config.yaml

###### general
fps: 2
videos_dir: "dataset/videos"
video_feature_dir: "dataset/features"
situation_split_dir: "dataset/situations/splits"
situation_feature_dir: "dataset/situations/features"
situation_clips_dir: "dataset/situations/clips"
speaker_split_dir: "dataset/speaker/splits"
speaker_feature_dir: "dataset/speaker/features"
speaker_clips_dir: "dataset/speaker/clips"
######

###### features.py
wikifier_key: ""
whisper_model: "large-v2"
tmp_dir: "feature_extraction/tmp/"
spk_diarization_script: "whisper-diarization-main/diarize.py"
CLIP_queries: "feature_extraction/CLIP_queries.json"
pos_dict:
  ADJ: 0
  ADP: 1
  ADV: 2
  AUX: 3
  CONJ: 4
  CCONJ: 4
  SCONJ: 4
  DET: 5
  INTJ: 6
  NOUN: 7
  NUM: 8  
  PART: 9
  PRON: 10
  PROPN: 11
  VERB: 12
  X: 13
######

source_abbreviations:
  - com
  - bild
  - tag

speaker_hierarchy_mapping_0:
  anchor: anchor
  reporter: reporter
  doctor-nurse: other
  expert-medicine: other
  expert-other: other
  layperson: other
  layperson-4-covid: other
  layperson-x-covid: other
  police-frwr: other
  politician-ampel: other
  politician-other: other
  politician-n-de: other
  politician-right: other
  celeb: other

speaker_hierarchy_mapping_1:
  anchor: anchor
  reporter: reporter
  doctor-nurse: expert
  expert-medicine: expert
  expert-other: expert
  layperson: layperson
  layperson-4-covid: layperson
  layperson-x-covid: layperson
  police-frwr: other
  politician-ampel: politician
  politician-other: politician
  politician-n-de: politician
  politician-right: politician
  celeb: other

groundtruth_numerical_speaker:
  anchor: 0
  reporter: 1
  other: 2
  expert: 3
  layperson: 4
  politician: 5

groundtruth_numerical_situations:
  talking-head: 0
  voiceover: 1
  interview: 2
  commenting: 3
  speech: 4

window_lengths:
  - 5
  - 10
  - 20
  - 40

context_size: 1

seed: 100
k: 10

labels_0:
  - anchor
  - reporter
  - other

labels_1:
  - anchor
  - reporter
  - other
  - expert
  - layperson
  - politician

labels_situations:
  - talking-head
  - voiceover
  - interview
  - commenting
  - speech

feature_names_speaker:
  - clip_Anchor
  - clip_Reporter
  - clip_PoliceFrwr
  - clip_DoctorNurse
  - clip_Xcov
  - clip_4cov
  - clip_Layperson
  - clip_ExpMedicine
  - clip_ExpOth
  - clip_Polit
  - clip_PolitAmpel
  - clip_PolitRight
  - clip_PolitOther
  - clip_Polit-n-de
  - clip_Celeb
  - clip_Talkinghead
  - clip_Interview
  - clip_Speech
  - clip_Commenting
  - clip_Voiceover
  - clip_Studio
  - clip_Indoor
  - clip_Outdoor
  - LengthOfSpeech
  - Avg_Shotdensity
  - NumberOfShots
  - PositiveSentProb
  - NegativeSentProb
  - NeutralSentProb
  - pos_ADJ 
  - pos_ADP 
  - pos_ADV 
  - pos_AUX 
  - pos_CONJ
  - pos_DET 
  - pos_INTJ 
  - pos_NOUN 
  - pos_NUM   
  - pos_PART 
  - pos_PRON
  - pos_PROPN
  - pos_VERB
  - pos_X
  - ner_EPER
  - ner_LPER
  - ner_LOC
  - ner_ORG
  - ner_MISC
  - speaker_before_img_sim
  - speaker_after_img_sim
  - speaker_before_sent_sim
  - speaker_after_sent_sim

feature_names_situations:
  - clip_Anchor
  - clip_Reporter
  - clip_PoliceFrwr
  - clip_DoctorNurse
  - clip_Xcov
  - clip_4cov
  - clip_Layperson
  - clip_ExpMedicine
  - clip_ExpOth
  - clip_Polit
  - clip_PolitAmpel
  - clip_PolitRight
  - clip_PolitOther
  - clip_Polit-n-de
  - clip_Celeb
  - clip_Talkinghead
  - clip_Interview
  - clip_Speech
  - clip_Commenting
  - clip_Voiceover
  - clip_Studio
  - clip_Indoor
  - clip_Outdoor
  - LengthOfSpeech
  - Avg_Shotdensity
  - NumberOfShots
  - PositiveSentProb
  - NegativeSentProb
  - NeutralSentProb
  - pos_ADJ 
  - pos_ADP 
  - pos_ADV 
  - pos_AUX 
  - pos_CONJ
  - pos_DET 
  - pos_INTJ 
  - pos_NOUN 
  - pos_NUM   
  - pos_PART 
  - pos_PRON
  - pos_PROPN
  - pos_VERB
  - pos_X
  - ner_EPER
  - ner_LPER
  - ner_LOC
  - ner_ORG
  - ner_MISC
  - speaker_before_img_sim
  - speaker_after_img_sim
  - speaker_before_sent_sim
  - speaker_after_sent_sim
######