",
        "BK",
        "BL",
        "BM",
        "BN",
        "BO",
        "BP",
        "BQ",
        "BR",
        "BS",
        "BT",
    ],
    "positive_reframing": None,
}

prompts_templates = {
    "power": """

Based on the conversation is {$speaker} in a position of power? 

Constraint: Even if you are uncertain, you must pick either True or False with without using any other words.""",
    "stance": """

Does {$speaker} support {$title}? 
True
False

Constraint: Even if you are uncertain, you must pick either True or False with without using any other words.
""",  # This Dataset is strange currently, too long of context
    "politeness": """

Based on formal workplace social norms, which of the following best describes the above conversation? 
A: Polite
B: Neutral
C: Impolite

Constraint: Answer with only the option above that is most accurate and nothing else.
""",
    "wikievents": """

{$blanks}

Replace the BLANKs with the extracted information about the event in <tgr> . Leave the keys of the JSON unchanged.

JSON Output: 
""",
    "persuasion": """

If you were the original poster, would this reply convince you?
True
False

Constraint: Even if you are uncertain, you must pick either True or False with without using any other words.
""",
    "discourse": """

Which of the following best characterizes the discourse type of the previous statement?
A: Question
B: Answer
C: Agreement
D: Disagreement
E: Appreciation
F: Elaboration
G: Humor

Constraint: Answer with only the option above that is most accurate and nothing else.
""",
    "hate": """

Which of the following categories of hate speech best describes the sentence above?
A: White Grievance (frustration over a minority group’s perceived privilege and casting majority groups as the real victims of racism)
B: Incitement to Violence (flaunting in-group unity and power or elevating known hate
groups and ideologies)
C: Inferiority Language (implies one group or individual is inferior to another, including dehumanization and toxification)
D: Irony (the use of sarcasm, humor, and satire to attack or demean a protected class or individual)
E: Stereotypes and Misinformation (associating a protected class with negative attributes)
F: Threatening and Intimidation (conveys a speaker commitment to a target’s pain, injury, damage, loss, or violation of rights)

Constraint: Answer with one or more of the options above that is most accurate and nothing else. Always choose at least one of the options.

""",
    "reframe": """

Reframe the previous sentence positively, without changing the core meaning of the sentence.
""",
    "humor": """ 

Constraint: Is the above joke humorous to most of the people? You must pick between \"True\" or \"False\" You cannot use any other words except for \"True\" or \"False\" 
""",
    "toxicity": """

Will the previouse conversation eventually derail into a personal attack?  

Constraint: Even if you are uncertain, you must pick either \"True\" or \"False\" with without using any other words.
""",
    "flute-explanation": """

Which of the following describes the relationship between the premise and the hypothesis?
A: Entailment
B: Contradiction

Constraint: Explain the figurative language in the hypothesis in one sentence and then answer with which option is the most accurate.
""",
    "flute-classification": """

Which of the following best describes the type of figurative language used in the hypothesis?
A: Idiom
B: Metaphor
C: Sarcasm
D: Simile

Constraint: Answer with only the option above that is most accurate and nothing else.
""",
    "media_ideology": """" 

Which of the following leanings would a political scientist say that the above article has?
A: Left
B: Right
C: Center

Constraint: Answer with only the option above that is most accurate and nothing else.
""",
    "hippocorpus": """This is an Event Extraction task. Which sentences above indicate new events?""",
    "indian_english_dialect": """
    
Which of the following features would a linguist say that the above sentence has?
A: Article Omission (e.g., 'Person I like most is here.')
B: Copula Omission (e.g., 'Everything busy in our life.')
C: Direct Object Pronoun Drop (e.g., 'He didn’t give me.')
D: Extraneous Article (e.g, 'Educated people get a good money.')
E: Focus Itself (e.g, 'I did it in the month of June itself.')
F: Focus Only (e.g, 'I was there yesterday only'.)
G: General Extender "and all" (e.g, 'My parents and siblings and all really enjoy it'.)
H: Habitual Progressive (e.g., 'They are getting H1B visas.')
I: Invariant Tag "isn’t it, no, na" (e.g., 'It’s come from me, no?')
J: Inversion In Embedded Clause (e.g., 'The school called to ask when are you going back.')
K: Lack Of Agreement (e.g., 'He talk to them.')
L: Lack Of Inversion In Wh-questions (e.g., 'What are you doing?')
M: Left Dislocation (e.g., 'My parents, they really enjoy playing board games.')
N: Mass Nouns As Count Nouns (e.g., 'They use proper grammars there.')
O: Non-initial Existential "is / are there" (e.g., 'Every year inflation is there.')
P: Object Fronting (e.g., 'In fifteen years, lot of changes we have seen.')
Q: Prepositional Phrase Fronting With Reduction (e.g., 'First of all, right side we can see a plate.')
R: Preposition Omission (e.g., 'I stayed alone two years.')
S: Resumptive Object Pronoun (e.g., 'Some teachers when I was in school I liked them very much.')
T: Resumptive Subject Pronoun (e.g., 'A person living in Calcutta, which he didn’t know Hindi earlier, when he comes to Delhi he has to learn English.')
U: Stative Progressive (e.g., 'We will be knowing how much the structure is getting deflected.')
V: Topicalized Non-argument Constituent (e.g., 'in the daytime I work for the courier service')
W: None of the above

Constraint: Answer with only the option above that is most accurate and nothing else.
""",
    "ibc": """" 

Which of the following leanings would a political scientist say that the above article has?
A: Liberal
B: Conservative
C: Neutral

Constraint: Answer with only the option above that is most accurate and nothing else.""",
    "semeval_stance": """

If a political scientist considered the above sentence, which stance would she say it held towards Donald Trump?
A: Against
B: Favor
C: None

Constraint: Answer with only the option above that is most accurate and nothing else.""",
    "tempowic": """
    
If a linguist considered the word above in text1 and text2, would she consider the meaning of this word to be the
A: Same
B: Different

Constraint: Answer with only the option above that is most accurate and nothing else.""",
    "sbic": """
    
Explain the stereotype that this sentence communicates? Please use a single sentence like "[GROUP] are ______"
""",
    "talklife": """

Explorations are when a mental health counselor shows active interest in a seeker by asking about unstated experiences. What level of exploration is expressed in the counselor's message above?
A: Strong exploration (specifically labels the seeker’s experiences and feelings, like "Are you feeling alone right now?")
B: Weak exploration (a generic question, like "What happened?")
C: No exploration

Constraint: Answer with only the option above that is most accurate and nothing else.""",
    "raop": """

If a user in a community for giving away free pizza saw the above sentence, they would consider it to use which of the following persuasion strategies?
A: Evidence (Providing concrete facts or evidence for the narrative or request, like "There is a Pizza Hut and a Dominos near me.")
B: Politeness (The usage of polite language in requests, like "Thank you so much!")
C: Reciprocity (Responding to a positive action with another positive action. People are more likely to help if they have received help themselves. Example messages are like "I’ll pay it forward with my first check")
D: Impact (Emphasizing the importance or impact of the request, like "I'll use this pizza to celebrate")
E: Emotion (Making requests full of emotional valence and arousal affect to influence others, like "I’ve been in the lowest depressive state of my life")
F: Scarcity (People emphasizing on the urgency, rare of their needs, like "I haven’t ate a meal in two days")
G: Other

Constraint: Answer with only the option above that is most accurate and nothing else.""",
    "emotion": """

If a mental health professional saw the above text, what emotion would they categorize it to be (using the following six basic emotions according to Paul Ekman)?
A: Fear
B: Anger
C: Joy
D: Sadness
E: Love
F: Surprise

Constraint: Answer with only the option above that is most accurate and nothing else.""",
    "mrf-explanation": """
    
What is the implied message of the above news headline? 

Constraint: Answer with a short phrase like "some masks are better than others."
""",
    "mrf-classification": """
    
Which of the following describes the above news headline?
A: Misinformation
B: Trustworthy

Constraint: Answer with only the option above that is most accurate and nothing else.""",
    "tropes": """

Given quotes from the character above, which of the following tropes would you say this character represents?

A: Absent Minded Professor
B: Adventurer Archaeologist
C: Arrogant Kungfu Guy
D: Big Man On Campus
E: Bounty Hunter
F: Brainless Beauty
G: Broken Bird
H: Bromantic Foil
I: Bruiser With A Soft Center
J: Bully
K: Byronic Hero
L: Casanova
M: Chanteuse
N: Charmer
O: Child Prodigy
P: Classy Cat Burglar
Q: Consummate Professional
R: Corrupt Corporate Executive
S: Coward
T: Crazy Jealous Guy
U: Crazy Survivalist
V: Cultured Badass
W: Dean Bitterman
X: Dirty Cop
Y: Ditz
Z: Doormat
AA: Drill Sargeant Nasty
AB: Dumb Blonde
AC: Dumb Muscle
AD: Eccentric Mentor
AE: Egomaniac Hunter
AF: Evil Prince
AG: Fastest Gun In The West
AH: Father To His Men
AI: Final Girl
AJ: Gadgeteer Genius
AK: Gentleman Thief
AL: Granola Person
AM: Grumpy Old Man
AN: Hardboiled Detective
AO: Heartbroken Badass
AP: Henpecked Husband
AQ: Hitman With A Heart
AR: Jerk Jock
AS: Junkie Prophet
AT: Klutz
AU: Loser Protagonist
AV: Loveable Rogue
AW: Master Swordsman
AX: Morally Bankrupt Banker
AY: Officer And A Gentleman
AZ: Ophelia
BA: Playful Hacker
BB: Prima Donna
BC: Psycho For Hire
BD: Pupil Turned To Evil
BE: Retired Outlaw
BF: Revenge
BG: Romantic Runnerup
BH: Self Made Man
BI: Slacker
BJ: Stoner
BK: Storyteller
BL: Stupid Crooks
BM: Surfer Dude
BN: The Chief
BO: The Editor
BP: Tranquil Fury
BQ: Trickster
BR: Valley Girl
BS: Warrior Poet
BT: Young Gun
""",
    "positive_reframing": """

Rephrase the above sentence to be more positive using the following strategies: {$strategy}

Strategies are defined below

growth: viewing a challenges as an opportunities for the author to grow or improve themselves
impermanence: saying bad things don't last forever, will get better soon, and/or that others have experienced similar struggles
neutralizing: replacing a negative word with a neutral word
optimism: focusing on things about the situation itself, in that moment, that are good
self-affirmation: talking about what strengths the author already has, or the values they admire
thankfulness: expressing gratitude

""",
}
""